Merge branch 'master' into xsn/qwen3-tts

common/arg.cpp

@@ -27,6 +27,7 @@
 #include <algorithm>
 #include <cinttypes>
 #include <climits>
+#include <cmath>
 #include <cstdarg>
 #include <filesystem>
 #include <fstream>
@@ -2021,7 +2022,13 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
         {"--repeat-penalty"}, "N",
         string_format("penalize repeat sequence of tokens (default: %.2f, 1.0 = disabled)", (double)params.sampling.penalty_repeat),
         [](common_params & params, const std::string & value) {
-            params.sampling.penalty_repeat = std::stof(value);
+            const float penalty_repeat = std::stof(value);
+            if (!std::isfinite(penalty_repeat) ||
+                penalty_repeat <= 0.0f ||
+                !std::isfinite(1.0f/penalty_repeat)) {
+                throw std::runtime_error("error: repeat-penalty must be finite and greater than 0\n");
+            }
+            params.sampling.penalty_repeat = penalty_repeat;
             params.sampling.user_sampling_config |= common_params_sampling_config::COMMON_PARAMS_SAMPLING_CONFIG_PENALTY_REPEAT;
         }
     ).set_sampling());
@@ -2029,14 +2036,22 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
         {"--presence-penalty"}, "N",
         string_format("repeat alpha presence penalty (default: %.2f, 0.0 = disabled)", (double)params.sampling.penalty_present),
         [](common_params & params, const std::string & value) {
-            params.sampling.penalty_present = std::stof(value);
+            const float penalty_present = std::stof(value);
+            if (!std::isfinite(penalty_present)) {
+                throw std::runtime_error("error: presence-penalty must be finite\n");
+            }
+            params.sampling.penalty_present = penalty_present;
         }
     ).set_sampling());
     add_opt(common_arg(
         {"--frequency-penalty"}, "N",
         string_format("repeat alpha frequency penalty (default: %.2f, 0.0 = disabled)", (double)params.sampling.penalty_freq),
         [](common_params & params, const std::string & value) {
-            params.sampling.penalty_freq = std::stof(value);
+            const float penalty_freq = std::stof(value);
+            if (!std::isfinite(penalty_freq)) {
+                throw std::runtime_error("error: frequency-penalty must be finite\n");
+            }
+            params.sampling.penalty_freq = penalty_freq;
         }
     ).set_sampling());
     add_opt(common_arg(
@@ -2552,7 +2567,7 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
             params.mtmd_batch_max_tokens = value;
         }
     ).set_examples({LLAMA_EXAMPLE_SERVER}).set_env("LLAMA_ARG_MTMD_BATCH_MAX_TOKENS"));
-    if (llama_supports_rpc()) {
+    if (params.is_gen_docs || llama_supports_rpc()) {
         add_opt(common_arg(
             {"--rpc"}, "SERVERS",
             "comma-separated list of RPC servers (host:port)",
@@ -3287,7 +3302,7 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
         {"--tools"}, "TOOL1,TOOL2,...",
         "experimental: whether to enable built-in tools for AI agents - do not enable in untrusted environments (default: no tools)\n"
         "specify \"all\" to enable all tools\n"
-        "available tools: read_file, file_glob_search, grep_search, exec_shell_command, write_file, edit_file, get_datetime\n"
+        "available tools: read_file, file_glob_search, grep_search, exec_shell_command, write_file, edit_file, get_datetime, get_info\n"
         "note: for security reasons, this will limit --cors-origins to localhost by default",
         [](common_params & params, const std::string & value) {
             params.server_tools = parse_csv_row(value);

common/common.cpp

@@ -998,6 +998,23 @@ bool fs_is_directory(const std::string & path) {
     return std::filesystem::exists(dir) && std::filesystem::is_directory(dir);
 }
 
+std::string common_get_env(const std::string & name) {
+    const char * value = std::getenv(name.c_str());
+    return value == nullptr ? "" : value;
+}
+
+void common_set_env(const std::string & name, const std::string & value) {
+#if defined(_WIN32)
+    _putenv_s(name.c_str(), value.c_str());
+#else
+    if (value.empty()) {
+        unsetenv(name.c_str());
+    } else {
+        setenv(name.c_str(), value.c_str(), 1);
+    }
+#endif
+}
+
 std::string fs_get_cache_directory() {
     std::string cache_directory = "";
     auto ensure_trailing_slash = [](std::string p) {
@@ -1299,8 +1316,9 @@ common_init_result::common_init_result(common_params & params, bool model_only)
     pimpl->samplers.resize(cparams.n_seq_max);
     pimpl->samplers_seq_config.resize(cparams.n_seq_max);
 
+    const int32_t n_ctx = cparams.n_ctx > 0 ? (int32_t) cparams.n_ctx : llama_model_n_ctx_train(model);
     for (int i = 0; i < (int) cparams.n_seq_max; ++i) {
-        pimpl->samplers[i].reset(common_sampler_init(model, params.sampling));
+        pimpl->samplers[i].reset(common_sampler_init(model, params.sampling, n_ctx));
         pimpl->samplers_seq_config[i] = { i, common_sampler_get(pimpl->samplers[i].get()) };
     }
 
@@ -1462,18 +1480,18 @@ common_init_result_ptr common_init_from_params(common_params & params, bool mode
 common_init_result::~common_init_result() = default;
 
 std::string common_get_model_endpoint() {
-    const char * model_endpoint_env = getenv("MODEL_ENDPOINT");
-    // We still respect the use of environment-variable "HF_ENDPOINT" for backward-compatibility.
-    const char * hf_endpoint_env = getenv("HF_ENDPOINT");
-    const char * endpoint_env = model_endpoint_env ? model_endpoint_env : hf_endpoint_env;
-    std::string model_endpoint = "https://huggingface.co/";
-    if (endpoint_env) {
-        model_endpoint = endpoint_env;
-        if (model_endpoint.back() != '/') {
-            model_endpoint += '/';
-        }
+    std::string endpoint = common_get_env("MODEL_ENDPOINT");
+    if (endpoint.empty()) {
+        // the HF_ENDPOINT variable is respected for backward compatibility
+        endpoint = common_get_env("HF_ENDPOINT");
     }
-    return model_endpoint;
+    if (endpoint.empty()) {
+        return "https://huggingface.co/";
+    }
+    if (endpoint.back() != '/') {
+        endpoint += '/';
+    }
+    return endpoint;
 }
 
 char * common_get_model_or_exit(int argc, char * argv[]) {

common/common.h

@@ -734,6 +734,8 @@ struct common_params {
     // TTS params
     std::string tts_lang = "";
     std::string tts_speaker_file = "";
+
+    bool is_gen_docs = false; // whether we are running inside llama-gen-docs
 };
 
 // call once at the start of a program if it uses libcommon
@@ -858,6 +860,15 @@ std::string string_from(const struct llama_context * ctx, const struct llama_bat
 
 bool glob_match(const std::string & pattern, const std::string & str);
 
+//
+// Environment utils
+//
+
+// portable environment access, an unset variable reads as an empty string
+// and setting an empty value unsets the variable
+std::string common_get_env(const std::string & name);
+void        common_set_env(const std::string & name, const std::string & value);
+
 //
 // Filesystem utils
 //

common/sampling.cpp

@@ -184,9 +184,26 @@ std::string common_params_sampling::print() const {
     return std::string(result);
 }
 
-struct common_sampler * common_sampler_init(const struct llama_model * model, struct common_params_sampling & params) {
-    const llama_vocab * vocab = llama_model_get_vocab(model);
+struct common_sampler * common_sampler_init(
+        const struct llama_model * model,
+        struct common_params_sampling & params,
+        int32_t n_ctx) {
+    if (!std::isfinite(params.penalty_repeat) ||
+        params.penalty_repeat <= 0.0f ||
+        !std::isfinite(1.0f/params.penalty_repeat)) {
+        throw std::invalid_argument("penalty_repeat must be finite and greater than 0");
+    }
+    if (!std::isfinite(params.penalty_freq)) {
+        throw std::invalid_argument("penalty_freq must be finite");
+    }
+    if (!std::isfinite(params.penalty_present)) {
+        throw std::invalid_argument("penalty_present must be finite");
+    }
+    if (params.penalty_last_n == -1) {
+        params.penalty_last_n = n_ctx > 0 ? n_ctx : llama_model_n_ctx_train(model);
+    }
 
+    const llama_vocab * vocab = llama_model_get_vocab(model);
     llama_sampler_chain_params lparams = llama_sampler_chain_default_params();
 
     lparams.no_perf = params.no_perf;

common/sampling.h

@@ -37,7 +37,10 @@ struct common_sampler;
 // llama_sampler API overloads
 
 // note: can mutate params in some cases
-struct common_sampler * common_sampler_init(const struct llama_model * model, struct common_params_sampling & params);
+struct common_sampler * common_sampler_init(
+        const struct llama_model * model,
+        struct common_params_sampling & params,
+        int32_t n_ctx = 0);
 
 void common_sampler_free(struct common_sampler * gsmpl);
 

conversion/deepseek.py

@@ -447,12 +447,43 @@ class DeepseekV2Model(TextModel):
 class DeepseekV32Model(DeepseekV2Model):
     model_arch = gguf.MODEL_ARCH.DEEPSEEK32
     skip_mtp = False
+    supports_mtp_export = True
+    _n_main_layers: int | None = None
 
     def __init__(self, *args, **kwargs):
         super().__init__(*args, **kwargs)
-        self.block_count = self.hparams["num_hidden_layers"] + self.hparams.get("num_nextn_predict_layers", 0)
+        self.block_count = self.hparams["num_hidden_layers"]
+        if not self.no_mtp:
+            self.block_count += self.hparams.get("num_nextn_predict_layers", 0)
         self.tensor_map = gguf.get_tensor_name_map(self.model_arch, self.block_count)
 
+    def index_tensors(self, remote_hf_model_id: str | None = None):
+        type(self)._n_main_layers = self.hparams["num_hidden_layers"]
+        return super().index_tensors(remote_hf_model_id=remote_hf_model_id)
+
+    @classmethod
+    def filter_tensors(cls, item: tuple[str, Callable[[], Tensor]]) -> tuple[str, Callable[[], Tensor]] | None:
+        if (titem := super().filter_tensors(item)) is None:
+            return None
+        name, gen = titem
+
+        # DeepSeek V3.2 appends the NextN/MTP block past num_hidden_layers
+        # (model.layers.61 -> blk.61 in the 62-block file).
+        assert cls._n_main_layers is not None
+        is_mtp = (m := re.match(r"model\.layers\.(\d+)\.", name)) is not None and int(m.group(1)) >= cls._n_main_layers
+
+        # --no-mtp: drop the appended NextN block entirely.
+        if is_mtp and cls.no_mtp:
+            return None
+        # --mtp: keep ONLY NextN-block tensors plus the shared embeddings/
+        # norm/lm_head (so the resulting GGUF carries just the draft head).
+        if cls.mtp_only and not is_mtp and name not in (
+            "model.embed_tokens.weight", "model.norm.weight", "lm_head.weight",
+        ):
+            return None
+
+        return name, gen
+
     def set_vocab(self):
         from transformers import AutoTokenizer
         tokenizer = AutoTokenizer.from_pretrained(self.dir_model)
@@ -463,7 +494,7 @@ class DeepseekV32Model(DeepseekV2Model):
         super().set_gguf_parameters()
 
         # NextN/MTP prediction layers
-        if (num_nextn_predict_layers := self.hparams.get("num_nextn_predict_layers")) is not None:
+        if not self.no_mtp and (num_nextn_predict_layers := self.hparams.get("num_nextn_predict_layers")) is not None:
             self.gguf_writer.add_nextn_predict_layers(num_nextn_predict_layers)
 
         # DSA indexer parameters

conversion/glm.py

@@ -206,10 +206,70 @@ class Glm4MoeModel(TextModel):
 @ModelBase.register("Glm4MoeLiteForCausalLM")
 class Glm4MoeLiteModel(DeepseekV2Model):
     model_arch = gguf.MODEL_ARCH.DEEPSEEK2
+    skip_mtp = False
+    supports_mtp_export = True
+    _n_main_layers: int | None = None
 
     def set_vocab(self):
         return self._set_vocab_glm()
 
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+
+        num_hidden_layers = self.hparams["num_hidden_layers"]
+        self.num_nextn_predict_layers = self.hparams.get("num_nextn_predict_layers", 0)
+        self.skip_mtp = self.no_mtp or self.num_nextn_predict_layers == 0
+
+        if self.skip_mtp:
+            self.block_count = num_hidden_layers
+        else:
+            self.block_count = num_hidden_layers + self.num_nextn_predict_layers
+
+        self.tensor_map = gguf.get_tensor_name_map(self.model_arch, self.block_count)
+
+    def set_gguf_parameters(self):
+        super().set_gguf_parameters()
+
+        if self.skip_mtp:
+            return
+
+        self.gguf_writer.add_nextn_predict_layers(self.num_nextn_predict_layers)
+
+    def index_tensors(self, remote_hf_model_id: str | None = None):
+        type(self)._n_main_layers = self.hparams["num_hidden_layers"]
+        return super().index_tensors(remote_hf_model_id=remote_hf_model_id)
+
+    @classmethod
+    def filter_tensors(cls, item):
+        if (titem := super().filter_tensors(item)) is None:
+            return None
+        name, gen = titem
+
+        if cls._n_main_layers is not None:
+            match = re.match(r"model\.layers\.(\d+)\.", name)
+            is_mtp = match is not None and int(match.group(1)) >= cls._n_main_layers
+            if is_mtp and cls.no_mtp:
+                return None
+            if cls.mtp_only and not is_mtp and name not in (
+                "model.embed_tokens.weight", "model.norm.weight", "lm_head.weight",
+            ):
+                return None
+
+        return name, gen
+
+    def prepare_metadata(self, vocab_only: bool):
+        from_dir = self.fname_out.is_dir()
+        super().prepare_metadata(vocab_only=vocab_only)
+
+        if not self.mtp_only or not from_dir:
+            return
+
+        output_type: str = self.ftype.name.partition("_")[2]
+        fname_default: str = gguf.naming_convention(
+            self.metadata.name, self.metadata.basename, self.metadata.finetune,
+            self.metadata.version, size_label=None, output_type=output_type, model_type=None)
+        self.fname_out = self.fname_out.parent / f"mtp-{fname_default}.gguf"
+
 
 @ModelBase.register("GlmMoeDsaForCausalLM")
 class GlmMoeDsaModel(DeepseekV2Model):

conversion/qwen.py

@@ -268,8 +268,101 @@ class Qwen3MoeModel(Qwen2MoeModel):
         super().set_vocab()
 
 
+class _QwenMtpMixin:
+    """Shared MTP wiring for Qwen3-Next and Qwen3.5/3.6 text variants. The HF
+    config carries the MTP block under `mtp_num_hidden_layers` (computed from
+    the checkpoint when absent, e.g. Qwen3-Next) and the tensors under
+    `mtp.*`; we extend block_count, emit the nextn metadata key, and remap
+    `mtp.*` to the standard layer-indexed nextn naming so the existing
+    tensor_map handles them."""
+
+    supports_mtp_export = True
+    hparams: dict[str, Any]
+    model_arch: gguf.MODEL_ARCH
+    gguf_writer: gguf.GGUFWriter
+    block_count: int
+    tensor_map: gguf.TensorNameMap
+    no_mtp: bool
+    mtp_only: bool
+    _original_block_count: int | None = None
+    opt_num_mtp_layers: int = 0
+
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+        self.block_count = self.hparams["num_hidden_layers"]
+        if not self.no_mtp:
+            n_mtp = self.hparams.get("mtp_num_hidden_layers", 0)
+            # Qwen-3-Next doesn't include `mtp_num_hidden_layers` in config.
+            if n_mtp == 0:
+                assert self.opt_num_mtp_layers != 0
+                n_mtp = self.opt_num_mtp_layers
+            self.block_count += n_mtp
+        self.tensor_map = gguf.get_tensor_name_map(self.model_arch, self.block_count)
+
+    def index_tensors(self, remote_hf_model_id: str | None = None) -> dict[str, Callable[[], Tensor]]:
+        hparams = {**self.hparams, **self.hparams.get("text_config", {})}
+        key = next((k for k in ["n_layers", "num_hidden_layers", "n_layer", "num_layers"] if k in hparams), None)
+        type(self)._original_block_count = hparams.get(key)
+        type(self).opt_num_mtp_layers = 0
+        return super().index_tensors(remote_hf_model_id=remote_hf_model_id)  # ty: ignore[unresolved-attribute]
+
+    @classmethod
+    def filter_tensors(cls, item):
+        assert cls._original_block_count is not None
+        # TODO: change TextModel to super()
+        if (titem := TextModel.filter_tensors(item)) is None:
+            return None
+        name, gen = titem
+        if name.startswith("model.mtp."):
+            name = name.replace("model.", "", 1)
+        if name.startswith("mtp."):
+            if cls.no_mtp:
+                return None
+            remapper = {
+                "fc":                    "eh_proj",
+                "pre_fc_norm_embedding": "enorm",
+                "pre_fc_norm_hidden":    "hnorm",
+                "norm":                  "shared_head.norm",
+            }
+            parts = name.split(".", 3)
+            if len(parts) == 4 and parts[1] == "layers" and parts[2].isdecimal():
+                mtp_idx = int(parts[2])
+                name = f"model.layers.{cls._original_block_count + mtp_idx}.{parts[3]}"
+                cls.opt_num_mtp_layers = max(cls.opt_num_mtp_layers, mtp_idx + 1)
+            elif len(parts) == 3 and parts[1] in remapper:
+                name = f"model.layers.{cls._original_block_count}.{remapper[parts[1]]}.{parts[2]}"
+        elif cls.mtp_only:
+            keep = name in (
+                "model.embed_tokens.weight", "model.norm.weight", "lm_head.weight",
+                "embed_tokens.weight", "norm.weight",
+            )
+            if not keep:
+                return None
+        return name, gen
+
+    def set_gguf_parameters(self):
+        super().set_gguf_parameters()  # ty: ignore[unresolved-attribute]
+        if self.no_mtp:
+            return
+        if (n := self.block_count - self.hparams["num_hidden_layers"]) > 0:
+            self.gguf_writer.add_nextn_predict_layers(n)
+
+    def prepare_metadata(self, vocab_only: bool):
+        from_dir = self.fname_out.is_dir()
+        super().prepare_metadata(vocab_only=vocab_only)  # ty: ignore[unresolved-attribute]
+
+        if not self.mtp_only or not from_dir:
+            return
+
+        output_type: str = self.ftype.name.partition("_")[2]  # pyright: ignore[reportAttributeAccessIssue] # ty: ignore[unresolved-attribute]
+        fname_default: str = gguf.naming_convention(
+            self.metadata.name, self.metadata.basename, self.metadata.finetune,                  # pyright: ignore[reportAttributeAccessIssue] # ty: ignore[unresolved-attribute]
+            self.metadata.version, size_label=None, output_type=output_type, model_type=None)    # pyright: ignore[reportAttributeAccessIssue] # ty: ignore[unresolved-attribute]
+        self.fname_out = self.fname_out.parent / f"mtp-{fname_default}.gguf"
+
+
 @ModelBase.register("Qwen3NextForCausalLM")
-class Qwen3NextModel(Qwen2MoeModel):
+class Qwen3NextModel(_QwenMtpMixin, Qwen2MoeModel):
     model_arch = gguf.MODEL_ARCH.QWEN3NEXT
 
     def set_gguf_parameters(self):
@@ -284,16 +377,6 @@ class Qwen3NextModel(Qwen2MoeModel):
             rope_dim = self.hparams["hidden_size"] // self.hparams["num_attention_heads"]
         self.gguf_writer.add_rope_dimension_count(int(rope_dim * self.rope_parameters.get("partial_rotary_factor", 0.25)))
 
-    @classmethod
-    def filter_tensors(cls, item: tuple[str, Callable[[], Tensor]]) -> tuple[str, Callable[[], Tensor]] | None:
-        name, gen = item
-
-        if name.startswith("mtp"):
-            # ignore MTP layers for now
-            return None
-
-        return super().filter_tensors(item)
-
     def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
         if name.endswith(".A_log"):
             data_torch = -torch.exp(data_torch)
@@ -536,97 +619,13 @@ class _Qwen35MRopeMixin:
             self.gguf_writer.add_rope_dimension_sections(self._QWEN35_DEFAULT_MROPE_SECTION)
 
 
-class _Qwen35MtpMixin:
-    """Shared MTP wiring for Qwen3.5/3.6 text variants. The HF config carries
-    the MTP block under `mtp_num_hidden_layers` and the tensors under
-    `mtp.*`; we extend block_count, emit the nextn metadata key, and remap
-    `mtp.*` to the standard layer-indexed nextn naming so the existing
-    tensor_map handles them."""
-
-    supports_mtp_export = True
-    hparams: dict[str, Any]
-    model_arch: gguf.MODEL_ARCH
-    gguf_writer: gguf.GGUFWriter
-    block_count: int
-    tensor_map: gguf.TensorNameMap
-    no_mtp: bool
-    mtp_only: bool
-    _original_block_count: int | None = None
-
-    def __init__(self, *args, **kwargs):
-        super().__init__(*args, **kwargs)
-        self.block_count = self.hparams["num_hidden_layers"]
-        if not self.no_mtp:
-            self.block_count += self.hparams.get("mtp_num_hidden_layers", 0)
-        self.tensor_map = gguf.get_tensor_name_map(self.model_arch, self.block_count)
-
-    def index_tensors(self, remote_hf_model_id: str | None = None) -> dict[str, Callable[[], Tensor]]:
-        hparams = {**self.hparams, **self.hparams.get("text_config", {})}
-        key = next((k for k in ["n_layers", "num_hidden_layers", "n_layer", "num_layers"] if k in hparams), None)
-        type(self)._original_block_count = hparams.get(key)
-        return super().index_tensors(remote_hf_model_id=remote_hf_model_id)  # ty: ignore[unresolved-attribute]
-
-    @classmethod
-    def filter_tensors(cls, item):
-        assert cls._original_block_count is not None
-        # TODO: change TextModel to super()
-        if (titem := TextModel.filter_tensors(item)) is None:
-            return None
-        name, gen = titem
-        if name.startswith("model.mtp."):
-            name = name.replace("model.", "", 1)
-        if name.startswith("mtp."):
-            if cls.no_mtp:
-                return None
-            remapper = {
-                "fc":                    "eh_proj",
-                "pre_fc_norm_embedding": "enorm",
-                "pre_fc_norm_hidden":    "hnorm",
-                "norm":                  "shared_head.norm",
-            }
-            parts = name.split(".", 3)
-            if len(parts) == 4 and parts[1] == "layers" and parts[2].isdecimal():
-                mtp_idx = int(parts[2])
-                name = f"model.layers.{cls._original_block_count + mtp_idx}.{parts[3]}"
-            elif len(parts) == 3 and parts[1] in remapper:
-                name = f"model.layers.{cls._original_block_count}.{remapper[parts[1]]}.{parts[2]}"
-        elif cls.mtp_only:
-            keep = name in (
-                "model.embed_tokens.weight", "model.norm.weight", "lm_head.weight",
-                "embed_tokens.weight", "norm.weight",
-            )
-            if not keep:
-                return None
-        return name, gen
-
-    def set_gguf_parameters(self):
-        super().set_gguf_parameters()  # ty: ignore[unresolved-attribute]
-        if self.no_mtp:
-            return
-        if (n := self.hparams.get("mtp_num_hidden_layers", 0)) > 0:
-            self.gguf_writer.add_nextn_predict_layers(n)
-
-    def prepare_metadata(self, vocab_only: bool):
-        from_dir = self.fname_out.is_dir()
-        super().prepare_metadata(vocab_only=vocab_only)  # ty: ignore[unresolved-attribute]
-
-        if not self.mtp_only or not from_dir:
-            return
-
-        output_type: str = self.ftype.name.partition("_")[2]  # pyright: ignore[reportAttributeAccessIssue] # ty: ignore[unresolved-attribute]
-        fname_default: str = gguf.naming_convention(
-            self.metadata.name, self.metadata.basename, self.metadata.finetune,                  # pyright: ignore[reportAttributeAccessIssue] # ty: ignore[unresolved-attribute]
-            self.metadata.version, size_label=None, output_type=output_type, model_type=None)    # pyright: ignore[reportAttributeAccessIssue] # ty: ignore[unresolved-attribute]
-        self.fname_out = self.fname_out.parent / f"mtp-{fname_default}.gguf"
-
-
 @ModelBase.register("Qwen3_5ForConditionalGeneration", "Qwen3_5ForCausalLM")
-class Qwen3_5TextModel(_Qwen35MtpMixin, _Qwen35MRopeMixin, _LinearAttentionVReorderBase):
+class Qwen3_5TextModel(_Qwen35MRopeMixin, _LinearAttentionVReorderBase):
     model_arch = gguf.MODEL_ARCH.QWEN35
 
 
 @ModelBase.register("Qwen3_5MoeForConditionalGeneration", "Qwen3_5MoeForCausalLM")
-class Qwen3_5MoeTextModel(_Qwen35MtpMixin, _Qwen35MRopeMixin, _LinearAttentionVReorderBase):
+class Qwen3_5MoeTextModel(_Qwen35MRopeMixin, _LinearAttentionVReorderBase):
     model_arch = gguf.MODEL_ARCH.QWEN35MOE
 
 

examples/gen-docs/gen-docs.cpp

@@ -70,6 +70,8 @@ static void write_table(std::ostringstream & ss, std::vector<common_arg *> & opt
 
 static void write_help(std::ostringstream & ss, const md_file & md) {
     common_params params;
+    params.is_gen_docs = true;
+
     auto ctx_arg = common_params_parser_init(params, md.ex);
 
     std::vector<common_arg *> common_options;

ggml/src/ggml-backend.cpp

@@ -765,8 +765,9 @@ struct ggml_backend_sched_split {
     int backend_id;
     int i_start;
     int i_end;
-    struct ggml_tensor * inputs[GGML_SCHED_MAX_SPLIT_INPUTS];
+    struct ggml_tensor ** inputs;
     int n_inputs;
+    int inputs_capacity;
     // graph view of this split
     struct ggml_cgraph graph;
 };
@@ -805,8 +806,9 @@ struct ggml_backend_sched {
     int cur_copy;
     int next_copy;
     ggml_backend_event_t events[GGML_SCHED_MAX_BACKENDS][GGML_SCHED_MAX_COPIES];
-    struct ggml_tensor * graph_inputs[GGML_SCHED_MAX_SPLIT_INPUTS];
+    struct ggml_tensor ** graph_inputs;
     int n_graph_inputs;
+    int graph_inputs_capacity;
 
     struct ggml_context * ctx;
 
@@ -832,6 +834,36 @@ struct ggml_backend_sched {
 #define tensor_id_copy(id, backend_id, copy_id) sched->hv_tensor_copies[(id) * sched->n_backends * sched->n_copies + (backend_id) * sched->n_copies + (copy_id)]
 #define tensor_copy(tensor, backend_id, copy_id) tensor_id_copy(hash_id(tensor), backend_id, copy_id)
 
+static void ggml_backend_sched_split_inputs_grow(struct ggml_backend_sched_split * split) {
+    int new_cap = GGML_SCHED_MAX_SPLIT_INPUTS;
+    if (split->inputs_capacity > 0) {
+        new_cap = 2*split->inputs_capacity;
+        GGML_LOG_WARN("%s: increasing split inputs capacity from %d to %d\n", __func__, split->inputs_capacity, new_cap);
+    }
+    auto * pnew = (struct ggml_tensor **) realloc((void *) split->inputs, new_cap * sizeof(struct ggml_tensor *));
+    if (pnew == NULL) {
+        GGML_LOG_ERROR("%s: failed to allocate %zu bytes\n", __func__, new_cap * sizeof(struct ggml_tensor *));
+        GGML_ABORT("failed to grow split inputs container");
+    }
+    split->inputs = pnew;
+    split->inputs_capacity = new_cap;
+}
+
+static void ggml_backend_sched_graph_inputs_grow(ggml_backend_sched_t sched) {
+    int new_cap = GGML_SCHED_MAX_SPLIT_INPUTS;
+    if (sched->graph_inputs_capacity > 0) {
+        new_cap = 2*sched->graph_inputs_capacity;
+        GGML_LOG_WARN("%s: increasing graph inputs capacity from %d to %d\n", __func__, sched->graph_inputs_capacity, new_cap);
+    }
+    auto * pnew = (struct ggml_tensor **) realloc((void *) sched->graph_inputs, new_cap * sizeof(struct ggml_tensor *));
+    if (pnew == NULL) {
+        GGML_LOG_ERROR("%s: failed to allocate %zu bytes\n", __func__, new_cap * sizeof(struct ggml_tensor *));
+        GGML_ABORT("failed to grow graph inputs container");
+    }
+    sched->graph_inputs = pnew;
+    sched->graph_inputs_capacity = new_cap;
+}
+
 // returns the priority of the backend, lower id is higher priority
 static int ggml_backend_sched_backend_id(ggml_backend_sched_t sched, ggml_backend_t backend) {
     for (int i = 0; i < sched->n_backends; i++) {
@@ -1297,7 +1329,7 @@ void ggml_backend_sched_split_graph(ggml_backend_sched_t sched, struct ggml_cgra
                     }
                     // check if the split has too many inputs
                     // FIXME: count the number of inputs instead of only checking when full
-                    if (split->n_inputs == GGML_SCHED_MAX_SPLIT_INPUTS) {
+                    if (split->n_inputs >= split->inputs_capacity) {
                         const size_t id = hash_id(src);
                         int src_backend_id = sched->hv_tensor_backend_ids[id];
                         bool supported = ggml_backend_sched_buffer_supported(sched, src, cur_backend_id);
@@ -1313,10 +1345,14 @@ void ggml_backend_sched_split_graph(ggml_backend_sched_t sched, struct ggml_cgra
                 split->i_end = i;
                 i_split++;
                 if (i_split >= sched->splits_capacity) {
+                    int old_cap = sched->splits_capacity;
                     sched->splits_capacity *= 2;
                     sched->splits = (ggml_backend_sched_split *)
                         realloc(sched->splits, sched->splits_capacity * sizeof(struct ggml_backend_sched_split));
                     GGML_ASSERT(sched->splits != NULL);
+                    for (int k = old_cap; k < sched->splits_capacity; k++) {
+                        memset(&sched->splits[k], 0, sizeof(struct ggml_backend_sched_split));
+                    }
                 }
                 split = &sched->splits[i_split];
                 split->backend_id = node_backend_id;
@@ -1353,7 +1389,9 @@ void ggml_backend_sched_split_graph(ggml_backend_sched_t sched, struct ggml_cgra
                             SET_CAUSE(tensor_copy, "4.cpy");
                         }
                         int n_graph_inputs = sched->n_graph_inputs++;
-                        GGML_ASSERT(n_graph_inputs < GGML_SCHED_MAX_SPLIT_INPUTS);
+                        if (n_graph_inputs >= sched->graph_inputs_capacity) {
+                            ggml_backend_sched_graph_inputs_grow(sched);
+                        }
                         sched->graph_inputs[n_graph_inputs] = src;
                     }
                 }
@@ -1373,7 +1411,9 @@ void ggml_backend_sched_split_graph(ggml_backend_sched_t sched, struct ggml_cgra
                             SET_CAUSE(tensor_copy, "4.cpy");
                         }
                         int n_inputs = split->n_inputs++;
-                        GGML_ASSERT(n_inputs < GGML_SCHED_MAX_SPLIT_INPUTS);
+                        if (n_inputs >= split->inputs_capacity) {
+                            ggml_backend_sched_split_inputs_grow(split);
+                        }
                         split->inputs[n_inputs] = src;
                     }
                     node->src[j] = tensor_id_copy(src_id, cur_backend_id, sched->cur_copy);
@@ -1399,7 +1439,11 @@ void ggml_backend_sched_split_graph(ggml_backend_sched_t sched, struct ggml_cgra
         sched->prev_leaf_backend_ids = tmp;
     }
 
-    int graph_size = std::max(graph->n_nodes, graph->n_leafs) + sched->n_splits*GGML_SCHED_MAX_SPLIT_INPUTS*2*sched->n_copies;
+    int total_inputs = sched->n_graph_inputs;
+    for (int i = 0; i < sched->n_splits; i++) {
+        total_inputs += sched->splits[i].n_inputs;
+    }
+    int graph_size = std::max(graph->n_nodes, graph->n_leafs) + total_inputs * 2 * sched->n_copies;
 
     // remember the actual graph_size for performing reallocation checks later [GGML_SCHED_DEBUG_REALLOC]
     sched->debug_prev_graph_size = sched->debug_graph_size;
@@ -1782,6 +1826,9 @@ ggml_backend_sched_t ggml_backend_sched_new(
     sched->splits = (ggml_backend_sched_split *) calloc(initial_splits_capacity, sizeof(sched->splits[0]));
     sched->splits_capacity = initial_splits_capacity;
 
+    sched->graph_inputs_capacity = GGML_SCHED_MAX_SPLIT_INPUTS;
+    sched->graph_inputs = (struct ggml_tensor **) calloc(sched->graph_inputs_capacity, sizeof(struct ggml_tensor *));
+
     for (int b = 0; b < n_backends; b++) {
         sched->backends[b] = backends[b];
         sched->bufts[b] = bufts ? bufts[b] : ggml_backend_get_default_buffer_type(backends[b]);
@@ -1814,7 +1861,11 @@ void ggml_backend_sched_free(ggml_backend_sched_t sched) {
     ggml_gallocr_free(sched->galloc);
     ggml_free(sched->ctx);
     ggml_hash_set_free(&sched->hash_set);
+    for (int i = 0; i < sched->splits_capacity; i++) {
+        free(sched->splits[i].inputs);
+    }
     free(sched->splits);
+    free(sched->graph_inputs);
     free(sched->hv_tensor_backend_ids);
     free(sched->hv_tensor_copies);
     free(sched->node_backend_ids);

ggml/src/ggml-cuda/common.cuh

@@ -627,7 +627,8 @@ template <typename T> struct block_reduce_policy<block_reduce_method::MAX, T> {
 };
 
 template <block_reduce_method reduce_method_t, const unsigned int block_size_template = 0, typename T>
-static __device__ T block_reduce(T val, T * shared_vals) {
+static __device__ T block_reduce(T val, [[maybe_unused]] T * shared_vals) {
+    // for multi-warp reductions, callers must not reuse shared_vals until all reads from this invocation have completed
     val                           = block_reduce_policy<reduce_method_t, T>::reduce(val);
     const unsigned int block_size = block_size_template == 0 ? blockDim.x : block_size_template;
     if (block_size > WARP_SIZE) {

ggml/src/ggml-cuda/norm.cu

@@ -64,7 +64,7 @@ static __global__ void group_norm_f32(const float * x, float * dst, const int gr
         tmp += xi * xi;
     }
 
-    tmp = block_reduce<block_reduce_method::SUM, block_size>(tmp, s_sum);
+    tmp = block_reduce<block_reduce_method::SUM, block_size>(tmp, s_sum + 32);
 
     const float variance = tmp / group_size;
     const float scale = rsqrtf(variance + eps);
@@ -297,7 +297,7 @@ static void group_norm_f32_cuda(
         group_norm_f32<WARP_SIZE><<<num_groups, block_dims, 0, stream>>>(x, dst, group_size, ne_elements, eps);
     } else {
         const dim3 block_dims(1024, 1, 1);
-        group_norm_f32<1024><<<num_groups, block_dims, block_dims.x > WARP_SIZE ? 32 * sizeof(float): 0, stream>>>(x, dst, group_size, ne_elements, eps);
+        group_norm_f32<1024><<<num_groups, block_dims, block_dims.x > WARP_SIZE ? 2 * 32 * sizeof(float): 0, stream>>>(x, dst, group_size, ne_elements, eps);
     }
 }
 

ggml/src/ggml-cuda/softmax.cu

@@ -116,6 +116,11 @@ static __global__ void soft_max_f32(
         vals[col] = val;
     }
 
+    if (block_size > WARP_SIZE) {
+        // sync is needed as we reuse buf_iw across block_reduce invocations, see #26385
+        // for block_size <= WARP_SIZE, block_reduce does not access buf_iw
+        __syncthreads();
+    }
     // find the sum of exps in the block
     tmp = block_reduce<block_reduce_method::SUM, block_size_template>(tmp, buf_iw);
 
@@ -142,6 +147,8 @@ static __device__ void soft_max_f32_parallelize_cols_single_row(const float * __
                                                                 float * __restrict__ dst,
                                                                 float * __restrict__ tmp_maxs,
                                                                 float * __restrict__ tmp_sums,
+                                                                float * shared_vals_max,
+                                                                float * shared_vals_sum,
                                                                 const soft_max_params p) {
     namespace cg = cooperative_groups;
 
@@ -154,7 +161,6 @@ static __device__ void soft_max_f32_parallelize_cols_single_row(const float * __
     float     local_vals[n_elem_per_thread] = { -INFINITY, -INFINITY, -INFINITY, -INFINITY };
     float     local_max                     = -INFINITY;
     const int step_size                     = gridDim.x * blockDim.x;
-    __shared__ float shared_vals[32];
 
     // Compute thread-local max
     for (int col = col_start; col < p.ncols;) {
@@ -171,7 +177,7 @@ static __device__ void soft_max_f32_parallelize_cols_single_row(const float * __
     }
 
     // Compute CTA-level max
-    local_max = block_reduce<block_reduce_method::MAX>(local_max, shared_vals);
+    local_max = block_reduce<block_reduce_method::MAX>(local_max, shared_vals_max);
 
     // Store CTA-level max to GMEM
     if (tid == 0) {
@@ -186,7 +192,7 @@ static __device__ void soft_max_f32_parallelize_cols_single_row(const float * __
     } else {
         local_max = -INFINITY;
     }
-    local_max = block_reduce<block_reduce_method::MAX>(local_max, shared_vals);
+    local_max = block_reduce<block_reduce_method::MAX>(local_max, shared_vals_max);
 
     // Compute softmax dividends, accumulate divisor
     float tmp_expf = 0.0f;
@@ -209,7 +215,7 @@ static __device__ void soft_max_f32_parallelize_cols_single_row(const float * __
     }
 
     // Reduce divisor within CTA
-    tmp_expf = block_reduce<block_reduce_method::SUM>(tmp_expf, shared_vals);
+    tmp_expf = block_reduce<block_reduce_method::SUM>(tmp_expf, shared_vals_sum);
 
     // Store CTA-level sum to GMEM
     if (tid == 0) {
@@ -223,7 +229,7 @@ static __device__ void soft_max_f32_parallelize_cols_single_row(const float * __
     } else {
         tmp_expf = 0.0f;
     }
-    tmp_expf = block_reduce<block_reduce_method::SUM>(tmp_expf, shared_vals);
+    tmp_expf = block_reduce<block_reduce_method::SUM>(tmp_expf, shared_vals_sum);
 
     // Divide dividend by global sum + store data
     for (int col = col_start; col < p.ncols;) {
@@ -310,9 +316,11 @@ __launch_bounds__(8*WARP_SIZE, 1) static __global__ void soft_max_f32_paralleliz
 // https://docs.nvidia.com/cuda/cuda-programming-guide/05-appendices/device-callable-apis.html#grid-synchronization
 // https://docs.nvidia.com/cuda/cuda-programming-guide/05-appendices/device-callable-apis.html#class-cluster-group
 {
+    __shared__ float shared_vals[2][32];
+
     for (int rowx = 0; rowx < p.ne01 * p.ne02 * p.ne03; rowx++) {
         soft_max_f32_parallelize_cols_single_row(x + int64_t(rowx) * p.ncols, dst + int64_t(rowx) * p.ncols, tmp_maxs,
-                                                 tmp_sums, p);
+                                                 tmp_sums, shared_vals[0], shared_vals[1], p);
     }
 }
 

ggml/src/ggml-metal/ggml-metal-device.cpp

@@ -477,6 +477,23 @@ ggml_metal_pipeline_with_params ggml_metal_library_get_pipeline_soft_max(ggml_me
     return res;
 }
 
+ggml_metal_pipeline_with_params ggml_metal_library_get_pipeline_lightning_indexer(
+        ggml_metal_library_t lib,
+        const ggml_tensor * op) {
+    GGML_ASSERT(op->op == GGML_OP_LIGHTNING_INDEXER);
+
+    char name[256];
+
+    snprintf(name, 256, "kernel_lightning_indexer_%s", ggml_type_name(op->src[1]->type));
+
+    ggml_metal_pipeline_with_params res = ggml_metal_library_get_pipeline(lib, name);
+    if (!res.pipeline) {
+        res = ggml_metal_library_compile_pipeline(lib, name, name, nullptr);
+    }
+
+    return res;
+}
+
 ggml_metal_pipeline_with_params ggml_metal_library_get_pipeline_dsv4_hc(ggml_metal_library_t lib, ggml_op op) {
     const char * name = nullptr;
 

ggml/src/ggml-metal/ggml-metal-device.h

@@ -125,6 +125,7 @@ struct ggml_metal_pipeline_with_params ggml_metal_library_get_pipeline_cumsum_bl
 struct ggml_metal_pipeline_with_params ggml_metal_library_get_pipeline_cumsum_add        (ggml_metal_library_t lib, const struct ggml_tensor * op);
 struct ggml_metal_pipeline_with_params ggml_metal_library_get_pipeline_tri               (ggml_metal_library_t lib, const struct ggml_tensor * op);
 struct ggml_metal_pipeline_with_params ggml_metal_library_get_pipeline_soft_max          (ggml_metal_library_t lib, const struct ggml_tensor * op);
+struct ggml_metal_pipeline_with_params ggml_metal_library_get_pipeline_lightning_indexer (ggml_metal_library_t lib, const struct ggml_tensor * op);
 struct ggml_metal_pipeline_with_params ggml_metal_library_get_pipeline_dsv4_hc           (ggml_metal_library_t lib, enum ggml_op op);
 struct ggml_metal_pipeline_with_params ggml_metal_library_get_pipeline_ssm_conv          (ggml_metal_library_t lib, const struct ggml_tensor * op);
 struct ggml_metal_pipeline_with_params ggml_metal_library_get_pipeline_ssm_conv_batched  (ggml_metal_library_t lib, const struct ggml_tensor * op, int ssm_conv_bs);

ggml/src/ggml-metal/ggml-metal-device.m

@@ -2,6 +2,7 @@
 
 #import "ggml-impl.h"
 #import "ggml-backend-impl.h"
+#import "ggml-metal-impl.h"
 
 #include <Foundation/Foundation.h>
 
@@ -1308,6 +1309,36 @@ bool ggml_metal_device_supports_op(ggml_metal_device_t dev, const struct ggml_te
                     return false;
             }
             return has_simdgroup_mm; // TODO: over-restricted for vec-kernels
+        case GGML_OP_LIGHTNING_INDEXER:
+            if (op->src[0]->ne[0] != OP_LIGHTNING_INDEXER_DK ||
+                op->src[0]->ne[1] != OP_LIGHTNING_INDEXER_NH) {
+                return false;
+            }
+            if (!has_simdgroup_mm ||
+                op->src[0]->type != GGML_TYPE_F32 ||
+                op->src[2]->type != GGML_TYPE_F32 ||
+                op->src[3]->type != GGML_TYPE_F16 ||
+                op->type         != GGML_TYPE_F32 ||
+                !ggml_is_contiguous_rows(op->src[0]) ||
+                !ggml_is_contiguous_rows(op->src[1]) ||
+                !ggml_is_contiguous_rows(op->src[2]) ||
+                !ggml_is_contiguous_rows(op->src[3])) {
+                return false;
+            }
+            switch (op->src[1]->type) {
+                case GGML_TYPE_F32:
+                case GGML_TYPE_F16:
+                case GGML_TYPE_Q4_0:
+                case GGML_TYPE_Q4_1:
+                case GGML_TYPE_Q5_0:
+                case GGML_TYPE_Q5_1:
+                case GGML_TYPE_Q8_0:
+                    return true;
+                case GGML_TYPE_BF16:
+                    return has_bfloat;
+                default:
+                    return false;
+            }
         case GGML_OP_DSV4_HC_COMB:
             return has_simdgroup_reduction &&
                 op->src[0]->type == GGML_TYPE_F32 &&

ggml/src/ggml-metal/ggml-metal-impl.h

@@ -112,6 +112,13 @@
 #define OP_FLASH_ATTN_EXT_VEC_NQPSG 1
 #define OP_FLASH_ATTN_EXT_VEC_NCPSG 32
 
+#define OP_LIGHTNING_INDEXER_DK    128
+#define OP_LIGHTNING_INDEXER_NH     64
+#define OP_LIGHTNING_INDEXER_NHPTG   8
+#define OP_LIGHTNING_INDEXER_NKPSG   8
+#define OP_LIGHTNING_INDEXER_NSG     8
+#define OP_LIGHTNING_INDEXER_NBPTG   8
+
 #define OP_UNARY_NUM_SCALE      10
 #define OP_UNARY_NUM_FILL       11
 #define OP_UNARY_NUM_CLAMP      12
@@ -1171,6 +1178,23 @@ typedef struct {
     int64_t val;
 } ggml_metal_kargs_memset;
 
+typedef struct {
+    int32_t  n_kv;
+    int32_t  n_batch;
+    int32_t  mask_ne3;
+    uint64_t nb1;
+    uint64_t nb3;
+    uint64_t nbq1;
+    uint64_t nbq2;
+    uint64_t nbq3;
+    uint64_t nbk2;
+    uint64_t nbk3;
+    uint64_t nbw1;
+    uint64_t nbw3;
+    uint64_t nbm1;
+    uint64_t nbm3;
+} ggml_metal_kargs_lightning_indexer;
+
 typedef struct {
     int32_t  n_tokens;
     int32_t  n_iter;

ggml/src/ggml-metal/ggml-metal-ops.cpp

@@ -320,6 +320,10 @@ static int ggml_metal_op_encode_impl(ggml_metal_op_t ctx, int idx) {
             {
                 n_fuse = ggml_metal_op_cumsum(ctx, idx);
             } break;
+        case GGML_OP_LIGHTNING_INDEXER:
+            {
+                n_fuse = ggml_metal_op_lightning_indexer(ctx, idx);
+            } break;
         case GGML_OP_DSV4_HC_COMB:
         case GGML_OP_DSV4_HC_PRE:
         case GGML_OP_DSV4_HC_POST:
@@ -1307,6 +1311,72 @@ int ggml_metal_op_diag(ggml_metal_op_t ctx, int idx) {
     return 1;
 }
 
+int ggml_metal_op_lightning_indexer(ggml_metal_op_t ctx, int idx) {
+    ggml_tensor * op = ctx->node(idx);
+
+    ggml_metal_encoder_t enc = ctx->enc;
+
+    GGML_ASSERT(op->op == GGML_OP_LIGHTNING_INDEXER);
+
+    const ggml_tensor * q = op->src[0];
+    const ggml_tensor * k = op->src[1];
+    const ggml_tensor * w = op->src[2];
+    const ggml_tensor * m = op->src[3];
+
+    GGML_ASSERT(q->type == GGML_TYPE_F32);
+    GGML_ASSERT(k->type == GGML_TYPE_F32  ||
+                k->type == GGML_TYPE_F16  ||
+                k->type == GGML_TYPE_BF16 ||
+                k->type == GGML_TYPE_Q4_0 ||
+                k->type == GGML_TYPE_Q4_1 ||
+                k->type == GGML_TYPE_Q5_0 ||
+                k->type == GGML_TYPE_Q5_1 ||
+                k->type == GGML_TYPE_Q8_0);
+    GGML_ASSERT(w->type == GGML_TYPE_F32);
+    GGML_ASSERT(m->type == GGML_TYPE_F16);
+    GGML_ASSERT(op->type == GGML_TYPE_F32);
+
+    GGML_ASSERT(q->ne[0] == OP_LIGHTNING_INDEXER_DK);
+    GGML_ASSERT(q->ne[1] == OP_LIGHTNING_INDEXER_NH);
+
+    ggml_metal_kargs_lightning_indexer args = {
+        /*.n_kv      =*/ (int32_t) k->ne[2],
+        /*.n_batch   =*/ (int32_t) q->ne[2],
+        /*.mask_ne3  =*/ (int32_t) m->ne[3],
+        /*.nb1       =*/ op->nb[1],
+        /*.nb3       =*/ op->nb[3],
+        /*.nbq1      =*/ q->nb[1],
+        /*.nbq2      =*/ q->nb[2],
+        /*.nbq3      =*/ q->nb[3],
+        /*.nbk2      =*/ k->nb[2],
+        /*.nbk3      =*/ k->nb[3],
+        /*.nbw1      =*/ w->nb[1],
+        /*.nbw3      =*/ w->nb[3],
+        /*.nbm1      =*/ m->nb[1],
+        /*.nbm3      =*/ m->nb[3],
+    };
+
+    ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(q),  1);
+    ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(k),  2);
+    ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(w),  3);
+    ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(m),  4);
+    ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(op), 5);
+
+    const int nsg   = OP_LIGHTNING_INDEXER_NSG;
+    const int nkptg = OP_LIGHTNING_INDEXER_NKPSG*nsg;
+    const int nbptg = OP_LIGHTNING_INDEXER_NBPTG;
+
+    auto pipeline = ggml_metal_library_get_pipeline_lightning_indexer(ctx->lib, op);
+    ggml_metal_encoder_set_pipeline(enc, pipeline);
+    ggml_metal_encoder_set_bytes(enc, &args, sizeof(args), 0);
+    ggml_metal_encoder_dispatch_threadgroups(enc,
+            (k->ne[2] + nkptg - 1)/nkptg,
+            (q->ne[2] + nbptg - 1)/nbptg,
+            q->ne[3], 32, nsg, 1);
+
+    return 1;
+}
+
 int ggml_metal_op_dsv4_hc(ggml_metal_op_t ctx, int idx) {
     ggml_tensor * op = ctx->node(idx);
 

ggml/src/ggml-metal/ggml-metal-ops.h

@@ -54,6 +54,7 @@ int ggml_metal_op_cumsum            (ggml_metal_op_t ctx, int idx);
 int ggml_metal_op_get_rows          (ggml_metal_op_t ctx, int idx);
 int ggml_metal_op_set_rows          (ggml_metal_op_t ctx, int idx);
 int ggml_metal_op_diag              (ggml_metal_op_t ctx, int idx);
+int ggml_metal_op_lightning_indexer (ggml_metal_op_t ctx, int idx);
 int ggml_metal_op_dsv4_hc           (ggml_metal_op_t ctx, int idx);
 int ggml_metal_op_soft_max          (ggml_metal_op_t ctx, int idx);
 int ggml_metal_op_ssm_conv          (ggml_metal_op_t ctx, int idx);

ggml/src/ggml-metal/ggml-metal.metal

@@ -11295,6 +11295,155 @@ typedef decltype(kernel_count_equal<int32_t>) kernel_count_equal_t;
 
 template [[host_name("kernel_count_equal_i32")]] kernel kernel_count_equal_t kernel_count_equal<int32_t>;
 
+template<
+    typename kd4x4_t,
+    short nl_k,
+    void (*deq_k)(device const kd4x4_t *, short, thread half4x4 &)>
+kernel void kernel_lightning_indexer(
+        constant ggml_metal_kargs_lightning_indexer & args,
+        device const char * q,
+        device const char * k,
+        device const char * w,
+        device const char * m,
+        device       char * dst,
+        uint3  tgpig[[threadgroup_position_in_grid]],
+        ushort tiitg[[thread_index_in_threadgroup]],
+        ushort tiisg[[thread_index_in_simdgroup]],
+        ushort sgitg[[simdgroup_index_in_threadgroup]]) {
+    constexpr short DK    = OP_LIGHTNING_INDEXER_DK;
+    constexpr short NH    = OP_LIGHTNING_INDEXER_NH;
+    constexpr short NHPTG = OP_LIGHTNING_INDEXER_NHPTG;
+    constexpr short NKPSG = OP_LIGHTNING_INDEXER_NKPSG;
+    constexpr short NSG   = OP_LIGHTNING_INDEXER_NSG;
+    constexpr short NBPTG = OP_LIGHTNING_INDEXER_NBPTG;
+
+    constexpr short DK4  = DK/4;
+    constexpr short DK8  = DK/8;
+    constexpr short DK16 = DK/16;
+
+    constexpr short NK  = NKPSG*NSG; // keys    per threadgroup
+    constexpr short NTG = 32*NSG;    // threads per threadgroup
+
+    const int i_stream = tgpig.z;
+    const int i_kv_0   = tgpig.x*NK;            // first key of this threadgroup
+    const int i_kv     = i_kv_0 + sgitg*NKPSG;  // first key of this simdgroup
+
+    threadgroup half4x4   sk4x4[NK*DK16];
+    threadgroup half    * sk = (threadgroup half *) sk4x4;
+
+    for (short i = tiitg; i < NK*DK16; i += NTG) {
+        const short ik  = i/DK16;
+        const short i16 = i%DK16;
+
+        half4x4 tmp;
+
+        if (i_kv_0 + ik < args.n_kv) {
+            device const kd4x4_t * kr = (device const kd4x4_t *) (k + (i_kv_0 + ik)*args.nbk2 + i_stream*args.nbk3);
+
+            deq_k(kr + i16/nl_k, i16%nl_k, tmp);
+        } else {
+            FOR_UNROLL (short j = 0; j < 4; ++j) {
+                tmp[j] = half4(0.0h);
+            }
+        }
+
+        sk4x4[i] = tmp;
+    }
+
+    threadgroup_barrier(mem_flags::mem_threadgroup);
+
+    // K tile of this simdgroup, transposed to [DK, NKPSG]
+    simdgroup_half8x8 mk[DK8];
+
+    FOR_UNROLL (short i = 0; i < DK8; ++i) {
+        simdgroup_load(mk[i], sk + sgitg*NKPSG*DK + 8*i, DK, 0, true);
+    }
+
+    threadgroup half4   sq4[NHPTG*DK4];
+    threadgroup half  * sq = (threadgroup half *) sq4;
+
+    threadgroup float sw [NHPTG];
+    threadgroup float sqk[NSG*NHPTG*NKPSG];
+
+    const int i_batch_0 = tgpig.y*NBPTG;
+    const int n_batch   = min((int) NBPTG, args.n_batch - i_batch_0);
+
+    for (short ib = 0; ib < n_batch; ++ib) {
+        const int i_batch = i_batch_0 + ib;
+
+        device const char * pq = q + i_batch*args.nbq2 + i_stream*args.nbq3;
+        device const char * pw = w + i_batch*args.nbw1 + i_stream*args.nbw3;
+
+        float score = 0.0f;
+
+        FOR_UNROLL (short i_head = 0; i_head < NH; i_head += NHPTG) {
+            // stage the Q tile [DK, NHPTG] and the (prescaled) head weights
+            for (short i = tiitg; i < NHPTG*DK4; i += NTG) {
+                const short ih = i/DK4;
+                const short i4 = i%DK4;
+
+                device const float4 * q4 = (device const float4 *) (pq + (i_head + ih)*args.nbq1);
+
+                sq4[ih*DK4 + i4] = half4(q4[i4]);
+            }
+
+            if (tiitg < NHPTG) {
+                sw[tiitg] = ((device const float *) pw)[i_head + tiitg];
+            }
+
+            threadgroup_barrier(mem_flags::mem_threadgroup);
+
+            simdgroup_float8x8 mqk = make_filled_simdgroup_matrix<float, 8>(0.0f);
+
+            FOR_UNROLL (short i = 0; i < DK8; ++i) {
+                simdgroup_half8x8 mq;
+
+                simdgroup_load(mq, sq + 8*i, DK, 0, false);
+                simdgroup_multiply_accumulate(mqk, mq, mk[i], mqk);
+            }
+
+            threadgroup float * pqk = sqk + sgitg*NHPTG*NKPSG;
+
+            simdgroup_store(mqk, pqk, NKPSG, 0, false);
+            simdgroup_barrier(mem_flags::mem_threadgroup);
+
+            // one lane per key: ReLU, apply the head weight and accumulate over the head tile
+            if (tiisg < NKPSG) {
+                FOR_UNROLL (short ih = 0; ih < NHPTG; ++ih) {
+                    score += max(pqk[ih*NKPSG + tiisg], 0.0f)*sw[ih];
+                }
+            }
+
+            threadgroup_barrier(mem_flags::mem_threadgroup);
+        }
+
+        if (tiisg < NKPSG) {
+            const int ik = i_kv + tiisg;
+            if (ik < args.n_kv) {
+                device const half  * pm = (device const half  *) (m   + i_batch*args.nbm1 + (i_stream % args.mask_ne3)*args.nbm3);
+                device       float * pd = (device       float *) (dst + i_batch*args.nb1  + i_stream*args.nb3);
+
+                pd[ik] = score + (float) pm[ik];
+            }
+        }
+    }
+}
+
+typedef decltype(kernel_lightning_indexer<half4x4, 1, dequantize_f16>) kernel_lightning_indexer_t;
+
+template [[host_name("kernel_lightning_indexer_f32")]]  kernel kernel_lightning_indexer_t kernel_lightning_indexer<float4x4, 1, dequantize_f32>;
+template [[host_name("kernel_lightning_indexer_f16")]]  kernel kernel_lightning_indexer_t kernel_lightning_indexer<half4x4,  1, dequantize_f16>;
+
+#if defined(GGML_METAL_HAS_BF16)
+template [[host_name("kernel_lightning_indexer_bf16")]] kernel kernel_lightning_indexer_t kernel_lightning_indexer<bfloat4x4, 1, dequantize_bf16>;
+#endif
+
+template [[host_name("kernel_lightning_indexer_q4_0")]] kernel kernel_lightning_indexer_t kernel_lightning_indexer<block_q4_0, 2, dequantize_q4_0>;
+template [[host_name("kernel_lightning_indexer_q4_1")]] kernel kernel_lightning_indexer_t kernel_lightning_indexer<block_q4_1, 2, dequantize_q4_1>;
+template [[host_name("kernel_lightning_indexer_q5_0")]] kernel kernel_lightning_indexer_t kernel_lightning_indexer<block_q5_0, 2, dequantize_q5_0>;
+template [[host_name("kernel_lightning_indexer_q5_1")]] kernel kernel_lightning_indexer_t kernel_lightning_indexer<block_q5_1, 2, dequantize_q5_1>;
+template [[host_name("kernel_lightning_indexer_q8_0")]] kernel kernel_lightning_indexer_t kernel_lightning_indexer<block_q8_0, 2, dequantize_q8_0>;
+
 kernel void kernel_dsv4_hc_comb_f32(
         constant ggml_metal_kargs_dsv4_hc_comb & args,
         device const char * mixes,
@@ -11452,4 +11601,3 @@ kernel void kernel_dsv4_hc_post_f32(
         *(device float *) (dst + i0*args.nb_d0 + idst*args.nb_d1 + it*args.nb_d2) = result[idst];
     }
 }
-

ggml/src/ggml-opencl/ggml-opencl.cpp

@@ -7065,7 +7065,7 @@ static inline bool use_flat_gemv_for_large_m_q4_K(const ggml_tensor *tensor) {
     return tensor->ne[1] >= 32768 && tensor->ne[2] == 1 && tensor->ne[3] == 1;
 }
 
-static inline bool use_flat_gemv_for_large_m_q6_K(const ggml_tensor *tensor) {
+static inline bool use_flat_gemv_for_large_m_q6_K(const ggml_backend_opencl_context *backend_ctx, const ggml_tensor *tensor) {
     // gemv_noshuffle variant perf drops for large M, use flat variant for large M.
     // threshold is well above typical hidden/FFN dims, but below typical vocab sizes.
     // q6_K flat gemv is worse for smaller K; 2048 seems to be a reasonable threshold.
@@ -7083,7 +7083,15 @@ static inline bool use_flat_gemv_for_large_m_q6_K(const ggml_tensor *tensor) {
     if ((tensor->ne[1] % 128 != 0) && tensor->ne[2] == 1 && tensor->ne[3] == 1) {
         return true;
     }
-    return tensor->ne[1] >= 32768 && tensor->ne[0] >= 2048 && tensor->ne[2] == 1 && tensor->ne[3] == 1;
+
+    // The gemv_noshuffle slowdown tracks TOTAL weight size, not ne0 alone; ne0 >= 2048 is a
+    // proxy for "large weight" that misses a narrow-hidden vocab-scale lm_head.
+    // Add a direct size escape so such weights also take the flat path, without changing
+    // which weights ne0 >= 2048 already routes there.
+    // The size escape is not taken on the A7X since its compiler miscompiles the flat K-quant GEMV
+    return tensor->ne[1] >= 32768
+        && (tensor->ne[0] >= 2048 || (backend_ctx->adreno_gen != ADRENO_GPU_GEN::A7X && ggml_nbytes(tensor) >= (256ull << 20)))
+        && tensor->ne[2] == 1 && tensor->ne[3] == 1;
 }
 
 static bool ggml_opencl_supports_op(ggml_backend_dev_t dev, const struct ggml_tensor * op) {
@@ -9403,7 +9411,7 @@ static void ggml_backend_opencl_buffer_set_tensor(ggml_backend_buffer_t buffer,
         cl_kernel kernel;
 #ifdef GGML_OPENCL_USE_ADRENO_KERNELS
         kernel = backend_ctx->kernel_convert_block_q6_K;
-        if (use_adreno_kernels(backend_ctx, tensor) && !use_flat_gemv_for_large_m_q6_K(tensor)) {
+        if (use_adreno_kernels(backend_ctx, tensor) && !use_flat_gemv_for_large_m_q6_K(backend_ctx, tensor)) {
             kernel = backend_ctx->kernel_convert_block_q6_K_noshuffle;
         }
 #else
@@ -9436,7 +9444,7 @@ static void ggml_backend_opencl_buffer_set_tensor(ggml_backend_buffer_t buffer,
         tensor->extra  = extra;
 
 #ifdef GGML_OPENCL_USE_ADRENO_KERNELS
-        if (use_adreno_kernels(backend_ctx, tensor) && !use_flat_gemv_for_large_m_q6_K(tensor)) {
+        if (use_adreno_kernels(backend_ctx, tensor) && !use_flat_gemv_for_large_m_q6_K(backend_ctx, tensor)) {
             cl_int M = tensor->ne[1];   // ne01
             cl_int K = tensor->ne[0];   // ne00
 
@@ -10473,7 +10481,7 @@ static void ggml_backend_opencl_buffer_get_tensor(ggml_backend_buffer_t buffer,
             CL_CHECK(clReleaseMemObject(data_device));
             return;
         }
-        if (use_adreno_kernels(backend_ctx, tensor) && !use_flat_gemv_for_large_m_q6_K(tensor)) {
+        if (use_adreno_kernels(backend_ctx, tensor) && !use_flat_gemv_for_large_m_q6_K(backend_ctx, tensor)) {
             static ggml_cl_buffer buf_trans_ql;
             static ggml_cl_buffer buf_trans_qh;
             static ggml_cl_buffer buf_trans_s;
@@ -18895,7 +18903,7 @@ static void ggml_cl_mul_mat(ggml_backend_t backend, const ggml_tensor * src0, co
         }
 
         // q6_K x fp32
-        if (src0t == GGML_TYPE_Q6_K && src1t == GGML_TYPE_F32 && !use_flat_gemv_for_large_m_q6_K(src0)) {
+        if (src0t == GGML_TYPE_Q6_K && src1t == GGML_TYPE_F32 && !use_flat_gemv_for_large_m_q6_K(backend_ctx, src0)) {
             ggml_cl_mul_mat_q6_K_f32_adreno(backend, src0, src1, dst);
             return;
         }

gguf-py/gguf/constants.py

@@ -2518,7 +2518,13 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = {
         MODEL_TENSOR.SSM_NORM,
         MODEL_TENSOR.SSM_IN,
         MODEL_TENSOR.SSM_BETA_ALPHA,
-        MODEL_TENSOR.SSM_OUT
+        MODEL_TENSOR.SSM_OUT,
+        MODEL_TENSOR.NEXTN_EH_PROJ,
+        MODEL_TENSOR.NEXTN_EMBED_TOKENS,
+        MODEL_TENSOR.NEXTN_ENORM,
+        MODEL_TENSOR.NEXTN_HNORM,
+        MODEL_TENSOR.NEXTN_SHARED_HEAD_HEAD,
+        MODEL_TENSOR.NEXTN_SHARED_HEAD_NORM,
     ],
     MODEL_ARCH.QWEN3VL: [
         MODEL_TENSOR.TOKEN_EMBD,
@@ -3404,6 +3410,13 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = {
         MODEL_TENSOR.FFN_DOWN_SHEXP,
         MODEL_TENSOR.FFN_UP_SHEXP,
         MODEL_TENSOR.FFN_EXP_PROBS_B,
+        # NextN/MTP tensors
+        MODEL_TENSOR.NEXTN_EH_PROJ,
+        MODEL_TENSOR.NEXTN_EMBED_TOKENS,
+        MODEL_TENSOR.NEXTN_ENORM,
+        MODEL_TENSOR.NEXTN_HNORM,
+        MODEL_TENSOR.NEXTN_SHARED_HEAD_HEAD,
+        MODEL_TENSOR.NEXTN_SHARED_HEAD_NORM,
     ],
     MODEL_ARCH.DEEPSEEK2OCR: [
         MODEL_TENSOR.TOKEN_EMBD,

include/llama.h

@@ -1256,6 +1256,7 @@ extern "C" {
         struct ggml_tensor * probs;
         struct ggml_tensor * sampled;
         struct ggml_tensor * candidates;
+        int64_t              n_vocab;
     };
 
     // user code can implement the interface below in order to create custom llama_sampler
@@ -1425,9 +1426,9 @@ extern "C" {
     /// NOTE: Avoid using on the full vocabulary as searching for repeated tokens can become slow. For example, apply top-k or top-p sampling first.
     LLAMA_API struct llama_sampler * llama_sampler_init_penalties(
                              int32_t   penalty_last_n,   // last n tokens to penalize (0 = disable penalty, -1 = context size)
-                               float   penalty_repeat,   // 1.0 = disabled
-                               float   penalty_freq,     // 0.0 = disabled
-                               float   penalty_present); // 0.0 = disabled
+                               float   penalty_repeat,   // must be > 0.0, 1.0 = disabled
+                               float   penalty_freq,     // must be finite, 0.0 = disabled
+                               float   penalty_present); // must be finite, 0.0 = disabled
 
     ///  @details DRY sampler, designed by p-e-w, as described in: https://github.com/oobabooga/text-generation-webui/pull/5677, porting Koboldcpp implementation authored by pi6am: https://github.com/LostRuins/koboldcpp/pull/982
     LLAMA_API struct llama_sampler * llama_sampler_init_dry(

src/CMakeLists.txt

@@ -25,6 +25,7 @@ add_library(llama
             llama-kv-cache.cpp
             llama-kv-cache-iswa.cpp
             llama-kv-cache-dsa.cpp
+            llama-kv-cache-msa.cpp
             llama-kv-cache-dsv4.cpp
             llama-memory.cpp
             llama-memory-hybrid.cpp

src/llama-graph.cpp

@@ -8,6 +8,7 @@
 #include "llama-kv-cache.h"
 #include "llama-kv-cache-iswa.h"
 #include "llama-kv-cache-dsa.h"
+#include "llama-kv-cache-msa.h"
 #include "llama-kv-cache-dsv4.h"
 #include "llama-memory-hybrid.h"
 #include "llama-memory-hybrid-iswa.h"
@@ -518,6 +519,40 @@ bool llm_graph_input_attn_k::can_reuse(const llm_graph_params & params) {
     return res;
 }
 
+llm_graph_input_attn_kv_msa::llm_graph_input_attn_kv_msa(
+        const llama_hparams & hparams,
+        const llama_cparams & cparams,
+        const llama_kv_cache_msa_context * mctx) :
+    llm_graph_input_attn_kv(hparams, cparams, mctx->get_base()),
+    mctx_msa(mctx) {
+}
+
+void llm_graph_input_attn_kv_msa::set_input(const llama_ubatch * ubatch) {
+    llm_graph_input_attn_kv::set_input(ubatch);
+
+    if (self_k_idxs_idx) {
+        mctx_msa->get_idx()->set_input_k_idxs(self_k_idxs_idx, ubatch);
+    }
+}
+
+bool llm_graph_input_attn_kv_msa::can_reuse(const llm_graph_params & params) {
+    mctx_msa = static_cast<const llama_kv_cache_msa_context *>(params.mctx);
+
+    // the parent class operates on the base cache context
+    this->mctx = mctx_msa->get_base();
+
+    bool res = true;
+
+    res &= self_k_idxs->ne[0] == params.ubatch.n_tokens;
+    if (self_k_idxs_idx) {
+        res &= self_k_idxs_idx->ne[0] == params.ubatch.n_tokens;
+    }
+
+    res &= can_reuse_kq_mask(self_kq_mask, this->mctx, params.ubatch, params.cparams);
+
+    return res;
+}
+
 void llm_graph_input_attn_k_dsa::set_input(const llama_ubatch * ubatch) {
     mctx->get_mla()->set_input_k_idxs(self_k_idxs_mla, ubatch);
 
@@ -3187,6 +3222,34 @@ llm_graph_input_attn_k_dsa * llm_graph_context::build_attn_inp_k_dsa() const {
     return (llm_graph_input_attn_k_dsa *) res->add_input(std::move(inp));
 }
 
+llm_graph_input_attn_kv_msa * llm_graph_context::build_attn_inp_kv_msa(bool msa_enabled) const {
+    const auto * mctx_cur = static_cast<const llama_kv_cache_msa_context *>(mctx);
+
+    auto inp = std::make_unique<llm_graph_input_attn_kv_msa>(hparams, cparams, mctx_cur);
+
+    const auto * mctx_base = mctx_cur->get_base();
+    const auto * mctx_idx  = mctx_cur->get_idx();
+
+    {
+        GGML_ASSERT(hparams.swa_type == LLAMA_SWA_TYPE_NONE && "Use llama_kv_cache_iswa for SWA");
+
+        inp->self_k_idxs = mctx_base->build_input_k_idxs(ctx0, ubatch);
+        inp->self_v_idxs = mctx_base->build_input_v_idxs(ctx0, ubatch);
+
+        inp->self_kq_mask = build_attn_inp_kq_mask(ctx0, mctx_base, ubatch, cparams);
+        inp->self_kq_mask_cnv = inp->self_kq_mask;
+    }
+
+    inp->self_k_rot = mctx_base->build_input_k_rot(ctx0);
+    inp->self_v_rot = mctx_base->build_input_v_rot(ctx0);
+
+    if (msa_enabled) {
+        inp->self_k_idxs_idx = mctx_idx->build_input_k_idxs(ctx0, ubatch);
+    }
+
+    return (llm_graph_input_attn_kv_msa *) res->add_input(std::move(inp));
+}
+
 // TODO: maybe separate the inner implementation into a separate function
 //       like with the non-sliding window equivalent
 //       once sliding-window hybrid caches are a thing.
@@ -3620,6 +3683,7 @@ void llm_graph_context::build_sampling() const {
             /*.probs       =*/ nullptr,
             /*.sampled     =*/ nullptr,
             /*.candidates  =*/ nullptr,
+            /*.n_vocab     =*/ logits_seq->ne[0],
         };
 
         assert(sampler->iface->backend_apply);

src/llama-graph.h

@@ -23,6 +23,7 @@ struct llama_memory_context_i;
 
 class llama_kv_cache_context;
 class llama_kv_cache_dsa_context;
+class llama_kv_cache_msa_context;
 class llama_kv_cache_dsv4_raw_context;
 class llama_kv_cache_dsv4_context;
 class llama_kv_cache_iswa_context;
@@ -425,6 +426,26 @@ public:
     const llama_kv_cache_dsa_context * mctx;
 };
 
+// standard K/V attention input against the base cache, plus destination indices for the indexer key cache
+class llm_graph_input_attn_kv_msa : public llm_graph_input_attn_kv {
+public:
+    llm_graph_input_attn_kv_msa(
+            const llama_hparams & hparams,
+            const llama_cparams & cparams,
+            const llama_kv_cache_msa_context * mctx);
+    ~llm_graph_input_attn_kv_msa() = default;
+
+    void set_input(const llama_ubatch * ubatch) override;
+
+    bool can_reuse(const llm_graph_params & params) override;
+
+    ggml_tensor * get_k_idxs_idx() const { return self_k_idxs_idx; }
+
+    ggml_tensor * self_k_idxs_idx = nullptr; // I64 [n_batch]
+
+    const llama_kv_cache_msa_context * mctx_msa;
+};
+
 class llm_graph_input_attn_kv_iswa : public llm_graph_input_i {
 public:
     llm_graph_input_attn_kv_iswa(
@@ -1169,6 +1190,8 @@ struct llm_graph_context {
 
     llm_graph_input_attn_k_dsa * build_attn_inp_k_dsa() const;
 
+    llm_graph_input_attn_kv_msa * build_attn_inp_kv_msa(bool msa_enabled) const;
+
     ggml_tensor * build_attn(
             llm_graph_input_attn_k_dsa * inp,
             ggml_tensor * wo,

src/llama-hparams.cpp

@@ -180,16 +180,6 @@ uint32_t llama_hparams::n_embd_v_gqa_max() const {
     return val;
 }
 
-uint32_t llama_hparams::n_embd_k_idx(uint32_t il) const {
-    if (!indexer_kv || indexer_head_size == 0) {
-        return 0; // arch without a MSA indexer
-    }
-    if (il < n_layer_dense_lead) {
-        return 0; // leading dense layers carry no indexer
-    }
-    return indexer_head_size; // 128
-}
-
 uint32_t llama_hparams::n_embd_r() const {
     if (wkv_head_size != 0) {
         // for RWKV models

src/llama-hparams.h

@@ -230,8 +230,6 @@ struct llama_hparams {
     // MSA
     uint32_t indexer_block_size  = 0;
     uint32_t indexer_local_blocks = 0;
-    // MSA stores its indexer keys in the main KV cache (k_idx tensors);
-    bool indexer_kv = false;
 
     // Indexer is "full" (1) or "shared" (0)
     // Shared indexers reuse top-k from previous full layer
@@ -356,9 +354,6 @@ struct llama_hparams {
     uint32_t n_embd_k_gqa_max() const;
     uint32_t n_embd_v_gqa_max() const;
 
-    // dimension of the single-head MSA indexer key stream
-    uint32_t n_embd_k_idx(uint32_t il = 0) const;
-
     // dimension of the rolling state embeddings
     // corresponds to Mamba's conv_states size or RWKV's token_shift states size
     uint32_t n_embd_r() const;

src/llama-kv-cache-dsa.cpp

@@ -23,7 +23,8 @@ llama_kv_cache_dsa::llama_kv_cache_dsa(
                  uint32_t   n_pad,
                  uint32_t   n_swa,
            llama_swa_type   swa_type,
-    const layer_filter_cb & filter,
+    const layer_filter_cb & filter_mla,
+    const layer_filter_cb & filter_lid,
     const  layer_reuse_cb & reuse) :
     hparams_lid(model.hparams), n_stream(unified ? 1 : n_seq_max) {
 
@@ -32,7 +33,7 @@ llama_kv_cache_dsa::llama_kv_cache_dsa(
     kv_mla = std::make_unique<llama_kv_cache>(
             model, model.hparams, type_k, type_v,
             v_trans, offload, unified, kv_size, n_seq_max, n_pad,
-            n_swa, swa_type, nullptr, filter, reuse, nullptr);
+            n_swa, swa_type, nullptr, filter_mla, reuse, nullptr);
 
     // we use llama_kv_cache for caching indexer keys
     // by hand-tweaking some hparams we fool it to create
@@ -49,7 +50,7 @@ llama_kv_cache_dsa::llama_kv_cache_dsa(
     kv_lid = std::make_unique<llama_kv_cache>(
             model, hparams_lid, type_k, type_v,
             v_trans, offload, unified, kv_size, n_seq_max, n_pad,
-            n_swa, swa_type, nullptr, filter, reuse, nullptr);
+            n_swa, swa_type, nullptr, filter_lid, reuse, nullptr);
 }
 
 void llama_kv_cache_dsa::clear(bool data) {

src/llama-kv-cache-dsa.h

@@ -26,7 +26,8 @@ public:
                      uint32_t   n_pad,
                      uint32_t   n_swa,
                llama_swa_type   swa_type,
-        const layer_filter_cb & filter,
+        const layer_filter_cb & filter_mla,
+        const layer_filter_cb & filter_lid,
         const  layer_reuse_cb & reuse);
 
     ~llama_kv_cache_dsa() = default;

src/llama-kv-cache-msa.cpp

@@ -0,0 +1,395 @@
+#include "llama-kv-cache-msa.h"
+
+#include "llama-impl.h"
+#include "llama-batch.h"
+#include "llama-model.h"
+
+#include <algorithm>
+#include <cassert>
+#include <cmath>
+
+// llama_kv_cache_msa
+
+llama_kv_cache_msa::llama_kv_cache_msa(
+        const llama_model & model,
+                ggml_type   type_k,
+                ggml_type   type_v,
+                     bool   v_trans,
+                     bool   offload,
+                     bool   unified,
+                 uint32_t   kv_size,
+                 uint32_t   n_seq_max,
+                 uint32_t   n_pad,
+                 uint32_t   n_swa,
+           llama_swa_type   swa_type,
+    const layer_filter_cb & filter,
+    const layer_filter_cb & filter_idx,
+    const  layer_reuse_cb & reuse) :
+    hparams_idx(model.hparams),
+    n_stream(unified ? 1 : n_seq_max), n_seq_max(n_seq_max), n_pad(n_pad),
+    n_swa(n_swa), swa_type(swa_type) {
+
+    LLAMA_LOG_INFO("%s: creating main KV cache, size = %u cells\n", __func__, kv_size);
+
+    kv_base = std::make_unique<llama_kv_cache>(
+            model, model.hparams, type_k, type_v,
+            v_trans, offload, unified, kv_size, n_seq_max, n_pad,
+            n_swa, swa_type, nullptr, filter, reuse, nullptr);
+
+    // the MSA indexer uses a single key head per layer
+    std::fill(hparams_idx.n_head_kv_arr.begin(), hparams_idx.n_head_kv_arr.end(), 1);
+    hparams_idx.n_embd_head_k_full = model.hparams.indexer_head_size;
+    // the rope parameters are kept identical to the main cache
+
+    LLAMA_LOG_INFO("%s: creating indexer KV cache, size = %u cells\n", __func__, kv_size);
+
+    kv_idx = std::make_unique<llama_kv_cache>(
+            model, hparams_idx, type_k, type_v,
+            v_trans, offload, unified, kv_size, n_seq_max, n_pad,
+            n_swa, swa_type, nullptr, filter_idx, reuse, nullptr);
+}
+
+void llama_kv_cache_msa::clear(bool data) {
+    kv_base->clear(data);
+    kv_idx ->clear(data);
+}
+
+bool llama_kv_cache_msa::seq_rm(llama_seq_id seq_id, llama_pos p0, llama_pos p1) {
+    bool res = true;
+
+    res = res & kv_base->seq_rm(seq_id, p0, p1);
+    res = res & kv_idx ->seq_rm(seq_id, p0, p1);
+
+    return res;
+}
+
+void llama_kv_cache_msa::seq_cp(llama_seq_id seq_id_src, llama_seq_id seq_id_dst, llama_pos p0, llama_pos p1) {
+    kv_base->seq_cp(seq_id_src, seq_id_dst, p0, p1);
+    kv_idx ->seq_cp(seq_id_src, seq_id_dst, p0, p1);
+}
+
+void llama_kv_cache_msa::seq_keep(llama_seq_id seq_id) {
+    kv_base->seq_keep(seq_id);
+    kv_idx ->seq_keep(seq_id);
+}
+
+void llama_kv_cache_msa::seq_add(llama_seq_id seq_id, llama_pos p0, llama_pos p1, llama_pos shift) {
+    kv_base->seq_add(seq_id, p0, p1, shift);
+    kv_idx ->seq_add(seq_id, p0, p1, shift);
+}
+
+void llama_kv_cache_msa::seq_div(llama_seq_id seq_id, llama_pos p0, llama_pos p1, int d) {
+    kv_base->seq_div(seq_id, p0, p1, d);
+    kv_idx ->seq_div(seq_id, p0, p1, d);
+}
+
+llama_pos llama_kv_cache_msa::seq_pos_min(llama_seq_id seq_id) const {
+    return kv_base->seq_pos_min(seq_id);
+}
+
+llama_pos llama_kv_cache_msa::seq_pos_max(llama_seq_id seq_id) const {
+    return kv_base->seq_pos_max(seq_id);
+}
+
+std::map<ggml_backend_buffer_type_t, size_t> llama_kv_cache_msa::memory_breakdown() const {
+    std::map<ggml_backend_buffer_type_t, size_t> mb = kv_base->memory_breakdown();
+    for (const auto & buft_size : kv_idx->memory_breakdown()) {
+        mb[buft_size.first] += buft_size.second;
+    }
+    return mb;
+}
+
+llama_memory_context_ptr llama_kv_cache_msa::init_batch(
+            llama_batch_allocr & balloc,
+            uint32_t n_ubatch,
+            bool embd_all) {
+    GGML_UNUSED(embd_all);
+
+    do {
+        balloc.split_reset();
+
+        std::vector<llama_ubatch> ubatches;
+        while (true) {
+            auto ubatch = n_stream == 1 ? balloc.split_simple(n_ubatch) : balloc.split_equal(n_ubatch, true, 0);
+
+            if (ubatch.n_tokens == 0) {
+                break;
+            }
+
+            ubatches.push_back(std::move(ubatch));
+        }
+
+        if (balloc.get_n_used() < balloc.get_n_tokens()) {
+            // failed to find a suitable split
+            break;
+        }
+
+        auto sinfos_base = kv_base->prepare(ubatches);
+        if (sinfos_base.empty()) {
+            break;
+        }
+
+        auto sinfos_idx = kv_idx->prepare(ubatches);
+        if (sinfos_idx.empty()) {
+            break;
+        }
+
+        assert(sinfos_base.size() == sinfos_idx.size());
+
+        return std::make_unique<llama_kv_cache_msa_context>(
+                this, std::move(sinfos_base), std::move(sinfos_idx), std::move(ubatches));
+    } while (false);
+
+    return std::make_unique<llama_kv_cache_msa_context>(LLAMA_MEMORY_STATUS_FAILED_PREPARE);
+}
+
+llama_memory_context_ptr llama_kv_cache_msa::init_full() {
+    return std::make_unique<llama_kv_cache_msa_context>(this);
+}
+
+llama_memory_context_ptr llama_kv_cache_msa::init_update(llama_context * lctx, bool optimize) {
+    return std::make_unique<llama_kv_cache_msa_context>(this, lctx, optimize);
+}
+
+bool llama_kv_cache_msa::get_can_shift() const {
+    return kv_base->get_can_shift() &&
+           kv_idx ->get_can_shift() &&
+           kv_base->get_size() == kv_idx->get_size();
+}
+
+void llama_kv_cache_msa::state_write(llama_io_write_i & io, llama_seq_id seq_id, llama_state_seq_flags flags) const {
+    kv_base->state_write(io, seq_id, flags);
+    kv_idx ->state_write(io, seq_id, flags);
+}
+
+void llama_kv_cache_msa::state_read(llama_io_read_i & io, llama_seq_id seq_id, llama_state_seq_flags flags) {
+    kv_base->state_read(io, seq_id, flags);
+    kv_idx ->state_read(io, seq_id, flags);
+}
+
+llama_kv_cache * llama_kv_cache_msa::get_base() const {
+    return kv_base.get();
+}
+
+llama_kv_cache * llama_kv_cache_msa::get_idx() const {
+    return kv_idx.get();
+}
+
+// llama_kv_cache_msa_context
+
+llama_kv_cache_msa_context::llama_kv_cache_msa_context(llama_memory_status status) :
+    kv(nullptr), status(status) {}
+
+llama_kv_cache_msa_context::llama_kv_cache_msa_context(
+        llama_kv_cache_msa * kv) :
+    kv(kv),
+    ctx_base(kv->get_base()->init_full()),
+    ctx_idx (kv->get_idx ()->init_full()),
+    status(llama_memory_status_combine(ctx_base->get_status(), ctx_idx->get_status())) {
+}
+
+llama_kv_cache_msa_context::llama_kv_cache_msa_context(
+        llama_kv_cache_msa * kv,
+        llama_context * lctx,
+        bool optimize) :
+    kv(kv),
+    ctx_base(kv->get_base()->init_update(lctx, optimize)),
+    ctx_idx (kv->get_idx ()->init_update(lctx, optimize)),
+    status(llama_memory_status_combine(ctx_base->get_status(), ctx_idx->get_status())) {
+}
+
+llama_kv_cache_msa_context::llama_kv_cache_msa_context(
+        llama_kv_cache_msa * kv,
+        slot_info_vec_t sinfos_base,
+        slot_info_vec_t sinfos_idx,
+        std::vector<llama_ubatch> ubatches) :
+    kv(kv),
+    ubatches(std::move(ubatches)),
+    // here we copy the ubatches. not sure if this is ideal
+    ctx_base(new llama_kv_cache_context(kv->get_base(), std::move(sinfos_base), this->ubatches)),
+    ctx_idx (new llama_kv_cache_context(kv->get_idx (), std::move(sinfos_idx),  this->ubatches)),
+    status(llama_memory_status_combine(ctx_base->get_status(), ctx_idx->get_status())) {
+}
+
+llama_kv_cache_msa_context::~llama_kv_cache_msa_context() = default;
+
+bool llama_kv_cache_msa_context::next() {
+    assert(status == LLAMA_MEMORY_STATUS_SUCCESS);
+
+    ctx_base->next();
+    ctx_idx ->next();
+
+    if (++i_next >= ubatches.size()) {
+        return false;
+    }
+
+    return true;
+}
+
+bool llama_kv_cache_msa_context::apply() {
+    assert(!llama_memory_status_is_fail(status));
+
+    bool res = true;
+
+    res = res & ctx_base->apply();
+    res = res & ctx_idx ->apply();
+
+    return res;
+}
+
+llama_memory_status llama_kv_cache_msa_context::get_status() const {
+    return status;
+}
+
+const llama_ubatch & llama_kv_cache_msa_context::get_ubatch() const {
+    assert(status == LLAMA_MEMORY_STATUS_SUCCESS);
+
+    return ubatches[i_next];
+}
+
+const llama_kv_cache_context * llama_kv_cache_msa_context::get_base() const {
+    assert(status == LLAMA_MEMORY_STATUS_SUCCESS);
+
+    return static_cast<const llama_kv_cache_context *>(ctx_base.get());
+}
+
+const llama_kv_cache_context * llama_kv_cache_msa_context::get_idx() const {
+    assert(status == LLAMA_MEMORY_STATUS_SUCCESS);
+
+    return static_cast<const llama_kv_cache_context *>(ctx_idx.get());
+}
+
+uint32_t llama_kv_cache_msa_context::get_n_pos() const {
+    // pad the value so that the graph remains constant across batches and can be reused
+    const uint32_t n_pad_cur = std::max(kv->get_n_pad(), 256u);
+
+    llama_pos pos_max = -1;
+
+    for (llama_seq_id seq_id = 0; seq_id < (llama_seq_id) kv->get_n_seq_max(); ++seq_id) {
+        pos_max = std::max(pos_max, kv->seq_pos_max(seq_id));
+    }
+
+    return std::max(n_pad_cur, GGML_PAD((uint32_t) (pos_max + 1), n_pad_cur));
+}
+
+void llama_kv_cache_msa_context::set_input_cell_pos(ggml_tensor * dst, const llama_ubatch * ubatch, int32_t div) const {
+    GGML_ASSERT(ggml_backend_buffer_is_host(dst->buffer));
+    GGML_ASSERT(dst->type == GGML_TYPE_I32);
+    GGML_ASSERT(div > 0);
+
+    const int64_t n_tokens    = ubatch->n_tokens;
+    const int64_t n_kv        = dst->ne[0];
+    const int64_t n_stream_ub = dst->ne[1];
+
+    GGML_ASSERT(n_tokens % n_stream_ub == 0);
+    const int64_t n_tps = n_tokens/n_stream_ub;
+
+    int32_t * data = (int32_t *) dst->data;
+
+    for (int64_t s = 0; s < n_stream_ub; ++s) {
+        const llama_seq_id seq_id = ubatch->seq_id[s*n_tps][0];
+
+        const auto & cells = kv->get_base()->get_cells(seq_id);
+
+        for (int64_t j = 0; j < n_kv; ++j) {
+            // the value for empty or other-sequence cells is irrelevant as consumers mask them
+            data[s*n_kv + j] =
+                cells.is_empty(j) || !cells.seq_has(j, seq_id)
+                    ? 0
+                    : (int32_t) (cells.pos_get(j)/div);
+        }
+    }
+}
+
+void llama_kv_cache_msa_context::set_input_pos_slot(ggml_tensor * dst, const llama_ubatch * ubatch) const {
+    GGML_ASSERT(ggml_backend_buffer_is_host(dst->buffer));
+    GGML_ASSERT(dst->type == GGML_TYPE_I32 || dst->type == GGML_TYPE_F32);
+
+    const int64_t n_tokens    = ubatch->n_tokens;
+    const int64_t n_pos       = dst->ne[0];
+    const int64_t n_stream_ub = dst->ne[1];
+
+    GGML_ASSERT(n_tokens % n_stream_ub == 0);
+    const int64_t n_tps = n_tokens/n_stream_ub;
+
+    for (int64_t s = 0; s < n_stream_ub; ++s) {
+        const llama_seq_id seq_id = ubatch->seq_id[s*n_tps][0];
+
+        const auto & cells = kv->get_base()->get_cells(seq_id);
+
+        std::vector<int32_t> map(n_pos, 0);
+
+        for (uint32_t j = 0; j < cells.size(); ++j) {
+            if (cells.is_empty(j) || !cells.seq_has(j, seq_id)) {
+                continue;
+            }
+
+            const llama_pos p0 = cells.pos_get(j);
+
+            if (p0 < 0 || p0 >= n_pos) {
+                continue;
+            }
+
+            map[p0] = (int32_t) j;
+        }
+
+        if (dst->type == GGML_TYPE_I32) {
+            int32_t * data = (int32_t *) dst->data + s*n_pos;
+            std::copy(map.begin(), map.end(), data);
+        } else {
+            float * data = (float *) dst->data + s*n_pos;
+            for (int64_t p = 0; p < n_pos; ++p) {
+                data[p] = (float) map[p];
+            }
+        }
+    }
+}
+
+void llama_kv_cache_msa_context::set_input_pos_mask(ggml_tensor * dst, const llama_ubatch * ubatch) const {
+    GGML_ASSERT(ggml_backend_buffer_is_host(dst->buffer));
+    GGML_ASSERT(dst->type == GGML_TYPE_F32);
+
+    const int64_t n_tokens = ubatch->n_tokens;
+    const int64_t n_pos    = dst->ne[0];
+
+    GGML_ASSERT(dst->ne[1] == n_tokens);
+
+    const uint32_t       n_swa    = kv->get_n_swa();
+    const llama_swa_type swa_type = kv->get_swa_type();
+
+    float * data = (float *) dst->data;
+
+    std::fill(data, data + n_pos*n_tokens, -INFINITY);
+
+    for (int64_t i = 0; i < n_tokens; ++i) {
+        const llama_seq_id seq_id = ubatch->seq_id[i][0];
+
+        const auto & cells = kv->get_base()->get_cells(seq_id);
+
+        const llama_pos p1 = ubatch->pos[i];
+
+        for (uint32_t j = 0; j < cells.size(); ++j) {
+            if (cells.is_empty(j) || !cells.seq_has(j, seq_id)) {
+                continue;
+            }
+
+            const llama_pos p0 = cells.pos_get(j);
+
+            if (p0 < 0 || p0 >= n_pos) {
+                continue;
+            }
+
+            // causal mask
+            if (p0 > p1) {
+                continue;
+            }
+
+            // apply SWA if any
+            if (llama_hparams::is_masked_swa(n_swa, swa_type, p0, p1)) {
+                continue;
+            }
+
+            data[i*n_pos + p0] = 0.0f;
+        }
+    }
+}

src/llama-kv-cache-msa.h

@@ -0,0 +1,153 @@
+#pragma once
+
+#include "llama-kv-cache.h"
+
+#include <vector>
+
+// llama_kv_cache_msa
+
+// uses two instances of llama_kv_cache, one for K/V tensors, and one for the MSA indexer tensors
+// both receive identical sequence operations and identical ubatches, so their cell layouts stay in synced.
+// the context also exposes per-ubatch pos - cell translation maps populated from llama_kv_cells via
+// llama_kv_cache::get_cells(), which the model graph uses to run MSA block selection in position space
+
+class llama_kv_cache_msa : public llama_memory_i {
+public:
+    llama_kv_cache_msa(
+            const llama_model & model,
+                    ggml_type   type_k,
+                    ggml_type   type_v,
+                         bool   v_trans,
+                         bool   offload,
+                         bool   unified,
+                     uint32_t   kv_size,
+                     uint32_t   n_seq_max,
+                     uint32_t   n_pad,
+                     uint32_t   n_swa,
+               llama_swa_type   swa_type,
+        const layer_filter_cb & filter,
+        const layer_filter_cb & filter_idx,
+        const  layer_reuse_cb & reuse);
+
+    ~llama_kv_cache_msa() = default;
+
+    // llama_memory_i
+
+    llama_memory_context_ptr init_batch(
+            llama_batch_allocr & balloc,
+            uint32_t n_ubatch,
+            bool embd_all) override;
+
+    llama_memory_context_ptr init_full() override;
+
+    llama_memory_context_ptr init_update(llama_context * lctx, bool optimize) override;
+
+    bool get_can_shift() const override;
+
+    void clear(bool data) override;
+
+    bool seq_rm  (llama_seq_id seq_id,                              llama_pos p0, llama_pos p1) override;
+    void seq_cp  (llama_seq_id seq_id_src, llama_seq_id seq_id_dst, llama_pos p0, llama_pos p1) override;
+    void seq_keep(llama_seq_id seq_id)                                                          override;
+    void seq_add (llama_seq_id seq_id,                              llama_pos p0, llama_pos p1, llama_pos shift) override;
+    void seq_div (llama_seq_id seq_id,                              llama_pos p0, llama_pos p1, int d) override;
+
+    llama_pos seq_pos_min(llama_seq_id seq_id) const override;
+    llama_pos seq_pos_max(llama_seq_id seq_id) const override;
+
+    std::map<ggml_backend_buffer_type_t, size_t> memory_breakdown() const override;
+
+    // state write/load
+
+    void state_write(llama_io_write_i & io, llama_seq_id seq_id = -1, llama_state_seq_flags flags = 0) const override;
+    void state_read (llama_io_read_i  & io, llama_seq_id seq_id = -1, llama_state_seq_flags flags = 0) override;
+
+    // llama_kv_cache_msa specific API
+
+    llama_kv_cache * get_base() const;
+    llama_kv_cache * get_idx () const;
+
+    uint32_t       get_n_pad()    const { return n_pad; }
+    uint32_t       get_n_seq_max() const { return n_seq_max; }
+    uint32_t       get_n_swa()    const { return n_swa; }
+    llama_swa_type get_swa_type() const { return swa_type; }
+
+private:
+    // keep the indexer KV cache hparams instance here as llama_kv_cache stores only a reference
+    llama_hparams hparams_idx;
+
+    const uint32_t n_stream  = 1;
+    const uint32_t n_seq_max = 1;
+    const uint32_t n_pad     = 1;
+
+    const uint32_t       n_swa    = 0;
+    const llama_swa_type swa_type = LLAMA_SWA_TYPE_NONE;
+
+    std::unique_ptr<llama_kv_cache> kv_base;
+    std::unique_ptr<llama_kv_cache> kv_idx;
+};
+
+class llama_kv_cache_msa_context : public llama_memory_context_i {
+public:
+    using slot_info_vec_t = llama_kv_cache::slot_info_vec_t;
+
+    // used for errors
+    llama_kv_cache_msa_context(llama_memory_status status);
+
+    // used to create a full-cache context
+    llama_kv_cache_msa_context(
+            llama_kv_cache_msa * kv);
+
+    // used to create an update context
+    llama_kv_cache_msa_context(
+            llama_kv_cache_msa * kv,
+            llama_context * lctx,
+            bool optimize);
+
+    // used to create a batch processing context from a batch
+    llama_kv_cache_msa_context(
+            llama_kv_cache_msa * kv,
+            slot_info_vec_t sinfos_base,
+            slot_info_vec_t sinfos_idx,
+            std::vector<llama_ubatch> ubatches);
+
+    virtual ~llama_kv_cache_msa_context();
+
+    // llama_memory_context_i
+
+    bool next()  override;
+    bool apply() override;
+
+    llama_memory_status  get_status() const override;
+    const llama_ubatch & get_ubatch() const override;
+
+    // llama_kv_cache_msa_context specific API
+
+    const llama_kv_cache_context * get_base() const;
+    const llama_kv_cache_context * get_idx () const;
+
+    // max position currently present in the cache plus one, padded MSA blocks are defined over token positions
+    // so the block-selection tensors are sized by this value rather than by the number of cells
+    uint32_t get_n_pos() const;
+
+    // position <-> cell translation maps, populated from the base cache cells
+    // the model graph relates cache contents to token positions only through these per ubatch inputs
+    // value for empty or other-sequence cells is 0 so consumers must mask them
+    void set_input_cell_pos(ggml_tensor * dst, const llama_ubatch * ubatch, int32_t div) const;
+    // positions without a cell map to cell 0, consumers must mask them assumes one sequence per stream
+    void set_input_pos_slot(ggml_tensor * dst, const llama_ubatch * ubatch) const;
+    void set_input_pos_mask(ggml_tensor * dst, const llama_ubatch * ubatch) const;
+
+private:
+    llama_kv_cache_msa * kv;
+
+    // the index of the next ubatch to process
+    size_t i_next = 0;
+
+    std::vector<llama_ubatch> ubatches;
+
+    const llama_memory_context_ptr ctx_base;
+    const llama_memory_context_ptr ctx_idx;
+
+    const llama_memory_status status;
+};

src/llama-kv-cache.cpp

@@ -112,7 +112,7 @@ llama_kv_cache::llama_kv_cache(
         auto it = ctx_map.find(buft);
         if (it == ctx_map.end()) {
             ggml_init_params params = {
-                /*.mem_size   =*/ size_t(3u*(1 + n_stream)*n_layer*ggml_tensor_overhead()), //Reserve tensor metadata for up to 3 tensors per layer (K, V, and optional K_idx), plus one view per tensor per stream.
+                /*.mem_size   =*/ size_t(2u*(1 + n_stream)*n_layer*ggml_tensor_overhead()),
                 /*.mem_buffer =*/ NULL,
                 /*.no_alloc   =*/ true,
             };
@@ -242,25 +242,9 @@ llama_kv_cache::llama_kv_cache(
             v_stream.push_back(has_v ? ggml_view_2d(ctx, v, n_embd_v_gqa, kv_size, v->nb[1], s*v->nb[2]) : nullptr);
         }
 
-        const uint32_t n_embd_k_idx = hparams.n_embd_k_idx(il);
-        ggml_tensor * k_idx = n_embd_k_idx > 0
-            ? ggml_new_tensor_3d(ctx, GGML_TYPE_F32, n_embd_k_idx, kv_size, n_stream)
-            : nullptr;
-        if (k_idx) {
-            ggml_format_name(k_idx, "cache_k_idx_l%d", il);
-            msa_strict_slots = (n_stream == n_seq_max);
-        }
-
-        std::vector<ggml_tensor *> k_idx_stream;
-        for (uint32_t s = 0; s < n_stream; ++s) {
-            k_idx_stream.push_back(k_idx
-                ? ggml_view_2d(ctx, k_idx, n_embd_k_idx, kv_size, k_idx->nb[1], s*k_idx->nb[2])
-                : nullptr);
-        }
-
         map_layer_ids[il] = layers.size();
 
-        layers.push_back({ il, k, v, k_idx, k_stream, v_stream, k_idx_stream });
+        layers.push_back({ il, k, v, k_stream, v_stream, });
     }
 
     if (reuse) {
@@ -309,24 +293,13 @@ llama_kv_cache::llama_kv_cache(
     }
 
     {
-        const size_t memory_size_k     = size_k_bytes();
-        const size_t memory_size_v     = size_v_bytes();
-        const size_t memory_size_k_idx = size_k_idx_bytes();
-        const size_t memory_size_total = memory_size_k + memory_size_v + memory_size_k_idx;
+        const size_t memory_size_k = size_k_bytes();
+        const size_t memory_size_v = size_v_bytes();
 
-        constexpr float mib = 1024.0f * 1024.0f;
-
-        const std::string k_log = format(", K (%s): %7.2f MiB", ggml_type_name(type_k), (float) memory_size_k / mib);
-        const std::string v_log = format(", V (%s): %7.2f MiB", ggml_type_name(type_v), (float) memory_size_v / mib);
-
-        std::string k_idx_log;
-        if (memory_size_k_idx > 0) {
-            k_idx_log = format(", K_idx (%s): %7.2f MiB", ggml_type_name(GGML_TYPE_F32), (float) memory_size_k_idx / mib);
-        }
-
-        LLAMA_LOG_INFO("%s: size = %7.2f MiB (%6u cells, %3d layers, %2u/%u seqs)%s%s%s\n", __func__,
-                (float) memory_size_total / mib, kv_size, (int) layers.size(), n_seq_max, n_stream,
-                k_log.c_str(), v_log.c_str(), k_idx_log.c_str());
+        LLAMA_LOG_INFO("%s: size = %7.2f MiB (%6u cells, %3d layers, %2u/%u seqs), K (%s): %7.2f MiB, V (%s): %7.2f MiB\n", __func__,
+                (float)(memory_size_k + memory_size_v) / (1024.0f * 1024.0f), kv_size, (int) layers.size(), n_seq_max, n_stream,
+                ggml_type_name(type_k), (float)memory_size_k / (1024.0f * 1024.0f),
+                ggml_type_name(type_v), (float)memory_size_v / (1024.0f * 1024.0f));
     }
 
     // TODO: refactor [TAG_KV_CACHE_SHARE_CELLS]
@@ -419,39 +392,6 @@ bool llama_kv_cache::seq_rm(llama_seq_id seq_id, llama_pos p0, llama_pos p1) {
         p1 = std::numeric_limits<llama_pos>::max();
     }
 
-    // empty range - nothing to remove
-    if (p0 >= p1) {
-        return true;
-    }
-
-    // MSA anchors block selection to absolute cache slots (slot == position). Tail trim and full removal preserve this invariant, but removing a prefix
-    // or middle range would free slots while later cells survive, desynchronizing the indexer cache. Reject such removals before modifying the cache.
-    if (msa_strict_slots) {
-        for (llama_seq_id sid = 0; sid < (llama_seq_id) seq_to_stream.size(); ++sid) {
-            if (seq_id >= 0 && sid != seq_id) {
-                continue;
-            }
-
-            const auto & cells = v_cells[seq_to_stream[sid]];
-
-            const llama_pos pmin = cells.seq_pos_min(sid);
-            const llama_pos pmax = cells.seq_pos_max(sid);
-
-            if (pmin < 0) {
-                continue;   // empty sequence
-            }
-
-            const bool overlaps    = p0 <= pmax && p1 > pmin;   // the range removes something
-            const bool leaves_tail = p1 <= pmax;                // cells beyond the range survive
-
-            if (overlaps && leaves_tail) {
-                LLAMA_LOG_WARN("%s: MSA: partial (non-suffix) removal [%d, %d) for seq %d is not supported "
-                        "(block selection is anchored to cache slots) - rejected\n", __func__, p0, p1, sid);
-                return false;
-            }
-        }
-    }
-
     if (seq_id >= 0) {
         auto & cells = v_cells[seq_to_stream[seq_id]];
         auto & head  = v_heads[seq_to_stream[seq_id]];
@@ -906,10 +846,6 @@ bool llama_kv_cache::update(llama_context * lctx, bool do_shift, const stream_co
                 if (layer.v_stream[ssrc]) {
                     ggml_backend_tensor_copy(layer.v_stream[ssrc], layer.v_stream[sdst]);
                 }
-                if (layer.k_idx_stream[ssrc]) {
-                    GGML_ASSERT(layer.k_idx_stream[sdst]);
-                    ggml_backend_tensor_copy(layer.k_idx_stream[ssrc], layer.k_idx_stream[sdst]);
-                }
             }
         }
     }
@@ -1058,44 +994,6 @@ llama_kv_cache::slot_info llama_kv_cache::find_slot(const llama_ubatch & ubatch,
 
         const auto & cells = v_cells[seq_to_stream[seq_id]];
 
-        if (n_tokens > cells.size()) {
-            LLAMA_LOG_ERROR("%s: n_tokens = %d > size = %u\n", __func__, n_tokens, cells.size());
-            return { };
-        }
-
-        // MSA block selection assumes slot == logical position (append-only streams).
-        if (msa_strict_slots) {
-            for (uint32_t ii = 0; ii < n_tokens; ++ii) {
-                const llama_pos pos = ubatch.pos[s*n_tokens + ii];
-
-                if (pos < 0 || (uint64_t) pos >= cells.size()) {
-                    LLAMA_LOG_WARN("%s: MSA: position %d is outside the cache range [0, %u)\n",
-                            __func__, pos, cells.size());
-                    return { };
-                }
-
-                const uint32_t idx = (uint32_t) pos;
-
-                if (!cells.is_empty(idx)) {
-                    LLAMA_LOG_WARN("%s: MSA: required slot %u is already occupied (stream %u)\n",
-                            __func__, idx, seq_to_stream[seq_id]);
-                    return { };
-                }
-
-                // strictly increasing positions, rules out duplicates and, for contiguous requests, is tightened to exact adjacency
-                if (!res.idxs[s].empty() && (cont ? idx != res.idxs[s].back() + 1
-                                                  : idx <= res.idxs[s].back())) {
-                    LLAMA_LOG_WARN("%s: MSA: token positions are not %s within the ubatch\n",
-                            __func__, cont ? "contiguous" : "strictly increasing");
-                    return { };
-                }
-
-                res.idxs[s].push_back(idx);
-            }
-
-            continue;
-        }
-
         uint32_t head_cur = v_heads[seq_to_stream[seq_id]];
 
         // if we have enough unused cells before the current head ->
@@ -1104,6 +1002,11 @@ llama_kv_cache::slot_info llama_kv_cache::find_slot(const llama_ubatch & ubatch,
             head_cur = 0;
         }
 
+        if (n_tokens > cells.size()) {
+            LLAMA_LOG_ERROR("%s: n_tokens = %d > size = %u\n", __func__, n_tokens, cells.size());
+            return { };
+        }
+
         uint32_t n_tested = 0;
 
         // for continuous slots, we test that all tokens in the ubatch fit, starting from the current head
@@ -1210,15 +1113,6 @@ void llama_kv_cache::apply_ubatch(const slot_info & sinfo, const llama_ubatch &
 
             const auto idx = sinfo.idxs[s][ii];
 
-            if (msa_strict_slots && (llama_pos) idx != ubatch.pos[i]) {
-                LLAMA_LOG_ERROR("%s: MSA slot/position invariant violated: "
-                        "writing pos %d into cell %u (stream %u). The indexer cache "
-                        "would desync and block selection would silently corrupt. "
-                        "This is a bug, please report it with reproduction steps.\n",
-                        __func__, ubatch.pos[i], idx, sinfo.strm[s]);
-                GGML_ABORT("MSA: slot != pos");
-            }
-
             if (!cells.is_empty(idx)) {
                 assert(cells.seq_count(idx) == 1);
 
@@ -1262,8 +1156,7 @@ void llama_kv_cache::apply_ubatch(const slot_info & sinfo, const llama_ubatch &
             LLAMA_LOG_DEBUG("%s: purging positions [%d, %d] of sequence %d from KV cache\n",
                     __func__, cells.seq_pos_min(s), seq_pos_max_rm[s], s);
 
-            // under MSA strict slots this path should be unreachable, since strict MSA placement never selects occupied cells
-            GGML_ASSERT(seq_rm(s, cells.seq_pos_min(s), seq_pos_max_rm[s] + 1));
+            seq_rm(s, cells.seq_pos_min(s), seq_pos_max_rm[s] + 1);
         }
     }
 
@@ -1283,12 +1176,6 @@ bool llama_kv_cache::get_can_shift() const {
     if (hparams.n_pos_per_embd() > 1) {
         return false;
     }
-    // shifting would leave k_idx stale
-    for (const auto & layer : layers) {
-        if (layer.k_idx) {
-            return false;
-        }
-    }
     return true;
 }
 
@@ -1337,6 +1224,12 @@ ggml_tensor * llama_kv_cache::get_k_storage(int32_t il) const {
     return layers[ikv].k;
 }
 
+const llama_kv_cells & llama_kv_cache::get_cells(llama_seq_id seq_id) const {
+    GGML_ASSERT(seq_id >= 0 && (size_t) seq_id < seq_to_stream.size());
+
+    return v_cells[seq_to_stream[seq_id]];
+}
+
 uint32_t llama_kv_cache::get_n_kv(const slot_info & sinfo) const {
     uint32_t result = 0;
 
@@ -1405,23 +1298,6 @@ ggml_tensor * llama_kv_cache::get_v(ggml_context * ctx, int32_t il, uint32_t n_k
             ggml_row_size(v->type, kv_size*n_embd_v_gqa)*sinfo.s0);
 }
 
-ggml_tensor * llama_kv_cache::get_k_idx(ggml_context * ctx, int32_t il, uint32_t n_kv, const slot_info & sinfo) const {
-    const int32_t ikv = map_layer_ids.at(il);
-    auto * k_idx = layers[ikv].k_idx;
-    GGML_ASSERT(k_idx);
-
-    const uint64_t kv_size = get_size();
-    const int64_t  n_idx   = k_idx->ne[0];                 // 128
-    const uint32_t ns      = sinfo.s1 - sinfo.s0 + 1;
-
-    return ggml_view_4d(ctx, k_idx,
-            n_idx, 1, n_kv, ns,
-            ggml_row_size(k_idx->type, n_idx),             // nb1 (single head)
-            ggml_row_size(k_idx->type, n_idx),             // nb2 (per cell)
-            ggml_row_size(k_idx->type, n_idx*kv_size),     // nb3 (per stream)
-            ggml_row_size(k_idx->type, n_idx*kv_size)*sinfo.s0);
-}
-
 ggml_tensor * llama_kv_cache::cpy_k(ggml_context * ctx, ggml_tensor * k_cur, ggml_tensor * k_idxs, int32_t il, const slot_info & sinfo) const {
     GGML_UNUSED(sinfo);
 
@@ -1523,28 +1399,6 @@ ggml_tensor * llama_kv_cache::build_input_k_idxs(ggml_context * ctx, const llama
     return k_idxs;
 }
 
-ggml_tensor * llama_kv_cache::cpy_k_idx(ggml_context * ctx, ggml_tensor * k_idx_cur, ggml_tensor * k_idxs, int32_t il, const slot_info & sinfo) const {
-    GGML_UNUSED(sinfo);
-    const int32_t ikv = map_layer_ids.at(il);
-    ggml_tensor * k_idx = layers[ikv].k_idx;
-    GGML_ASSERT(k_idx && "cpy_k_idx on a layer with no indexer cache");
-
-    const int64_t n_embd_head = k_idx_cur->ne[0];          // 128
-    const int64_t n_head      = k_idx_cur->ne[1];          // 1
-    const int64_t n_tokens    = k_idx_cur->ne[2];
-    const int64_t n_embd_gqa  = n_embd_head*n_head;        // 128
-
-    GGML_ASSERT(ggml_row_size(k_idx_cur->type, n_embd_head) == k_idx_cur->nb[1]);
-    k_idx_cur = ggml_view_2d(ctx, k_idx_cur, n_embd_gqa, n_tokens, k_idx_cur->nb[2], 0);
-
-    const int64_t n_stream = k_idx->ne[2];
-    if (n_stream > 1) {
-        const int64_t kv_size = get_size();
-        k_idx = ggml_reshape_2d(ctx, k_idx, n_embd_gqa, kv_size*n_stream);
-    }
-    return ggml_set_rows(ctx, k_idx, k_idx_cur, k_idxs);   // same k_idxs as the K store
-}
-
 ggml_tensor * llama_kv_cache::build_input_v_idxs(ggml_context * ctx, const llama_ubatch & ubatch) const {
     const uint32_t n_tokens = ubatch.n_tokens;
 
@@ -1979,18 +1833,6 @@ size_t llama_kv_cache::size_v_bytes() const {
     return size_v_bytes;
 }
 
-size_t llama_kv_cache::size_k_idx_bytes() const {
-    size_t size_k_idx_bytes = 0;
-
-    for (const auto & layer : layers) {
-        if (layer.k_idx) {
-            size_k_idx_bytes += ggml_nbytes(layer.k_idx);
-        }
-    }
-
-    return size_k_idx_bytes;
-}
-
 ggml_tensor * llama_kv_cache::build_rope_shift(
         const llama_cparams & cparams,
                ggml_context * ctx,
@@ -2303,36 +2145,6 @@ void llama_kv_cache::state_write_data(llama_io_write_i & io, const cell_ranges_t
         }
     }
 
-    if (size_k_idx_bytes() > 0) {
-        const uint32_t has_k_idx_u32 = 1;
-        io.write(&has_k_idx_u32, sizeof(has_k_idx_u32));
-
-        for (const auto & layer : layers) {
-            const uint32_t layer_has_k_idx = layer.k_idx ? 1 : 0;
-            io.write(&layer_has_k_idx, sizeof(layer_has_k_idx));
-
-            if (!layer_has_k_idx) {
-                continue;
-            }
-
-            GGML_ASSERT(layer.k_idx_stream[cr.strm]);
-
-            const int32_t k_idx_type_i = (int32_t) layer.k_idx->type;
-            io.write(&k_idx_type_i, sizeof(k_idx_type_i));
-
-            const uint64_t k_idx_size_row = ggml_row_size(layer.k_idx->type, layer.k_idx->ne[0]);
-            io.write(&k_idx_size_row, sizeof(k_idx_size_row));
-
-            for (const auto & range : cr.data) {
-                const size_t range_size = range.second - range.first;
-                const size_t buf_size   = range_size * k_idx_size_row;
-                const size_t offset     = range.first * k_idx_size_row;
-
-                io.write_tensor(layer.k_idx_stream[cr.strm], offset, buf_size);
-            }
-        }
-    }
-
     if (!v_trans) {
         for (const auto & layer : layers) {
             const uint32_t il = layer.il;
@@ -2581,68 +2393,6 @@ bool llama_kv_cache::state_read_data(llama_io_read_i & io, uint32_t strm, uint32
         }
     }
 
-    if (size_k_idx_bytes() > 0) {
-        uint32_t has_k_idx_u32 = 0;
-        io.read(&has_k_idx_u32, sizeof(has_k_idx_u32));
-
-        if (has_k_idx_u32 != 1) {
-            LLAMA_LOG_ERROR("%s: missing k_idx data in KV cache state\n", __func__);
-            return false;
-        }
-
-        for (const auto & layer : layers) {
-            uint32_t layer_has_k_idx = 0;
-            io.read(&layer_has_k_idx, sizeof(layer_has_k_idx));
-
-            const uint32_t expected_layer_has_k_idx = layer.k_idx ? 1 : 0;
-
-            if (layer_has_k_idx != expected_layer_has_k_idx) {
-                LLAMA_LOG_ERROR(
-                    "%s: mismatched k_idx state for layer: got %u, expected %u\n",
-                    __func__, layer_has_k_idx, expected_layer_has_k_idx);
-                return false;
-            }
-
-            if (!layer_has_k_idx) {
-                continue;
-            }
-
-            GGML_ASSERT(layer.k_idx_stream[strm]);
-
-            int32_t k_idx_type_i = -1;
-            io.read(&k_idx_type_i, sizeof(k_idx_type_i));
-
-            if (k_idx_type_i != (int32_t) layer.k_idx->type) {
-                LLAMA_LOG_ERROR(
-                    "%s: mismatched k_idx type: got %d, expected %d\n",
-                    __func__, k_idx_type_i, (int32_t) layer.k_idx->type);
-                return false;
-            }
-
-            uint64_t k_idx_size_row = 0;
-            io.read(&k_idx_size_row, sizeof(k_idx_size_row));
-
-            const uint64_t expected_k_idx_size_row = ggml_row_size(layer.k_idx->type, layer.k_idx->ne[0]);
-
-            if (k_idx_size_row != expected_k_idx_size_row) {
-                LLAMA_LOG_ERROR(
-                    "%s: mismatched k_idx row size: got %zu, expected %zu\n",
-                    __func__, (size_t) k_idx_size_row, (size_t) expected_k_idx_size_row);
-                return false;
-            }
-
-            if (cell_count) {
-                if (sinfo.is_contiguous()) {
-                    io.read_tensor(layer.k_idx_stream[strm], sinfo.head() * k_idx_size_row, cell_count * k_idx_size_row);
-                } else {
-                    for (uint32_t i = 0; i < cell_count; ++i) {
-                        io.read_tensor(layer.k_idx_stream[strm], sinfo.idxs[0][i] * k_idx_size_row, k_idx_size_row);
-                    }
-                }
-            }
-        }
-    }
-
     if (!this->v_trans) {
         for (const auto & layer : layers) {
             const uint32_t il = layer.il;
@@ -2844,10 +2594,6 @@ ggml_tensor * llama_kv_cache_context::get_v(ggml_context * ctx, int32_t il) cons
     return kv->get_v(ctx, il, n_kv, sinfos[i_cur]);
 }
 
-ggml_tensor * llama_kv_cache_context::get_k_idx(ggml_context * ctx, int32_t il) const {
-    return kv->get_k_idx(ctx, il, n_kv, sinfos[i_cur]);
-}
-
 ggml_tensor * llama_kv_cache_context::cpy_k(ggml_context * ctx, ggml_tensor * k_cur, ggml_tensor * k_idxs, int32_t il) const {
     return kv->cpy_k(ctx, k_cur, k_idxs, il, sinfos[i_cur]);
 }
@@ -2856,10 +2602,6 @@ ggml_tensor * llama_kv_cache_context::cpy_v(ggml_context * ctx, ggml_tensor * v_
     return kv->cpy_v(ctx, v_cur, v_idxs, il, sinfos[i_cur]);
 }
 
-ggml_tensor * llama_kv_cache_context::cpy_k_idx(ggml_context * ctx, ggml_tensor * k_idx_cur, ggml_tensor * k_idxs, int32_t il) const {
-    return kv->cpy_k_idx(ctx, k_idx_cur, k_idxs, il, sinfos[i_cur]);
-}
-
 ggml_tensor * llama_kv_cache_context::build_input_k_idxs(ggml_context * ctx, const llama_ubatch & ubatch) const {
     return kv->build_input_k_idxs(ctx, ubatch);
 }

src/llama-kv-cache.h

@@ -164,6 +164,8 @@ public:
     std::vector<uint32_t> get_layer_ids() const;
     ggml_tensor * get_k_storage(int32_t il) const;
 
+    const llama_kv_cells & get_cells(llama_seq_id seq_id) const;
+
     //
     // graph_build API
     //
@@ -173,12 +175,10 @@ public:
     // get views of the current state of the cache
     ggml_tensor * get_k(ggml_context * ctx, int32_t il, uint32_t n_kv, const slot_info & sinfo) const;
     ggml_tensor * get_v(ggml_context * ctx, int32_t il, uint32_t n_kv, const slot_info & sinfo) const;
-    ggml_tensor * get_k_idx(ggml_context * ctx, int32_t il, uint32_t n_kv, const slot_info & sinfo) const;
 
     // store k_cur and v_cur in the cache based on the provided head location
     ggml_tensor * cpy_k(ggml_context * ctx, ggml_tensor * k_cur, ggml_tensor * k_idxs, int32_t il, const slot_info & sinfo) const;
     ggml_tensor * cpy_v(ggml_context * ctx, ggml_tensor * v_cur, ggml_tensor * v_idxs, int32_t il, const slot_info & sinfo) const;
-    ggml_tensor * cpy_k_idx(ggml_context * ctx, ggml_tensor * k_idx_cur, ggml_tensor * k_idxs, int32_t il, const slot_info & sinfo) const;
 
     //
     // preparation API
@@ -230,11 +230,9 @@ private:
 
         ggml_tensor * k;
         ggml_tensor * v;
-        ggml_tensor * k_idx;   // MSA single-head indexer keys, F32
 
         std::vector<ggml_tensor *> k_stream;
         std::vector<ggml_tensor *> v_stream;
-        std::vector<ggml_tensor *> k_idx_stream;
     };
 
     bool v_trans = true;  // the value tensor is transposed
@@ -263,9 +261,6 @@ private:
     // env: LLAMA_KV_CACHE_DEBUG
     int debug = 0;
 
-    // set when a k_idx (indexer) cache exists and the stream layout supports MSA (single seq, or one stream per seq)
-    bool msa_strict_slots = false;
-
     // this is the SWA type of the cache - not to be confused with the model SWA type
     const llama_swa_type swa_type = LLAMA_SWA_TYPE_NONE;
 
@@ -298,7 +293,6 @@ private:
 
     size_t size_k_bytes() const;
     size_t size_v_bytes() const;
-    size_t size_k_idx_bytes() const;
 
     ggml_tensor * build_rope_shift(
             const llama_cparams & cparams,
@@ -378,7 +372,6 @@ public:
     // get views of the current state of the cache
     ggml_tensor * get_k(ggml_context * ctx, int32_t il) const;
     ggml_tensor * get_v(ggml_context * ctx, int32_t il) const;
-    ggml_tensor * get_k_idx(ggml_context * ctx, int32_t il) const;
 
     // store k_cur and v_cur in the cache based on the provided head location
     // note: the heads in k_cur and v_cur should be laid out contiguously in memory
@@ -388,7 +381,6 @@ public:
     //   - v_idxs [n_tokens] or [n_tokens*n_embd_v_gqa] depending if V cache is transposed
     ggml_tensor * cpy_k(ggml_context * ctx, ggml_tensor * k_cur, ggml_tensor * k_idxs, int32_t il) const;
     ggml_tensor * cpy_v(ggml_context * ctx, ggml_tensor * v_cur, ggml_tensor * v_idxs, int32_t il) const;
-    ggml_tensor * cpy_k_idx(ggml_context * ctx, ggml_tensor * k_idx_cur, ggml_tensor * k_idxs, int32_t il) const;
 
     // create destination indices for each head of the current batch for where it would be written in the KV cache
     // the indices address the global KV cache (not per stream) - this is not relevant for the user of this API, but

src/llama-model.cpp

@@ -11,6 +11,7 @@
 #include "llama-kv-cache.h"
 #include "llama-kv-cache-iswa.h"
 #include "llama-kv-cache-dsa.h"
+#include "llama-kv-cache-msa.h"
 #include "llama-kv-cache-dsv4.h"
 #include "llama-memory-hybrid.h"
 #include "llama-memory-hybrid-iswa.h"
@@ -2073,9 +2074,13 @@ llama_memory_i * llama_model::create_memory(const llama_memory_params & params,
             {
                 res = nullptr;
             } break;
-        case LLM_ARCH_DEEPSEEK32:
+        case LLM_ARCH_MINIMAX_M3:
             {
-                res = new llama_kv_cache_dsa(
+                // sparse (MSA) layers carry an indexer key cache, but leading dense layers do not
+                llama_kv_cache::layer_filter_cb filter_idx =
+                    [&](int32_t il) { return (uint32_t) il >= hparams.n_layer_dense_lead; };
+
+                res = new llama_kv_cache_msa(
                         *this,
                         params.type_k,
                         params.type_v,
@@ -2088,9 +2093,11 @@ llama_memory_i * llama_model::create_memory(const llama_memory_params & params,
                         hparams.n_swa,
                         hparams.swa_type,
                         nullptr,
+                        filter_idx,
                         nullptr);
             } break;
         case LLM_ARCH_GLM_DSA:
+        case LLM_ARCH_DEEPSEEK32:
             {
                 if (params.ctx_type == LLAMA_CONTEXT_TYPE_MTP && hparams.n_layer_nextn > 0) {
                     // The NextN/MTP draft head runs dense MLA (no DSA indexer), so the
@@ -2119,10 +2126,11 @@ llama_memory_i * llama_model::create_memory(const llama_memory_params & params,
                 } else {
                     // Main context: DSA cache for the trunk layers only - the nextn
                     // layer(s) are never attended by the trunk graph.
-                    llama_kv_cache::layer_filter_cb filter = nullptr;
+                    llama_kv_cache::layer_filter_cb filter_mla = nullptr;
                     if (hparams.n_layer_nextn > 0) {
-                        filter = [&](uint32_t il) { return il < hparams.n_layer(); };
+                        filter_mla = [&](uint32_t il) { return il < hparams.n_layer(); };
                     }
+                    llama_kv_cache::layer_filter_cb filter_lid = [&](uint32_t il) { return il < hparams.n_layer() && (arch != LLM_ARCH_GLM_DSA || hparams.is_indexer_full(il)); };
 
                     res = new llama_kv_cache_dsa(
                             *this,
@@ -2136,7 +2144,8 @@ llama_memory_i * llama_model::create_memory(const llama_memory_params & params,
                             1,
                             hparams.n_swa,
                             hparams.swa_type,
-                            filter,
+                            filter_mla,
+                            filter_lid,
                             nullptr);
                 }
             } break;
@@ -2213,11 +2222,11 @@ llama_memory_i * llama_model::create_memory(const llama_memory_params & params,
         // checks
         default:
             {
-                // The MTP head is dense-attention only on hybrid Qwen3.5/3.6, so use a plain
+                // The MTP head is dense-attention only on hybrid Qwen3-Next/3.5/3.6, so use a plain
                 // attention KV cache for the MTP context instead of the hybrid wrapper.
-                const bool mtp_on_hybrid_qwen35 =
+                const bool mtp_on_hybrid_qwen =
                     params.ctx_type == LLAMA_CONTEXT_TYPE_MTP &&
-                    (arch == LLM_ARCH_QWEN35 || arch == LLM_ARCH_QWEN35MOE);
+                    (arch == LLM_ARCH_QWEN3NEXT || arch == LLM_ARCH_QWEN35 || arch == LLM_ARCH_QWEN35MOE);
 
                 if (llm_arch_is_recurrent(arch)) {
                     res = new llama_memory_recurrent(
@@ -2229,7 +2238,7 @@ llama_memory_i * llama_model::create_memory(const llama_memory_params & params,
                             cparams.n_seq_max,
                             cparams.n_rs_seq,
                             nullptr);
-                } else if (llm_arch_is_hybrid(arch) && !mtp_on_hybrid_qwen35) {
+                } else if (llm_arch_is_hybrid(arch) && !mtp_on_hybrid_qwen) {
                     // The main difference between hybrid architectures is the
                     // layer filters, so pick the right one here
                     llama_memory_hybrid::layer_filter_cb filter_attn = nullptr;
@@ -2244,7 +2253,7 @@ llama_memory_i * llama_model::create_memory(const llama_memory_params & params,
                         filter_recr = [&](uint32_t il) {
                             return hparams.is_recr(il) && hparams.n_ff(il) == 0;
                         };
-                    } else if (arch == LLM_ARCH_QWEN35 || arch == LLM_ARCH_QWEN35MOE) {
+                    } else if (arch == LLM_ARCH_QWEN3NEXT || arch == LLM_ARCH_QWEN35 || arch == LLM_ARCH_QWEN35MOE) {
                         filter_attn = [&](uint32_t il) {
                             return il < hparams.n_layer() && !hparams.is_recr(il);
                         };
@@ -2310,12 +2319,12 @@ llama_memory_i * llama_model::create_memory(const llama_memory_params & params,
                         };
                     }
 
-                    if (mtp_on_hybrid_qwen35) {
+                    if (mtp_on_hybrid_qwen) {
                         filter = [&](uint32_t il) { return il >= hparams.n_layer(); };
                     }
 
                     if ((arch == LLM_ARCH_STEP35 || arch == LLM_ARCH_HY_V3 || arch == LLM_ARCH_GLM_DSA ||
-                            arch == LLM_ARCH_MIMO2) &&
+                            arch == LLM_ARCH_MIMO2 || arch == LLM_ARCH_DEEPSEEK32) &&
                             hparams.n_layer_nextn > 0) {
                         if (params.ctx_type == LLAMA_CONTEXT_TYPE_MTP) {
                             filter = [&](uint32_t il) { return il >= hparams.n_layer(); };

src/llama-sampler.cpp

@@ -589,6 +589,7 @@ static bool llama_sampler_backend_support(
         /*.probs      = */ nullptr,
         /*.sampled    = */ nullptr,
         /*.candidates = */ ggml_new_tensor_1d(ctx, GGML_TYPE_I32, n),
+        /*.n_vocab    = */ n,
     };
 
     ggml_cgraph * gf = ggml_new_graph(ctx);
@@ -2638,7 +2639,7 @@ struct llama_sampler * llama_sampler_init_grammar_lazy_patterns(
 
 // penalties
 
-struct llama_sampler_penalties {
+struct llama_sampler_penalties : public llama_sampler_backend {
     const int32_t penalty_last_n;
     const float   penalty_repeat;
     const float   penalty_freq;
@@ -2648,10 +2649,49 @@ struct llama_sampler_penalties {
 
     // a frequency map to count token occurrences
     std::unordered_map<llama_token, int> token_count;
+
+    // backend graph inputs
+    ggml_tensor * inp_token_ids = nullptr;
+    ggml_tensor * inp_counts    = nullptr;
+
+    // backend helpers
+    int32_t n_vocab = 0;
+    int32_t n_max   = 0;
+    bool has_candidates = false;
+
+    std::vector<int32_t> host_token_ids;
+    std::vector<int32_t> host_counts;
+
+    static bool is_disabled(
+            int32_t penalty_last_n,
+            float   penalty_repeat,
+            float   penalty_freq,
+            float   penalty_present) {
+        return penalty_last_n == 0 ||
+            (penalty_repeat == 1.0f && penalty_freq == 0.0f && penalty_present == 0.0f);
+    }
+
+    bool is_disabled() const {
+        return is_disabled(penalty_last_n, penalty_repeat, penalty_freq, penalty_present);
+    }
+
+    llama_sampler_penalties(
+            int32_t penalty_last_n,
+            float   penalty_repeat,
+            float   penalty_freq,
+            float   penalty_present)
+        : llama_sampler_backend("penalties")
+        , penalty_last_n  (penalty_last_n)
+        , penalty_repeat  (penalty_repeat)
+        , penalty_freq    (penalty_freq)
+        , penalty_present (penalty_present)
+        , prev            (penalty_last_n) {
+    }
 };
 
-static const char * llama_sampler_penalties_name(const struct llama_sampler * /*smpl*/) {
-    return "penalties";
+static const char * llama_sampler_penalties_name(const struct llama_sampler * smpl) {
+    auto * ctx = (llama_sampler_penalties *) smpl->ctx;
+    return ctx->get_name();
 }
 
 static void llama_sampler_penalties_accept(struct llama_sampler * smpl, llama_token token) {
@@ -2688,8 +2728,7 @@ static void llama_sampler_penalties_accept(struct llama_sampler * smpl, llama_to
 static void llama_sampler_penalties_apply(struct llama_sampler * smpl, llama_token_data_array * cur_p) {
     auto * ctx = (llama_sampler_penalties *) smpl->ctx;
 
-    if ((ctx->penalty_last_n == 0) ||
-        (ctx->penalty_repeat == 1.0f && ctx->penalty_freq == 0.0f && ctx->penalty_present == 0.0f)) {
+    if (ctx->is_disabled()) {
         return;
     }
 
@@ -2736,7 +2775,8 @@ static struct llama_sampler * llama_sampler_penalties_clone(const struct llama_s
     {
         auto * result_ctx = (llama_sampler_penalties *) result->ctx;
 
-        result_ctx->prev = ctx->prev;
+        result_ctx->prev        = ctx->prev;
+        result_ctx->token_count = ctx->token_count;
     }
 
     return result;
@@ -2746,6 +2786,171 @@ static void llama_sampler_penalties_free(struct llama_sampler * smpl) {
     delete (llama_sampler_penalties *) smpl->ctx;
 }
 
+static bool llama_sampler_penalties_backend_init(
+        struct llama_sampler       * smpl,
+        ggml_backend_buffer_type_t   buft) {
+    auto * sctx = (llama_sampler_penalties *) smpl->ctx;
+
+    const bool res = llama_sampler_backend_support(smpl, buft);
+
+    sctx->init(res);
+
+    return res;
+}
+
+static void llama_sampler_penalties_backend_apply(
+        struct llama_sampler      * smpl,
+        struct ggml_context       * ctx,
+        struct ggml_cgraph        * gf,
+        struct llama_sampler_data * data) {
+    GGML_UNUSED(gf);
+
+    auto * sctx = (llama_sampler_penalties *) smpl->ctx;
+
+    if (sctx->is_disabled()) {
+        return;
+    }
+
+    GGML_ASSERT(data->n_vocab > 0 && data->n_vocab <= INT32_MAX);
+
+    sctx->has_candidates = data->candidates != nullptr;
+    sctx->n_vocab = (int32_t) data->n_vocab;
+    sctx->n_max   = std::min(sctx->penalty_last_n, sctx->n_vocab);
+
+    sctx->inp_token_ids = ggml_new_tensor_1d(ctx, GGML_TYPE_I32, sctx->n_max);
+    ggml_set_name(sctx->inp_token_ids, "penalties_token_ids");
+    ggml_set_input(sctx->inp_token_ids);
+
+    sctx->inp_counts = ggml_new_tensor_1d(ctx, GGML_TYPE_I32, sctx->n_max);
+    ggml_set_name(sctx->inp_counts, "penalties_counts");
+    ggml_set_input(sctx->inp_counts);
+
+    if ((int32_t) sctx->host_token_ids.size() != sctx->n_max) {
+        sctx->host_token_ids.assign(sctx->n_max, 0);
+        sctx->host_counts.assign(sctx->n_max, 0);
+    }
+
+    // flatten
+    ggml_tensor * logits = ggml_reshape_1d(ctx, data->logits, ggml_nelements(data->logits));
+    ggml_tensor * gathered = logits;
+    ggml_tensor * counts_f32 = ggml_cast(ctx, sctx->inp_counts, GGML_TYPE_F32);
+
+    if (sctx->has_candidates) {
+        ggml_tensor * candidates = ggml_reshape_1d(
+                ctx, data->candidates, ggml_nelements(data->candidates));
+        const int64_t n_candidates = candidates->ne[0];
+        GGML_ASSERT(n_candidates == ggml_nelements(logits));
+
+        ggml_tensor * counts_rows = ggml_fill(
+                ctx, ggml_new_tensor_2d(ctx, GGML_TYPE_F32, 1, sctx->n_vocab), 0.0f);
+        ggml_tensor * scatter_rows = ggml_reshape_2d(ctx, counts_f32, 1, sctx->n_max);
+        counts_rows = ggml_set_rows(ctx, counts_rows, scatter_rows, sctx->inp_token_ids);
+        counts_f32 = ggml_get_rows(ctx, counts_rows, candidates);
+        counts_f32 = ggml_reshape_1d(ctx, counts_f32, n_candidates);
+    } else {
+        ggml_tensor * logits_rows = ggml_reshape_2d(ctx, logits, 1, ggml_nelements(logits));
+        gathered = ggml_get_rows(ctx, logits_rows, sctx->inp_token_ids);
+        gathered = ggml_reshape_1d(ctx, gathered, sctx->n_max);
+    }
+
+    ggml_tensor * active_mask = ggml_step(ctx, counts_f32);
+    ggml_tensor * inactive_mask = ggml_sub(ctx, ggml_fill(ctx, active_mask, 1.0f), active_mask);
+
+    ggml_tensor * penalized = gathered;
+
+    if (sctx->penalty_repeat != 1.0f) {
+        ggml_tensor * pos_mask = ggml_step(ctx, penalized);
+        ggml_tensor * neg_mask = ggml_sub(ctx, ggml_fill(ctx, pos_mask, 1.0f), pos_mask);
+
+        ggml_tensor * pos_scale = ggml_scale(ctx, pos_mask, 1.0f/sctx->penalty_repeat);
+        ggml_tensor * neg_scale = ggml_scale(ctx, neg_mask, sctx->penalty_repeat);
+        ggml_tensor * repeat_scale = ggml_add(ctx, pos_scale, neg_scale);
+
+        // scale inactive entries with 1 to avoid -INF * 0 = NaN for values masked by top-p
+        repeat_scale = ggml_mul(ctx, repeat_scale, active_mask);
+        repeat_scale = ggml_add(ctx, repeat_scale, inactive_mask);
+        penalized = ggml_mul(ctx, gathered, repeat_scale);
+    }
+
+    if (sctx->penalty_freq != 0.0f) {
+        ggml_tensor * penalty_freq = ggml_scale(ctx, counts_f32, sctx->penalty_freq);
+        penalized = ggml_sub(ctx, penalized, penalty_freq);
+    }
+
+    if (sctx->penalty_present != 0.0f) {
+        ggml_tensor * penalty_present = ggml_scale(ctx, active_mask, sctx->penalty_present);
+        penalized = ggml_sub(ctx, penalized, penalty_present);
+    }
+
+    if (sctx->has_candidates) {
+        data->logits = penalized;
+    } else {
+        ggml_tensor * logits_rows = ggml_reshape_2d(ctx, logits, 1, ggml_nelements(logits));
+        ggml_tensor * scatter_rows = ggml_reshape_2d(ctx, penalized, 1, sctx->n_max);
+        logits_rows = ggml_set_rows(ctx, logits_rows, scatter_rows, sctx->inp_token_ids);
+        data->logits = ggml_reshape_1d(ctx, logits_rows, ggml_nelements(logits));
+    }
+}
+
+static void llama_sampler_penalties_backend_set_input(struct llama_sampler * smpl) {
+    auto * sctx = (llama_sampler_penalties *) smpl->ctx;
+
+    if (!sctx->inp_token_ids || !sctx->inp_counts || sctx->n_max <= 0 || sctx->n_vocab <= 0) {
+        return;
+    }
+
+    if (sctx->is_disabled()) {
+        return;
+    }
+
+    // fill active entries from the map
+    int32_t n_active = 0;
+
+    for (const auto & it : sctx->token_count) {
+        GGML_ASSERT(n_active < sctx->n_max);
+        sctx->host_token_ids[n_active] = it.first;
+        sctx->host_counts   [n_active] = it.second;
+        ++n_active;
+    }
+
+    // Sorting is required because backend_apply uses ggml_set_rows (a scatter-back operation)
+    std::vector<std::pair<int32_t, int32_t>> entries;
+    entries.reserve(n_active);
+    for (int32_t i = 0; i < n_active; ++i) {
+        entries.emplace_back(sctx->host_token_ids[i], sctx->host_counts[i]);
+    }
+    std::sort(entries.begin(), entries.end(), [](const auto & a, const auto & b) {
+        return a.first < b.first;
+    });
+    for (int32_t i = 0; i < n_active; ++i) {
+        sctx->host_token_ids[i] = entries[i].first;
+        sctx->host_counts   [i] = entries[i].second;
+    }
+
+    // Padding: Finds a filler token id that is not present in token_count.
+    // Use it to do padding for the arrays, it avoids resizing every time.
+    // The arrays must always have exactly n_max entries (the GPU tensor is a fixed size).
+    int32_t filler = 0;
+    if (n_active < sctx->n_max) {
+        while (sctx->token_count.find(filler) != sctx->token_count.end()) {
+            ++filler;
+        }
+        GGML_ASSERT(filler < sctx->n_vocab);
+    }
+
+    // Fill the rest of the arrays with the filler token id and count 0.
+    // Inactive slots are padded with a unique dummy token ID (count = 0).
+    // The uniqueness matters because ggml_set_rows with duplicate indices can produce non-deterministic or incorrect results.
+    // Using a filler token with count 0 that isn't in the active set is safe, because the active_mask step in backend_apply filters them out via ggml_step(counts_f32)
+    for (int32_t i = n_active; i < sctx->n_max; ++i) {
+        sctx->host_token_ids[i] = filler;
+        sctx->host_counts   [i] = 0;
+    }
+
+    ggml_backend_tensor_set(sctx->inp_token_ids, sctx->host_token_ids.data(), 0, sctx->n_max * sizeof(int32_t));
+    ggml_backend_tensor_set(sctx->inp_counts,    sctx->host_counts.data(),    0, sctx->n_max * sizeof(int32_t));
+}
+
 static struct llama_sampler_i llama_sampler_penalties_i = {
     /* .name              = */ llama_sampler_penalties_name,
     /* .accept            = */ llama_sampler_penalties_accept,
@@ -2753,10 +2958,10 @@ static struct llama_sampler_i llama_sampler_penalties_i = {
     /* .reset             = */ llama_sampler_penalties_reset,
     /* .clone             = */ llama_sampler_penalties_clone,
     /* .free              = */ llama_sampler_penalties_free,
-    /* .backend_init      = */ nullptr,
+    /* .backend_init      = */ llama_sampler_penalties_backend_init,
     /* .backend_accept    = */ nullptr,
-    /* .backend_apply     = */ nullptr,
-    /* .backend_set_input = */ nullptr,
+    /* .backend_apply     = */ llama_sampler_penalties_backend_apply,
+    /* .backend_set_input = */ llama_sampler_penalties_backend_set_input,
 };
 
 struct llama_sampler * llama_sampler_init_penalties(
@@ -2766,22 +2971,18 @@ struct llama_sampler * llama_sampler_init_penalties(
         float penalty_present) {
     penalty_last_n = std::max(penalty_last_n, 0);
 
-    const bool is_empty = (penalty_last_n == 0 || (penalty_repeat == 1.0f && penalty_freq == 0.0f && penalty_present == 0.0f));
-
-    if (is_empty) {
+    if (llama_sampler_penalties::is_disabled(
+                penalty_last_n, penalty_repeat, penalty_freq, penalty_present)) {
         return llama_sampler_init_empty("?penalties");
     }
 
     return llama_sampler_init(
         /* .iface = */ &llama_sampler_penalties_i,
-        /* .ctx   = */ new llama_sampler_penalties {
-            /* .penalty_last_n  = */ penalty_last_n,
-            /* .penalty_repeat  = */ penalty_repeat,
-            /* .penalty_freq    = */ penalty_freq,
-            /* .penalty_present = */ penalty_present,
-            /* .prev            = */ ring_buffer<llama_token>(penalty_last_n),
-            /* .token_count     = */ {},
-        }
+        /* .ctx   = */ new llama_sampler_penalties(
+            penalty_last_n,
+            penalty_repeat,
+            penalty_freq,
+            penalty_present)
     );
 }
 

src/llama-vocab.cpp

@@ -2532,6 +2532,12 @@ void llama_vocab::impl::load(llama_model_loader & ml, const LLM_KV & kv) {
             const std::string & key = kv(std::get<0>(it));
             int32_t & id = std::get<1>(it);
 
+            if (id >= 0 && static_cast<size_t>(id) >= id_to_token.size()) {
+                LLAMA_LOG_WARN("%s: default special token '%s' = %d out of vocab range, disabling\n",
+                    __func__, key.c_str(), id);
+                id = LLAMA_TOKEN_NULL;
+            }
+
             uint32_t new_id;
             if (!ml.get_key(std::get<0>(it), new_id, false)) {
                 continue;

src/models/deepseek2.cpp

@@ -37,6 +37,11 @@ void llama_model_deepseek2::load_arch_hparams(llama_model_loader & ml) {
         hparams.rope_yarn_log_mul /= 0.1f;
     }
 
+    // NextN/MTP
+    ml.get_key(LLM_KV_NEXTN_PREDICT_LAYERS, hparams.n_layer_nextn, false);
+    GGML_ASSERT(hparams.n_layer_nextn == 0 ||
+        hparams.n_layer() + hparams.n_layer_nextn == hparams.n_layer_all);
+
     // (optional) temperature tuning - used by mistral-large
     ml.get_key(LLM_KV_ATTENTION_TEMPERATURE_SCALE,  hparams.f_attn_temp_scale,       false);
     ml.get_key(LLM_KV_ATTENTION_TEMPERATURE_LENGTH, hparams.n_attn_temp_floor_scale, false); // FIXME why not use temperature_length?
@@ -52,10 +57,20 @@ void llama_model_deepseek2::load_arch_hparams(llama_model_loader & ml) {
     }
 }
 
-void llama_model_deepseek2::load_arch_tensors(llama_model_loader &) {
+void llama_model_deepseek2::load_arch_tensors(llama_model_loader & ml) {
     LLAMA_LOAD_LOCALS;
     const int64_t n_expert_shared = hparams.n_expert_shared;
 
+    const bool mtp_only = (hparams.n_layer_nextn > 0) && (ml.get_weight("blk.0.attn_norm.weight") == nullptr);
+    const std::string mtp_probe = "blk." + std::to_string(n_layer) + ".nextn.eh_proj.weight";
+    const bool trunk_only = (hparams.n_layer_nextn > 0) && (ml.get_weight(mtp_probe.c_str()) == nullptr);
+    const int trunk_flags = mtp_only  ? TENSOR_NOT_REQUIRED : 0;
+    int       mtp_flags   = trunk_only ? TENSOR_NOT_REQUIRED : 0;
+
+    if (!ml.load_mtp) {
+        mtp_flags |= TENSOR_SKIP;
+    }
+
     const bool is_mla = hparams.is_mla();
 
     // note: these are the actual head sizes you get when treating as MHA or after "decompression" using wv_b for MLA
@@ -81,44 +96,45 @@ void llama_model_deepseek2::load_arch_tensors(llama_model_loader &) {
         output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED);
     }
 
-    for (int i = 0; i < n_layer; ++i) {
+    for (int i = 0; i < n_layer_all; ++i) {
         auto & layer = layers[i];
+        const int flags = i < n_layer ? trunk_flags : mtp_flags;
 
-        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
+        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, flags);
         if (q_lora_rank > 0) {
-            layer.attn_q_a_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_A_NORM, "weight", i), {q_lora_rank}, 0);
+            layer.attn_q_a_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_A_NORM, "weight", i), {q_lora_rank}, flags);
         }
 
-        layer.attn_kv_a_norm = create_tensor(tn(LLM_TENSOR_ATTN_KV_A_NORM, "weight", i), {kv_lora_rank}, 0);
+        layer.attn_kv_a_norm = create_tensor(tn(LLM_TENSOR_ATTN_KV_A_NORM, "weight", i), {kv_lora_rank}, flags);
 
         if (q_lora_rank > 0) {
-            layer.wq_a = create_tensor(tn(LLM_TENSOR_ATTN_Q_A, "weight", i), {n_embd, q_lora_rank}, 0);
-            layer.wq_b = create_tensor(tn(LLM_TENSOR_ATTN_Q_B, "weight", i), {q_lora_rank, n_head * n_embd_head_k_mla}, 0);
+            layer.wq_a = create_tensor(tn(LLM_TENSOR_ATTN_Q_A, "weight", i), {n_embd, q_lora_rank}, flags);
+            layer.wq_b = create_tensor(tn(LLM_TENSOR_ATTN_Q_B, "weight", i), {q_lora_rank, n_head * n_embd_head_k_mla}, flags);
         } else {
-            layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_head * n_embd_head_k_mla}, 0);
+            layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_head * n_embd_head_k_mla}, flags);
         }
 
-        layer.wkv_a_mqa = create_tensor(tn(LLM_TENSOR_ATTN_KV_A_MQA, "weight", i), {n_embd, kv_lora_rank + n_embd_head_qk_rope}, 0);
+        layer.wkv_a_mqa = create_tensor(tn(LLM_TENSOR_ATTN_KV_A_MQA, "weight", i), {n_embd, kv_lora_rank + n_embd_head_qk_rope}, flags);
 
         // note: only old legacy GGUF files will have the unsplit wkv_b tensor in
         if (is_mla) {
-            layer.wk_b = create_tensor(tn(LLM_TENSOR_ATTN_K_B, "weight", i), {n_embd_head_qk_nope, kv_lora_rank, n_head}, 0);
-            layer.wv_b = create_tensor(tn(LLM_TENSOR_ATTN_V_B, "weight", i), {kv_lora_rank, n_embd_head_v_mla, n_head}, 0);
+            layer.wk_b = create_tensor(tn(LLM_TENSOR_ATTN_K_B, "weight", i), {n_embd_head_qk_nope, kv_lora_rank, n_head}, flags);
+            layer.wv_b = create_tensor(tn(LLM_TENSOR_ATTN_V_B, "weight", i), {kv_lora_rank, n_embd_head_v_mla, n_head}, flags);
         } else {
-            layer.wkv_b = create_tensor(tn(LLM_TENSOR_ATTN_KV_B, "weight", i), {kv_lora_rank, n_head * (n_embd_head_qk_nope + n_embd_head_v_mla)}, 0);
+            layer.wkv_b = create_tensor(tn(LLM_TENSOR_ATTN_KV_B, "weight", i), {kv_lora_rank, n_head * (n_embd_head_qk_nope + n_embd_head_v_mla)}, flags);
         }
 
-        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_head * n_embd_head_v_mla, n_embd}, 0);
+        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_head * n_embd_head_v_mla, n_embd}, flags);
 
-        layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
+        layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, flags);
 
         if (i < (int) hparams.n_layer_dense_lead) {
-            layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd,   n_ff}, 0);
-            layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
-            layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
+            layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd,   n_ff}, flags);
+            layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {  n_ff, n_embd}, flags);
+            layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff}, flags);
         } else {
-            layer.ffn_gate_inp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP, "weight", i), {n_embd, n_expert}, 0);
-            layer.ffn_exp_probs_b = create_tensor(tn(LLM_TENSOR_FFN_EXP_PROBS_B, "bias", i), {n_expert}, TENSOR_NOT_REQUIRED);
+            layer.ffn_gate_inp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP, "weight", i), {n_embd, n_expert}, flags);
+            layer.ffn_exp_probs_b = create_tensor(tn(LLM_TENSOR_FFN_EXP_PROBS_B, "bias", i), {n_expert}, TENSOR_NOT_REQUIRED | flags);
 
             if (n_expert == 0) {
                 throw std::runtime_error("n_expert must be > 0");
@@ -128,21 +144,281 @@ void llama_model_deepseek2::load_arch_tensors(llama_model_loader &) {
             }
 
             // MoE branch
-            layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), {n_ff_exp,   n_embd, n_expert}, 0);
-            create_tensor_gate_up_exps(layer, i, n_embd, n_ff_exp, n_expert, 0);
+            layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), {n_ff_exp,   n_embd, n_expert}, flags);
+            create_tensor_gate_up_exps(layer, i, n_embd, n_ff_exp, n_expert, flags);
 
             // Shared expert branch
-            layer.ffn_gate_shexp = create_tensor(tn(LLM_TENSOR_FFN_GATE_SHEXP, "weight", i), {n_embd, n_ff_exp * n_expert_shared}, 0);
-            layer.ffn_down_shexp = create_tensor(tn(LLM_TENSOR_FFN_DOWN_SHEXP, "weight", i), {        n_ff_exp * n_expert_shared, n_embd}, 0);
-            layer.ffn_up_shexp   = create_tensor(tn(LLM_TENSOR_FFN_UP_SHEXP,   "weight", i), {n_embd, n_ff_exp * n_expert_shared}, 0);
+            layer.ffn_gate_shexp = create_tensor(tn(LLM_TENSOR_FFN_GATE_SHEXP, "weight", i), {n_embd, n_ff_exp * n_expert_shared}, flags);
+            layer.ffn_down_shexp = create_tensor(tn(LLM_TENSOR_FFN_DOWN_SHEXP, "weight", i), {        n_ff_exp * n_expert_shared, n_embd}, flags);
+            layer.ffn_up_shexp   = create_tensor(tn(LLM_TENSOR_FFN_UP_SHEXP,   "weight", i), {n_embd, n_ff_exp * n_expert_shared}, flags);
+        }
+
+        // NextN/MTP tensors
+        if (i >= n_layer) {
+            layer.nextn.eh_proj          = create_tensor(tn(LLM_TENSOR_NEXTN_EH_PROJ, "weight", i), { 2 * n_embd, n_embd }, mtp_flags);
+            layer.nextn.enorm            = create_tensor(tn(LLM_TENSOR_NEXTN_ENORM, "weight", i), { n_embd }, mtp_flags);
+            layer.nextn.hnorm            = create_tensor(tn(LLM_TENSOR_NEXTN_HNORM, "weight", i), { n_embd }, mtp_flags);
+            layer.nextn.embed_tokens     = create_tensor(tn(LLM_TENSOR_NEXTN_EMBED_TOKENS, "weight", i), { n_embd, n_vocab }, TENSOR_NOT_REQUIRED | flags);
+            layer.nextn.shared_head_head = create_tensor(tn(LLM_TENSOR_NEXTN_SHARED_HEAD_HEAD, "weight", i), { n_embd, n_vocab }, TENSOR_NOT_REQUIRED | flags);
+            layer.nextn.shared_head_norm = create_tensor(tn(LLM_TENSOR_NEXTN_SHARED_HEAD_NORM, "weight", i), { n_embd }, TENSOR_NOT_REQUIRED | flags);
         }
     }
 }
 
 std::unique_ptr<llm_graph_context> llama_model_deepseek2::build_arch_graph(const llm_graph_params & params) const {
+    if (params.gtype == LLM_GRAPH_TYPE_DECODER_MTP) {
+        return std::make_unique<graph_mtp>(*this, params);
+    }
     return std::make_unique<graph>(*this, params);
 }
 
+llama_model_deepseek2::graph_mtp::graph_mtp(const llama_model & model, const llm_graph_params & params) :
+    llm_graph_context(params) {
+    GGML_ASSERT(hparams.n_layer_nextn > 0 && "GLM4 MTP requires n_layer_nextn > 0");
+    GGML_ASSERT(hparams.n_layer_nextn == 1 && "GLM4 MTP currently only supports a single MTP block");
+    GGML_ASSERT(hparams.is_mla() && "GLM4 MTP requires MLA");
+    GGML_ASSERT(hparams.f_attn_temp_scale == 0.0f && "GLM4 MTP does not support attention temperature scaling");
+
+    // The appended MTP block is stored immediately after the main decoder layers.
+    const int il = hparams.n_layer();
+    const auto & layer = model.layers[il];
+
+    GGML_ASSERT(layer.nextn.eh_proj && "MTP block missing nextn.eh_proj");
+    GGML_ASSERT(layer.nextn.enorm   && "MTP block missing nextn.enorm");
+    GGML_ASSERT(layer.nextn.hnorm   && "MTP block missing nextn.hnorm");
+
+    GGML_ASSERT((uint32_t) il >= hparams.n_layer_dense_lead && "GLM4 MTP block expected to use MoE FFN");
+
+    const int64_t n_embd_head_k_mla   = hparams.n_embd_head_k_mla();
+    const int64_t n_embd_head_qk_rope = hparams.n_rot();
+    const int64_t n_embd_head_qk_nope = n_embd_head_k_mla - n_embd_head_qk_rope;
+    const int64_t kv_lora_rank        = hparams.n_lora_kv;
+
+    GGML_ASSERT(n_embd_head_qk_nope >= 1);
+    GGML_ASSERT(hparams.n_lora_q > 0);
+    GGML_ASSERT(layer.wq_a);
+    GGML_ASSERT(layer.attn_q_a_norm);
+    GGML_ASSERT(layer.wq_b);
+    GGML_ASSERT(layer.wkv_a_mqa);
+    GGML_ASSERT(layer.attn_kv_a_norm);
+    GGML_ASSERT(layer.wk_b);
+
+    const bool has_split_exps =
+            layer.ffn_up_exps   != nullptr &&
+            layer.ffn_gate_exps != nullptr;
+
+    const bool has_fused_exps = layer.ffn_gate_up_exps != nullptr;
+
+    GGML_ASSERT(has_split_exps || has_fused_exps);
+    GGML_ASSERT(layer.ffn_norm);
+    GGML_ASSERT(layer.ffn_gate_inp);
+    GGML_ASSERT(layer.ffn_down_exps);
+    GGML_ASSERT(layer.ffn_gate_shexp);
+    GGML_ASSERT(layer.ffn_down_shexp);
+    GGML_ASSERT(layer.ffn_up_shexp);
+
+    auto inp = std::make_unique<llm_graph_input_embd_h>(hparams.n_embd);
+
+    inp->tokens = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, n_tokens);
+    ggml_set_input(inp->tokens);
+
+    inp->embd = ggml_new_tensor_2d(ctx0, GGML_TYPE_F32, hparams.n_embd_inp(), n_tokens);
+    ggml_set_input(inp->embd);
+
+    ggml_tensor * tok_embd;
+    if (ubatch.token) {
+        ggml_tensor * tok_embd_w = layer.nextn.embed_tokens
+                ? layer.nextn.embed_tokens
+                : model.tok_embd;
+
+        tok_embd = ggml_get_rows(ctx0, tok_embd_w, inp->tokens);
+    } else {
+        tok_embd = inp->embd;
+    }
+    cb(tok_embd, "mtp_tok_embd", il);
+
+    inp->h = ggml_new_tensor_2d(ctx0, GGML_TYPE_F32, hparams.n_embd, n_tokens);
+    ggml_set_input(inp->h);
+    ggml_set_name(inp->h, "mtp_h_input");
+
+    ggml_tensor * h_embd = inp->h;
+
+    res->add_input(std::move(inp));
+
+    ggml_tensor * inp_pos     = build_inp_pos();
+    ggml_tensor * inp_out_ids = build_inp_out_ids();
+
+    auto * inp_attn_k = build_attn_inp_k();
+
+    ggml_tensor * h_norm = build_norm(h_embd, layer.nextn.hnorm, nullptr, LLM_NORM_RMS, il);
+    cb(h_norm, "mtp_hnorm", il);
+
+    ggml_tensor * e_norm = build_norm(tok_embd, layer.nextn.enorm, nullptr, LLM_NORM_RMS, il);
+    cb(e_norm, "mtp_enorm", il);
+
+    ggml_tensor * concat = ggml_concat(ctx0, e_norm, h_norm, 0);
+    cb(concat, "mtp_concat", il);
+
+    ggml_tensor * cur = build_lora_mm(layer.nextn.eh_proj, concat, layer.nextn.eh_proj_s);
+    cb(cur, "mtp_eh_proj", il);
+
+    ggml_tensor * inpSA = cur;
+
+    cur = build_norm(cur, layer.attn_norm, nullptr, LLM_NORM_RMS, il);
+    cb(cur, "mtp_attn_norm", il);
+
+    ggml_tensor * q = ggml_mul_mat(ctx0, layer.wq_a, cur);
+    cb(q, "mtp_q_a", il);
+
+    q = build_norm(q, layer.attn_q_a_norm, nullptr, LLM_NORM_RMS, il);
+    cb(q, "mtp_q_a_norm", il);
+
+    q = ggml_mul_mat(ctx0, layer.wq_b, q);
+    cb(q, "mtp_q_b", il);
+
+    ggml_tensor * q_nope =
+        ggml_view_3d(ctx0, q, n_embd_head_qk_nope, n_head, n_tokens,
+                ggml_row_size(q->type, n_embd_head_k_mla),
+                ggml_row_size(q->type, n_embd_head_k_mla) * n_head, 0);
+    cb(q_nope, "mtp_q_nope", il);
+
+    ggml_tensor * q_pe =
+        ggml_view_3d(ctx0, q, n_embd_head_qk_rope, n_head, n_tokens,
+                ggml_row_size(q->type, n_embd_head_k_mla),
+                ggml_row_size(q->type, n_embd_head_k_mla) * n_head,
+                ggml_row_size(q->type, n_embd_head_qk_nope));
+    cb(q_pe, "mtp_q_pe", il);
+
+    ggml_tensor * kv_cmpr_pe = ggml_mul_mat(ctx0, layer.wkv_a_mqa, cur);
+    cb(kv_cmpr_pe, "mtp_kv_cmpr_pe", il);
+
+    ggml_tensor * kv_cmpr =
+        ggml_view_2d(ctx0, kv_cmpr_pe, kv_lora_rank, n_tokens,
+                ggml_row_size(kv_cmpr_pe->type, kv_lora_rank + n_embd_head_qk_rope), 0);
+    cb(kv_cmpr, "mtp_kv_cmpr", il);
+
+    ggml_tensor * k_pe =
+        ggml_view_3d(ctx0, kv_cmpr_pe, n_embd_head_qk_rope, 1, n_tokens,
+                ggml_row_size(kv_cmpr_pe->type, kv_lora_rank + n_embd_head_qk_rope),
+                ggml_row_size(kv_cmpr_pe->type, kv_lora_rank + n_embd_head_qk_rope),
+                ggml_row_size(kv_cmpr_pe->type, kv_lora_rank));
+    cb(k_pe, "mtp_k_pe", il);
+
+    kv_cmpr = build_norm(kv_cmpr, layer.attn_kv_a_norm, nullptr, LLM_NORM_RMS, il);
+    cb(kv_cmpr, "mtp_kv_cmpr_norm", il);
+
+    GGML_ASSERT(ext_factor >= 0.0f);
+
+    const float attn_factor_org =
+            attn_factor * (1.0f + 0.1f * logf(1.0f / freq_scale));
+
+    const float mscale =
+            attn_factor_org * (1.0f + 0.1f * hparams.rope_yarn_log_mul * logf(1.0f / freq_scale));
+
+    const float kq_scale =
+            1.0f * mscale * mscale / sqrtf(float(n_embd_head_k_mla));
+
+    q_pe = ggml_rope_ext(ctx0, q_pe, inp_pos, nullptr,
+            n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
+            ext_factor, attn_factor, beta_fast, beta_slow);
+    cb(q_pe, "mtp_q_pe_rope", il);
+
+    k_pe = ggml_rope_ext(ctx0, k_pe, inp_pos, nullptr,
+            n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
+            ext_factor, attn_factor, beta_fast, beta_slow);
+    cb(k_pe, "mtp_k_pe_rope", il);
+
+    q_nope = ggml_permute(ctx0, q_nope, 0, 2, 1, 3);
+    cb(q_nope, "mtp_q_nope_perm", il);
+
+    ggml_tensor * q_nope_absorbed = ggml_mul_mat(ctx0, layer.wk_b, q_nope);
+    cb(q_nope_absorbed, "mtp_q_nope_absorbed", il);
+
+    q_nope_absorbed = ggml_permute(ctx0, q_nope_absorbed, 0, 2, 1, 3);
+    cb(q_nope_absorbed, "mtp_q_nope_absorbed_perm", il);
+
+    ggml_tensor * Qcur = ggml_concat(ctx0, q_nope_absorbed, q_pe, 0);
+    cb(Qcur, "mtp_Qcur", il);
+
+    kv_cmpr = ggml_reshape_3d(ctx0, kv_cmpr, hparams.n_lora_kv, 1, n_tokens);
+    cb(kv_cmpr, "mtp_kv_cmpr_reshape", il);
+
+    ggml_tensor * Kcur = ggml_concat(ctx0, kv_cmpr, k_pe, 0);
+    cb(Kcur, "mtp_Kcur", il);
+
+    ggml_tensor * Vcur = kv_cmpr;
+    cb(Vcur, "mtp_Vcur", il);
+
+    cur = build_attn(inp_attn_k,
+            layer.wo, nullptr, layer.wo_s,
+            Qcur, Kcur, Vcur, nullptr, nullptr, layer.wv_b, kq_scale, il);
+    cb(cur, "mtp_attn_out", il);
+
+    ggml_tensor * ffn_inp = ggml_add(ctx0, cur, inpSA);
+    cb(ffn_inp, "mtp_ffn_inp", il);
+
+    cur = build_norm(ffn_inp, layer.ffn_norm, nullptr, LLM_NORM_RMS, il);
+    cb(cur, "mtp_ffn_norm", il);
+
+    ggml_tensor * moe_out = build_moe_ffn(cur,
+            layer.ffn_gate_inp,
+            layer.ffn_up_exps,
+            layer.ffn_gate_exps,
+            layer.ffn_down_exps,
+            layer.ffn_exp_probs_b,
+            n_expert, n_expert_used,
+            LLM_FFN_SILU, hparams.expert_weights_norm,
+            hparams.expert_weights_scale,
+            (llama_expert_gating_func_type) hparams.expert_gating_func,
+            il,
+            nullptr,
+            layer.ffn_gate_up_exps);
+    cb(moe_out, "mtp_ffn_moe_out", il);
+
+    ggml_tensor * ffn_shexp = build_ffn(cur,
+            layer.ffn_up_shexp, nullptr, nullptr,
+            layer.ffn_gate_shexp, nullptr, nullptr,
+            layer.ffn_down_shexp, nullptr, nullptr,
+            nullptr, LLM_FFN_SILU, LLM_FFN_PAR, il);
+    cb(ffn_shexp, "mtp_ffn_shexp", il);
+
+    cur = ggml_add(ctx0, moe_out, ffn_shexp);
+    cb(cur, "mtp_ffn_out", il);
+
+    cur = ggml_add(ctx0, cur, ffn_inp);
+    cb(cur, "mtp_post_ffn", il);
+
+    ggml_tensor * head_norm_w = layer.nextn.shared_head_norm
+            ? layer.nextn.shared_head_norm
+            : model.output_norm;
+    GGML_ASSERT(head_norm_w && "GLM4 MTP: missing both nextn.shared_head_norm and output_norm");
+
+    cur = build_norm(cur, head_norm_w, nullptr, LLM_NORM_RMS, -1);
+    cb(cur, "h_nextn", -1);
+    res->t_h_nextn = cur;
+
+    if (inp_out_ids) {
+        cur = ggml_get_rows(ctx0, cur, inp_out_ids);
+    }
+    cb(cur, "mtp_shared_head_norm", -1);
+
+    ggml_tensor * head_w = layer.nextn.shared_head_head
+            ? layer.nextn.shared_head_head
+            : model.output;
+
+    ggml_tensor * head_s = layer.nextn.shared_head_head
+            ? layer.nextn.shared_head_head_s
+            : model.output_s;
+
+    GGML_ASSERT(head_w && "GLM4 MTP: missing LM head (nextn.shared_head_head or model.output)");
+
+    cur = build_lora_mm(head_w, cur, head_s);
+    cb(cur, "result_output", -1);
+
+    res->t_logits = cur;
+    ggml_build_forward_expand(gf, cur);
+}
+
 llama_model_deepseek2::graph::graph(const llama_model & model, const llm_graph_params & params) :
     llm_graph_context(params) {
     // lite variants include DeepSeek-V2-Lite, GigaChat3-10B-A1.8B
@@ -365,7 +641,7 @@ llama_model_deepseek2::graph::graph(const llama_model & model, const llm_graph_p
                             Qcur, Kcur, Vcur, nullptr, nullptr, nullptr, kq_scale, il);
             }
         }
-        if (il == n_layer - 1 && inp_out_ids) {
+        if (il == n_layer - 1 && inp_out_ids && (!cparams.embeddings_nextn || cparams.embeddings_nextn_masked)) {
             cur   = ggml_get_rows(ctx0, cur, inp_out_ids);
             inpSA = ggml_get_rows(ctx0, inpSA, inp_out_ids);
         }
@@ -425,6 +701,13 @@ llama_model_deepseek2::graph::graph(const llama_model & model, const llm_graph_p
 
     cur = build_norm(cur, model.output_norm, NULL, LLM_NORM_RMS, -1);
 
+    cb(cur, "h_nextn", -1);
+    res->t_h_nextn = cur;
+
+    if (cparams.embeddings_nextn && !cparams.embeddings_nextn_masked && inp_out_ids) {
+        cur = ggml_get_rows(ctx0, cur, inp_out_ids);
+    }
+
     cb(cur, "result_norm", -1);
     res->t_embd = cur;
 

src/models/deepseek32.cpp

@@ -44,13 +44,24 @@ void llama_model_deepseek32::load_arch_hparams(llama_model_loader & ml) {
     GGML_ASSERT(hparams.n_layer_nextn < hparams.n_layer_all && "n_layer_nextn must be < n_layer");
 
     switch (hparams.n_layer()) {
-        case 62: type = LLM_TYPE_685B_A37B; break;
+        case 61: type = LLM_TYPE_685B_A37B; break;
         default: type = LLM_TYPE_UNKNOWN;
     }
 }
 
-void llama_model_deepseek32::load_arch_tensors(llama_model_loader &) {
+void llama_model_deepseek32::load_arch_tensors(llama_model_loader & ml) {
     LLAMA_LOAD_LOCALS;
+
+    const bool mtp_only = (hparams.n_layer_nextn > 0) && (ml.get_weight("blk.0.attn_norm.weight") == nullptr);
+    const std::string mtp_probe = "blk." + std::to_string(n_layer) + ".nextn.eh_proj.weight";
+    const bool trunk_only = (hparams.n_layer_nextn > 0) && (ml.get_weight(mtp_probe.c_str()) == nullptr);
+    const int trunk_flags = mtp_only   ? TENSOR_NOT_REQUIRED : 0;
+    int       mtp_flags   = trunk_only ? TENSOR_NOT_REQUIRED : 0;
+
+    if (!ml.load_mtp) {
+        mtp_flags |= TENSOR_SKIP;
+    }
+
     const bool is_mla = hparams.is_mla();
     if (!is_mla) {
         throw std::runtime_error("DEEPSEEK32 architecture requires MLA");
@@ -80,12 +91,7 @@ void llama_model_deepseek32::load_arch_tensors(llama_model_loader &) {
     }
 
     for (int i = 0; i < n_layer_all; ++i) {
-        int flags = 0;
-        if (i >= n_layer) {
-            // skip all tensors in the NextN layers
-            // TODO @ngxson : TENSOR_NOT_REQUIRED was a hack, need to remove it later
-            flags |= TENSOR_SKIP | TENSOR_NOT_REQUIRED;
-        }
+        const int flags = (i >= n_layer) ? mtp_flags : trunk_flags;
 
         auto & layer = layers[i];
 
@@ -138,7 +144,7 @@ void llama_model_deepseek32::load_arch_tensors(llama_model_loader &) {
             layer.ffn_up_shexp   = create_tensor(tn(LLM_TENSOR_FFN_UP_SHEXP,   "weight", i), {n_embd, n_ff_exp * n_expert_shared}, flags);
         }
 
-        // NextN/MTP tensors (preserved but unused) - conditionally load for last nextn_predict_layers
+        // NextN/MTP tensors - conditionally load for last nextn_predict_layers
         if (i >= n_layer) {
             layer.nextn.eh_proj          = create_tensor(tn(LLM_TENSOR_NEXTN_EH_PROJ, "weight", i), { 2 * n_embd, n_embd }, flags);
             layer.nextn.enorm            = create_tensor(tn(LLM_TENSOR_NEXTN_ENORM, "weight", i), { n_embd }, flags);
@@ -153,6 +159,9 @@ void llama_model_deepseek32::load_arch_tensors(llama_model_loader &) {
 }
 
 std::unique_ptr<llm_graph_context> llama_model_deepseek32::build_arch_graph(const llm_graph_params & params) const {
+    if (params.gtype == LLM_GRAPH_TYPE_DECODER_MTP) {
+        return std::make_unique<graph_mtp>(*this, params);
+    }
     return std::make_unique<graph>(*this, params);
 }
 
@@ -430,7 +439,9 @@ llama_model_deepseek32::graph::graph(const llama_model & model, const llm_graph_
                         Qcur, Kcur, Vcur, nullptr, nullptr, model.layers[il].wv_b, top_k, kq_scale, il);
             }
         }
-        if (il == n_layer - 1 && inp_out_ids) {
+        // when unmasked nextn embeddings are requested, t_h_nextn must keep all rows,
+        // so the early output masking has to be skipped (it is applied after the final norm instead)
+        if (il == n_layer - 1 && inp_out_ids && (!cparams.embeddings_nextn || cparams.embeddings_nextn_masked)) {
             cur   = ggml_get_rows(ctx0, cur, inp_out_ids);
             inpSA = ggml_get_rows(ctx0, inpSA, inp_out_ids);
         }
@@ -493,6 +504,14 @@ llama_model_deepseek32::graph::graph(const llama_model & model, const llm_graph_
 
     cur = build_norm(cur, model.output_norm, NULL, LLM_NORM_RMS, -1);
 
+    // post-norm hidden state feeds the NextN/MTP draft head
+    cb(cur, "h_nextn", -1);
+    res->t_h_nextn = cur;
+
+    if (cparams.embeddings_nextn && !cparams.embeddings_nextn_masked && inp_out_ids) {
+        cur = ggml_get_rows(ctx0, cur, inp_out_ids);
+    }
+
     cb(cur, "result_norm", -1);
     res->t_embd = cur;
 
@@ -504,3 +523,243 @@ llama_model_deepseek32::graph::graph(const llama_model & model, const llm_graph_
 
     ggml_build_forward_expand(gf, cur);
 }
+
+// LLM_GRAPH_TYPE_DECODER_MTP draft head for DeepSeek V3.2 (DEEPSEEK32).
+// Semantics mirror the deepseek-family NextN/MTP layer:
+//   enorm(embed) + hnorm(prev_hidden) -> concat(e, h) -> eh_proj ->
+//   full deepseek32 decoder block (dense MLA attention + sigmoid-gated MoE FFN
+//   with shared expert, exactly as the trunk deepseek2 graph builds it) ->
+//   shared_head_norm (fallback output_norm) -> shared LM head.
+// The DSA indexer is not used at runtime.
+llama_model_deepseek32::graph_mtp::graph_mtp(const llama_model & model, const llm_graph_params & params)
+    : llm_graph_context(params) {
+    GGML_ASSERT(hparams.n_layer_nextn > 0 && "DEEPSEEK32 MTP requires n_layer_nextn > 0");
+    GGML_ASSERT(hparams.n_layer_nextn == 1 && "DEEPSEEK32 MTP currently only supports a single MTP block");
+    GGML_ASSERT(hparams.is_mla() && "DEEPSEEK32 MTP requires MLA");
+
+    const int il = hparams.n_layer() + cparams.nextn_layer_offset;
+    GGML_ASSERT(cparams.nextn_layer_offset >= 0 &&
+                cparams.nextn_layer_offset < (int) hparams.n_layer_nextn &&
+                "nextn_layer_offset out of range [0, n_layer_nextn)");
+    const auto & layer = model.layers[il];
+
+    GGML_ASSERT(layer.nextn.eh_proj && "MTP block missing nextn.eh_proj");
+    GGML_ASSERT(layer.nextn.enorm   && "MTP block missing nextn.enorm");
+    GGML_ASSERT(layer.nextn.hnorm   && "MTP block missing nextn.hnorm");
+    GGML_ASSERT(layer.ffn_gate_inp  && "MTP block missing ffn_gate_inp");
+
+    // note: these are the actual head sizes you get when treating as MHA or after "decompression" using wv_b for MLA
+    const int64_t n_embd_head_k = hparams.n_embd_head_k_mla();
+
+    const int64_t n_embd_head_qk_rope = hparams.n_rot();
+    const int64_t n_embd_head_qk_nope = n_embd_head_k - n_embd_head_qk_rope;
+
+    const uint32_t kv_lora_rank = hparams.n_lora_kv;
+
+    // We have to pre-scale kq_scale and attn_factor to make the YaRN RoPE work correctly.
+    // See the deepseek2 trunk graph for the detailed explanation - this must match it EXACTLY.
+    GGML_ASSERT(ext_factor >= 0.0f);
+    const float attn_factor_org = attn_factor * (1.0f + 0.1f * logf(1.0f / freq_scale));
+
+    const float mscale   = attn_factor_org * (1.0f + 0.1f * hparams.rope_yarn_log_mul * logf(1.0f / freq_scale));
+    const float kq_scale = 1.0f * mscale * mscale / sqrtf(float(n_embd_head_k));
+
+    // TODO: extract in a common llm_graph_context::build_inp_embd_h()
+    auto inp = std::make_unique<llm_graph_input_embd_h>(hparams.n_embd);
+
+    inp->tokens = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, n_tokens);
+    ggml_set_input(inp->tokens);
+
+    inp->embd = ggml_new_tensor_2d(ctx0, GGML_TYPE_F32, hparams.n_embd_inp(), n_tokens);
+    ggml_set_input(inp->embd);
+
+    ggml_tensor * tok_embd;
+    if (ubatch.token) {
+        ggml_tensor * tok_embd_w = layer.nextn.embed_tokens ? layer.nextn.embed_tokens : model.tok_embd;
+
+        tok_embd = ggml_get_rows(ctx0, tok_embd_w, inp->tokens);
+    } else {
+        tok_embd = inp->embd;
+    }
+    cb(tok_embd, "mtp_tok_embd", il);
+
+    inp->h = ggml_new_tensor_2d(ctx0, GGML_TYPE_F32, hparams.n_embd, n_tokens);
+    ggml_set_input(inp->h);
+    ggml_set_name(inp->h, "mtp_h_input");
+
+    ggml_tensor * h_embd = inp->h;
+
+    res->add_input(std::move(inp));
+
+    ggml_tensor * inp_pos     = build_inp_pos();
+    ggml_tensor * inp_out_ids = build_inp_out_ids();
+
+    // MLA with the absorption optimization uses a K-only cache (V is a view of K)
+    auto * inp_attn = build_attn_inp_k();
+
+    ggml_tensor * h_norm = build_norm(h_embd, layer.nextn.hnorm, nullptr, LLM_NORM_RMS, il);
+    cb(h_norm, "mtp_hnorm", il);
+
+    ggml_tensor * e_norm = build_norm(tok_embd, layer.nextn.enorm, nullptr, LLM_NORM_RMS, il);
+    cb(e_norm, "mtp_enorm", il);
+
+    ggml_tensor * concat = ggml_concat(ctx0, e_norm, h_norm, /*dim=*/ 0);
+    cb(concat, "mtp_concat", il);
+
+    ggml_tensor * cur = build_lora_mm(layer.nextn.eh_proj, concat, layer.nextn.eh_proj_s);
+    cb(cur, "mtp_eh_proj", il);
+
+    ggml_tensor * inpSA = cur;
+
+    cur = build_norm(cur, layer.attn_norm, nullptr, LLM_NORM_RMS, il);
+    cb(cur, "mtp_attn_norm", il);
+
+    // self-attention: dense MLA, same construction as the deepseek2 trunk graph
+    {
+        ggml_tensor * q = ggml_mul_mat(ctx0, layer.wq_a, cur);
+        cb(q, "mtp_q", il);
+
+        q = build_norm(q, layer.attn_q_a_norm, nullptr, LLM_NORM_RMS, il);
+        cb(q, "mtp_q", il);
+
+        q = ggml_mul_mat(ctx0, layer.wq_b, q);
+        cb(q, "mtp_q", il);
+
+        // split into {n_embd_head_qk_nope, n_head, n_tokens}
+        ggml_tensor * q_nope =
+            ggml_view_3d(ctx0, q, n_embd_head_qk_nope, n_head, n_tokens, ggml_row_size(q->type, n_embd_head_k),
+                         ggml_row_size(q->type, n_embd_head_k) * n_head, 0);
+        cb(q_nope, "mtp_q_nope", il);
+
+        // and {n_embd_head_qk_rope, n_head, n_tokens}
+        ggml_tensor * q_pe = ggml_view_3d(
+            ctx0, q, n_embd_head_qk_rope, n_head, n_tokens, ggml_row_size(q->type, n_embd_head_k),
+            ggml_row_size(q->type, n_embd_head_k) * n_head, ggml_row_size(q->type, n_embd_head_qk_nope));
+        cb(q_pe, "mtp_q_pe", il);
+
+        ggml_tensor * kv_cmpr_pe = ggml_mul_mat(ctx0, layer.wkv_a_mqa, cur);
+        cb(kv_cmpr_pe, "mtp_kv_cmpr_pe", il);
+
+        // split into {kv_lora_rank, n_tokens}
+        ggml_tensor * kv_cmpr =
+            ggml_view_2d(ctx0, kv_cmpr_pe, kv_lora_rank, n_tokens,
+                         ggml_row_size(kv_cmpr_pe->type, kv_lora_rank + n_embd_head_qk_rope), 0);
+        cb(kv_cmpr, "mtp_kv_cmpr", il);
+
+        // and {n_embd_head_qk_rope, 1, n_tokens}
+        ggml_tensor * k_pe = ggml_view_3d(ctx0, kv_cmpr_pe, n_embd_head_qk_rope, 1, n_tokens,
+                                          ggml_row_size(kv_cmpr_pe->type, kv_lora_rank + n_embd_head_qk_rope),
+                                          ggml_row_size(kv_cmpr_pe->type, kv_lora_rank + n_embd_head_qk_rope),
+                                          ggml_row_size(kv_cmpr_pe->type, kv_lora_rank));
+        cb(k_pe, "mtp_k_pe", il);
+
+        q_pe = ggml_rope_ext(ctx0, q_pe, inp_pos, nullptr, n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
+                             ext_factor, attn_factor, beta_fast, beta_slow);
+        cb(q_pe, "mtp_q_pe", il);
+
+        k_pe = ggml_rope_ext(ctx0, k_pe, inp_pos, nullptr, n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
+                             ext_factor, attn_factor, beta_fast, beta_slow);
+        cb(k_pe, "mtp_k_pe", il);
+
+        kv_cmpr = build_norm(kv_cmpr, layer.attn_kv_a_norm, nullptr, LLM_NORM_RMS, il);
+        cb(kv_cmpr, "mtp_kv_cmpr", il);
+
+        // {n_embd_head_qk_nope, n_tokens, n_head}
+        q_nope = ggml_permute(ctx0, q_nope, 0, 2, 1, 3);
+        cb(q_nope, "mtp_q_nope_perm", il);
+
+        // {n_embd_head_qk_nope, kv_lora_rank, n_head} x {n_embd_head_qk_nope, n_tokens, n_head}
+        ggml_tensor * q_nope_absorbed = ggml_mul_mat(ctx0, layer.wk_b, q_nope);
+        cb(q_nope_absorbed, "mtp_q_nope_absorbed", il);
+
+        // {kv_lora_rank, n_head, n_tokens}
+        q_nope_absorbed = ggml_permute(ctx0, q_nope_absorbed, 0, 2, 1, 3);
+        cb(q_nope_absorbed, "mtp_q_nope_absorbed_perm", il);
+
+        // {n_embd_head_qk_rope + kv_lora_rank, n_head, n_tokens}
+        // note: rope must go first for in-place context shifting in build_rope_shift()
+        ggml_tensor * Qcur = ggml_concat(ctx0, q_nope_absorbed, q_pe, 0);
+        cb(Qcur, "mtp_Qcur", il);
+
+        kv_cmpr = ggml_reshape_3d(ctx0, kv_cmpr, kv_lora_rank, 1, n_tokens);
+        cb(kv_cmpr, "mtp_kv_cmpr_reshape", il);
+
+        // {n_embd_head_qk_rope + kv_lora_rank, 1, n_tokens}
+        ggml_tensor * Kcur = ggml_concat(ctx0, kv_cmpr, k_pe, 0);
+        cb(Kcur, "mtp_Kcur", il);
+
+        // {kv_lora_rank, 1, n_tokens}
+        ggml_tensor * Vcur = kv_cmpr;
+        cb(Vcur, "mtp_Vcur", il);
+
+        // note: MLA with the absorption optimization converts into MQA (ie: GQA with 1 group)
+        cur = build_attn(inp_attn,
+                layer.wo, NULL, layer.wo_s,
+                Qcur, Kcur, Vcur, nullptr, nullptr, layer.wv_b, kq_scale, il);
+        cb(cur, "mtp_attn_out", il);
+    }
+
+    ggml_tensor * ffn_inp = ggml_add(ctx0, cur, inpSA);
+    cb(ffn_inp, "mtp_ffn_inp", il);
+
+    cur = build_norm(ffn_inp, layer.ffn_norm, NULL, LLM_NORM_RMS, il);
+    cb(cur, "mtp_ffn_norm", il);
+
+    // MoE FFN with shared expert - same construction as the deepseek2 trunk graph
+    ggml_tensor * moe_out = build_moe_ffn(cur,
+        layer.ffn_gate_inp,
+        layer.ffn_up_exps,
+        layer.ffn_gate_exps,
+        layer.ffn_down_exps,
+        layer.ffn_exp_probs_b,
+        n_expert, n_expert_used,
+        LLM_FFN_SILU, hparams.expert_weights_norm,
+        hparams.expert_weights_scale,
+        (llama_expert_gating_func_type) hparams.expert_gating_func,
+        il,
+        nullptr,
+        layer.ffn_gate_up_exps,
+        layer.ffn_up_exps_s,
+        layer.ffn_gate_exps_s,
+        layer.ffn_down_exps_s);
+    cb(moe_out, "mtp_ffn_moe_out", il);
+
+    // FFN shared expert
+    ggml_tensor * ffn_shexp =
+        build_ffn(cur,
+            layer.ffn_up_shexp, NULL, layer.ffn_up_shexp_s,
+            layer.ffn_gate_shexp, NULL, layer.ffn_gate_shexp_s,
+            layer.ffn_down_shexp, NULL, layer.ffn_down_shexp_s,
+            NULL, LLM_FFN_SILU, LLM_FFN_PAR, il);
+    cb(ffn_shexp, "mtp_ffn_shexp", il);
+
+    cur = ggml_add(ctx0, moe_out, ffn_shexp);
+    cb(cur, "mtp_ffn_out", il);
+
+    cur = ggml_add(ctx0, cur, ffn_inp);
+    cb(cur, "mtp_post_ffn", il);
+
+    // shared_head_norm applied after the decoder block, before the shared LM head.
+    // The post-norm hidden state seeds the next MTP step.
+    ggml_tensor * head_norm_w = layer.nextn.shared_head_norm
+            ? layer.nextn.shared_head_norm
+            : model.output_norm;
+    GGML_ASSERT(head_norm_w && "DEEPSEEK32 MTP: missing both nextn.shared_head_norm and output_norm");
+    cur = build_norm(cur, head_norm_w, nullptr, LLM_NORM_RMS, -1);
+
+    cb(cur, "h_nextn", -1);
+    res->t_h_nextn = cur;
+
+    cur = ggml_get_rows(ctx0, cur, inp_out_ids);
+    cb(cur, "mtp_shared_head_norm", -1);
+
+    ggml_tensor * head_w = layer.nextn.shared_head_head ? layer.nextn.shared_head_head : model.output;
+    ggml_tensor * head_s = layer.nextn.shared_head_head ? layer.nextn.shared_head_head_s : model.output_s;
+    GGML_ASSERT(head_w && "DEEPSEEK32 MTP: missing LM head (nextn.shared_head_head or model.output)");
+    cur = build_lora_mm(head_w, cur, head_s);
+    cb(cur, "result_output", -1);
+
+    res->t_logits = cur;
+    ggml_build_forward_expand(gf, cur);
+}
+

src/models/minimax-m3.cpp

@@ -1,5 +1,5 @@
 #include "models.h"
-#include "llama-kv-cache.h"
+#include "llama-kv-cache-msa.h"
 #include <cmath>
 #include <vector>
 #include <cstdint>
@@ -7,7 +7,8 @@
 // MiniMax-M3: MiniMax-M2 style GQA (per-head QK-norm, partial rotary) with
 // DeepSeek-V3 leading-dense + routed/shared experts (sigmoid gating, routed scaling),
 // swigluoai activation, and MiniMax Sparse Attention (MSA). MTP is not in released model weights.
-// Notes: Blocks are anchored to absolute KV cache slots.
+// MSA blocks are defined over token positions. The graph translates between position space (block
+// selection) and cell space (K/V/indexer storage) via per-ubatch pos<->cell maps populated from llama_kv_cells
 
 void llama_model_minimax_m3::load_arch_hparams(llama_model_loader & ml) {
     ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
@@ -23,7 +24,6 @@ void llama_model_minimax_m3::load_arch_hparams(llama_model_loader & ml) {
     ml.get_key(LLM_KV_ATTENTION_INDEXER_BLOCK_SIZE,    hparams.indexer_block_size);
     ml.get_key(LLM_KV_ATTENTION_INDEXER_LOCAL_BLOCKS,  hparams.indexer_local_blocks);
     msa_p = { (int) hparams.indexer_block_size, (int) hparams.indexer_top_k, (int) hparams.indexer_local_blocks };
-    hparams.indexer_kv = true;
 
     switch (hparams.n_layer()) {
         case 60: type = LLM_TYPE_428B_A23B; break;
@@ -86,43 +86,83 @@ std::unique_ptr<llm_graph_context> llama_model_minimax_m3::build_arch_graph(cons
     return std::make_unique<graph>(*this, params);
 }
 
-// per-query local-force bias for MSA selection
-// local window always wins a slot
-class llm_graph_input_msa_local : public llm_graph_input_i {
+class llm_graph_input_msa : public llm_graph_input_i {
 public:
-    llm_graph_input_msa_local(int blk, int local, int64_t nblk) : blk(blk), local(local), nblk(nblk) {}
+    llm_graph_input_msa(const llama_kv_cache_msa_context * mctx, int blk, int local) :
+        mctx(mctx), blk(blk), local(local) {}
 
     void set_input(const llama_ubatch * ubatch) override {
-        if (!bias || !ubatch->pos) {
-            return;
-        }
-        const int64_t n_tokens = ubatch->n_tokens;
-        std::vector<float> data((size_t) nblk * n_tokens, 0.0f);
-        for (int64_t i = 0; i < n_tokens; ++i) {
-            const int64_t L = ubatch->pos[i] / blk;
-            for (int l = 0; l < local && L - l >= 0; ++l) {
-                if (L - l < nblk) {
-                    data[(size_t) i * nblk + (L - l)] = 1e30f;
+        if (pos_slot_i) { mctx->set_input_pos_slot(pos_slot_i, ubatch); }
+        if (pos_slot_f) { mctx->set_input_pos_slot(pos_slot_f, ubatch); }
+        if (cell_blk)   { mctx->set_input_cell_pos(cell_blk, ubatch, blk); }
+        if (pos_mask)   { mctx->set_input_pos_mask(pos_mask, ubatch); }
+
+        // local-force bias over position blocks
+        if (bias && ubatch->pos) {
+            const int64_t n_tokens = ubatch->n_tokens;
+            const int64_t nblk     = bias->ne[0];
+            std::vector<float> data((size_t) nblk * n_tokens, 0.0f);
+            for (int64_t i = 0; i < n_tokens; ++i) {
+                const int64_t L = ubatch->pos[i] / blk;
+                for (int l = 0; l < local && L - l >= 0; ++l) {
+                    if (L - l < nblk) {
+                        data[(size_t) i * nblk + (L - l)] = 1e30f;
+                    }
                 }
             }
+            ggml_backend_tensor_set(bias, data.data(), 0, data.size() * sizeof(float));
         }
-        ggml_backend_tensor_set(bias, data.data(), 0, data.size() * sizeof(float));
     }
 
-    // valid as long as the bias tensor dims still match the new ubatch/cache window
+    // valid as long as the tensor dims still match the new ubatch/cache window and the
+    // ubatch is in the same regime (decode graphs have pos_slot_f, batch graphs cell_blk)
     bool can_reuse(const llm_graph_params & params) override {
-        const auto * mctx = static_cast<const llama_kv_cache_context *>(params.mctx);
+        const auto * mctx_new = static_cast<const llama_kv_cache_msa_context *>(params.mctx);
+
+        this->mctx = mctx_new;
+
+        const int64_t n_ps = GGML_PAD((int64_t) mctx_new->get_n_pos(), blk);
+        const int64_t ns   = params.cparams.kv_unified ? 1 : params.ubatch.n_seqs_unq;
+
+        const bool decode = params.ubatch.n_tokens == ns;   // one token per stream
 
         bool res = true;
-        res &= bias->ne[1] == params.ubatch.n_tokens;
-        res &= bias->ne[0] * blk == (int64_t) mctx->get_n_kv();
+
+        res &= bias->ne[0] * blk == n_ps;
+        res &= bias->ne[1]       == params.ubatch.n_tokens;
+
+        res &= pos_mask->ne[0] == n_ps;
+        res &= pos_mask->ne[1] == params.ubatch.n_tokens;
+
+        res &= pos_slot_i->ne[0] == n_ps;
+        res &= pos_slot_i->ne[1] == ns;
+
+        res &= decode == (pos_slot_f != nullptr);
+        res &= decode == (cell_blk   == nullptr);
+
+        if (pos_slot_f) {
+            res &= pos_slot_f->ne[0] == n_ps;
+            res &= pos_slot_f->ne[1] == ns;
+        }
+
+        if (cell_blk) {
+            res &= cell_blk->ne[0] == (int64_t) mctx_new->get_base()->get_n_kv();
+            res &= cell_blk->ne[1] == ns;
+        }
+
         return res;
     }
 
-    ggml_tensor * bias = nullptr;
-    int     blk;
-    int     local;
-    int64_t nblk;
+    ggml_tensor * bias       = nullptr; // F32 [nblk, n_tokens] local-force bias (position blocks)
+    ggml_tensor * pos_mask   = nullptr; // F32 [n_ps, n_tokens] 0/-inf visibility, by position
+    ggml_tensor * pos_slot_i = nullptr; // I32 [n_ps, ns]       pos -> cell (get_rows index)
+    ggml_tensor * pos_slot_f = nullptr; // F32 [n_ps, ns]       pos -> cell (gatherable values, decode)
+    ggml_tensor * cell_blk   = nullptr; // I32 [n_kv, ns]       cell -> position block (batch)
+
+    const llama_kv_cache_msa_context * mctx;
+
+    int blk;
+    int local;
 };
 
 // One FA call for all GQA groups (and at multi-stream decode, all streams) by mapping them onto the FA sequence dim (ne[3])
@@ -173,7 +213,9 @@ llama_model_minimax_m3::graph::graph(const llama_model & model, const llm_graph_
     inpL = build_inp_embd(model.tok_embd);
 
     ggml_tensor * inp_pos = build_inp_pos();
-    auto inp_attn = build_attn_inp_kv();
+
+    // ==========================================
+    // TODO: avoid such kind of complexity in the model graphs
 
     // MSA calls ggml_flash_attn_ext directly and assumes the non-transposed V layout that
     // llama.cpp only provides when flash attention is enabled. Block selection is anchored
@@ -185,6 +227,8 @@ llama_model_minimax_m3::graph::graph(const llama_model & model, const llm_graph_
     const bool streams_ok  = cparams.n_seq_max == 1 || !cparams.kv_unified;
     const bool msa_enabled = fa_on && streams_ok;
 
+    auto * inp_attn = build_attn_inp_kv_msa(msa_enabled);
+
     static bool warned_no_fa = false;
     if (!fa_on && !warned_no_fa) {
         LLAMA_LOG_WARN("%s: flash attention disabled; MSA requires it -> running DENSE attention "
@@ -197,36 +241,54 @@ llama_model_minimax_m3::graph::graph(const llama_model & model, const llm_graph_
                        "-> running DENSE attention. Output may be degraded. Drop --kv-unified to enable MSA.\n", __func__);
         warned_unified = true;
     }
+    // ==========================================
 
     // hoisted per-graph MSA state (shared by every sparse layer)
-    llm_graph_input_msa_local * msa_loc = nullptr;
+    llm_graph_input_msa * msa = nullptr;
     ggml_tensor * msa_kqm = nullptr;
-    ggml_tensor * msa_mf  = nullptr;
-    int64_t n_kv = 0, nblk = 0, ns = 1, n_tps = 0;
+    ggml_tensor * msa_mf  = nullptr;   // F32 copy of the FA mask for the final mask add
+    int64_t n_kv = 0, n_ps = 0, nblk = 0, ns = 1, n_tps = 0;
     bool msa_decode = false;           // gather (1 token per stream) vs mask
     const int     blk = mm.msa_p.blk;
     const int64_t Hd  = hparams.indexer_n_head;   // one indexer head per GQA group
 
     if (msa_enabled) {
+        const auto * mctx_msa = static_cast<const llama_kv_cache_msa_context *>(mctx);
+
         msa_kqm = inp_attn->get_kq_mask();
         n_kv  = msa_kqm->ne[0];
         n_tps = msa_kqm->ne[1];        // tokens per stream
         ns    = msa_kqm->ne[3];        // streams in this ubatch
         GGML_ASSERT(msa_kqm->type == GGML_TYPE_F16 && "MSA requires the FA (f16) mask");
         GGML_ASSERT(n_tps*ns == n_tokens);
-        GGML_ASSERT(n_kv % blk == 0 &&
-            "MSA: KV/mask n_kv must be a multiple of indexer.block_size (128); "
-            "the flash-attention KV padding must be a multiple of the block size. "
-            "A non-multiple would silently drop the partial tail block.");
-        nblk = n_kv / blk;
+
+        // the position axis covers every position currently in the cache and is padded to whole blocks
+        n_ps = GGML_PAD((int64_t) mctx_msa->get_n_pos(), blk);
+        nblk = n_ps / blk;
         msa_decode = n_tps == 1;
 
-        msa_mf = ggml_cast(ctx0, msa_kqm, GGML_TYPE_F32);
+        auto inp = std::make_unique<llm_graph_input_msa>(mctx_msa, blk, mm.msa_p.local);
 
-        auto loc = std::make_unique<llm_graph_input_msa_local>(blk, mm.msa_p.local, nblk);
-        loc->bias = ggml_new_tensor_2d(ctx0, GGML_TYPE_F32, nblk, n_tokens);  // stream-grouped tokens
-        ggml_set_input(loc->bias);
-        msa_loc = (llm_graph_input_msa_local *) res->add_input(std::move(loc));
+        inp->bias = ggml_new_tensor_2d(ctx0, GGML_TYPE_F32, nblk, n_tokens);  // stream-grouped tokens
+        ggml_set_input(inp->bias);
+
+        inp->pos_mask = ggml_new_tensor_2d(ctx0, GGML_TYPE_F32, n_ps, n_tokens);
+        ggml_set_input(inp->pos_mask);
+
+        inp->pos_slot_i = ggml_new_tensor_2d(ctx0, GGML_TYPE_I32, n_ps, ns);
+        ggml_set_input(inp->pos_slot_i);
+
+        if (msa_decode) {
+            inp->pos_slot_f = ggml_new_tensor_2d(ctx0, GGML_TYPE_F32, n_ps, ns);
+            ggml_set_input(inp->pos_slot_f);
+        } else {
+            inp->cell_blk = ggml_new_tensor_2d(ctx0, GGML_TYPE_I32, n_kv, ns);
+            ggml_set_input(inp->cell_blk);
+
+            msa_mf = ggml_cast(ctx0, msa_kqm, GGML_TYPE_F32);
+        }
+
+        msa = (llm_graph_input_msa *) res->add_input(std::move(inp));
     }
 
     ggml_tensor * inp_out_ids = build_inp_out_ids();
@@ -283,9 +345,11 @@ llama_model_minimax_m3::graph::graph(const llama_model & model, const llm_graph_
                 ik = ggml_rope_ext(ctx0, ik, inp_pos, nullptr, n_rot, rope_type, n_ctx_orig,
                                    freq_base, freq_scale, ext_factor, attn_factor, beta_fast, beta_slow);
 
-                const auto * mctx_cur = inp_attn->mctx;
-                ggml_build_forward_expand(gf, mctx_cur->cpy_k_idx(ctx0, ik, inp_attn->get_k_idxs(), il));
-                ggml_tensor * ik_kv = mctx_cur->get_k_idx(ctx0, il);
+                const auto * mctx_msa_l = static_cast<const llama_kv_cache_msa_context *>(mctx);
+                const auto * mctx_cur = mctx_msa_l->get_base();
+                const auto * mctx_idx = mctx_msa_l->get_idx();
+                ggml_build_forward_expand(gf, mctx_idx->cpy_k(ctx0, ik, inp_attn->get_k_idxs_idx(), il));
+                ggml_tensor * ik_kv = mctx_idx->get_k(ctx0, il);
 
                 if (inp_attn->self_k_rot) {
                     Qcur = llama_mul_mat_hadamard(ctx0, Qcur, inp_attn->self_k_rot);
@@ -316,42 +380,52 @@ llama_model_minimax_m3::graph::graph(const llama_model & model, const llm_graph_
 
                 if (msa_decode) {
                     // decode: batched over streams top-k + gather, one grouped FA
-                    // scores: per-stream batched matmul over the stream dim (ne[3]).
-                    // the cache views are not contiguous across streams (stride = kv_size, not n_kv)
-                    ggml_tensor * ikv4 = ggml_view_4d(ctx0, ik_kv, n_idx_dim, n_kv, 1, ns,
-                            ik_kv->nb[2], ik_kv->nb[3], ik_kv->nb[3], 0);
+                    // gather the indexer keys through the pos -> cell map
+                    ggml_tensor * ik3 = ggml_view_3d(ctx0, ik_kv, n_idx_dim, n_kv, ns,
+                            ik_kv->nb[2], ik_kv->nb[3], 0);
+                    ggml_tensor * ikp = ggml_get_rows(ctx0, ik3, msa->pos_slot_i);   // [n_idx_dim, n_ps, ns]
                     ggml_tensor * iq4 = ggml_reshape_4d(ctx0, iq, n_idx_dim, Hd, 1, ns);
-                    ggml_tensor * sc  = ggml_mul_mat(ctx0, ikv4, iq4);
+                    ggml_tensor * sc  = ggml_mul_mat(ctx0,
+                            ggml_reshape_4d(ctx0, ikp, n_idx_dim, n_ps, 1, ns), iq4);
                     ggml_mul_mat_set_prec(sc, GGML_PREC_F32);
-                    sc = ggml_add_inplace(ctx0, sc, msa_mf);
+                    // unmapped positions come out -inf, so they can never rank into the top-k
+                    sc = ggml_add_inplace(ctx0, sc,
+                            ggml_reshape_4d(ctx0, msa->pos_mask, n_ps, 1, 1, ns));
                     ggml_tensor * bs = ggml_pool_2d(ctx0, sc, GGML_OP_POOL_MAX, blk, 1, blk, 1, 0, 0);
                     cb(bs, "msa_bs", il);
 
                     ggml_tensor * bsf = ggml_add(ctx0, bs,
-                            ggml_reshape_4d(ctx0, msa_loc->bias, nblk, 1, 1, ns));
-                    ggml_tensor * idx = ggml_top_k(ctx0, bsf, K);
+                            ggml_reshape_4d(ctx0, msa->bias, nblk, 1, 1, ns));
+                    ggml_tensor * idx = ggml_top_k(ctx0, bsf, K);   // position blocks
 
-                    // token idx: tj[t,k,h,s] = blk*idx[k,h,s] + t   (for the mask gather)
-                    // row   idx: tr[t,k,h,s] = tj*HKV + h           (for the per-stream K/V gather)
+                    // pos idx:  tj[t,k,h,s] = blk*idx[k,h,s] + t   (positions - mask gather)
+                    // cell idx: cs[t,k,h,s] = pos_slot[tj]         (pos -> cell translation)
+                    // row idx:  tr[t,k,h,s] = cs*HKV + h           (per-stream K/V gather)
                     ggml_tensor * a = ggml_scale(ctx0, ggml_cast(ctx0, idx, GGML_TYPE_F32), (float) blk);
                     a = ggml_reshape_4d(ctx0, a, 1, K, Hd, ns);
                     ggml_tensor * tj = ggml_add(ctx0,
                             ggml_repeat_4d(ctx0, a, blk, K, Hd, ns),
                             ggml_reshape_3d(ctx0, ggml_arange(ctx0, 0.0f, (float) blk, 1.0f), blk, 1, 1));
-                    ggml_tensor * tr = ggml_add(ctx0,
-                            ggml_scale(ctx0, tj, (float) HKV),
-                            ggml_reshape_3d(ctx0, ggml_arange(ctx0, 0.0f, (float) HKV, 1.0f), 1, 1, Hd));
 
                     ggml_tensor * tokj = ggml_cast(ctx0, ggml_reshape_2d(ctx0, tj, (int64_t) blk*K*Hd, ns), GGML_TYPE_I32);
+
+                    ggml_tensor * cs = ggml_get_rows(ctx0,
+                            ggml_reshape_3d(ctx0, msa->pos_slot_f, 1, n_ps, ns), tokj);   // [1, blk*K*Hd, ns]
+                    cs = ggml_reshape_4d(ctx0, cs, blk, K, Hd, ns);
+
+                    ggml_tensor * tr = ggml_add(ctx0,
+                            ggml_scale(ctx0, cs, (float) HKV),
+                            ggml_reshape_3d(ctx0, ggml_arange(ctx0, 0.0f, (float) HKV, 1.0f), 1, 1, Hd));
+
                     ggml_tensor * tokr = ggml_cast(ctx0, ggml_reshape_2d(ctx0, tr, (int64_t) blk*K*Hd, ns), GGML_TYPE_I32);
 
                     ggml_tensor * k3 = ggml_view_3d(ctx0, k, D, HKV*n_kv, ns, k->nb[1], k->nb[3], 0);
                     ggml_tensor * v3 = ggml_view_3d(ctx0, v, D, HKV*n_kv, ns, v->nb[1], v->nb[3], 0);
-                    ggml_tensor * m3 = ggml_reshape_3d(ctx0, msa_kqm, 1, n_kv, ns);
+                    ggml_tensor * mp = ggml_reshape_3d(ctx0, msa->pos_mask, 1, n_ps, ns);
 
                     ggml_tensor * kg = ggml_get_rows(ctx0, k3, tokr);
                     ggml_tensor * vg = ggml_get_rows(ctx0, v3, tokr);
-                    ggml_tensor * mg = ggml_get_rows(ctx0, m3, tokj);
+                    ggml_tensor * mg = ggml_get_rows(ctx0, mp, tokj);
 
                     // fold (group, stream) onto the FA channel dim
                     const ggml_type kt = ggml_is_quantized(k->type) ? GGML_TYPE_F16 : k->type;
@@ -372,12 +446,16 @@ llama_model_minimax_m3::graph::graph(const llama_model & model, const llm_graph_
                                 iq->nb[1], iq->nb[2], st*n_tps*iq->nb[2]);
                         ggml_tensor * ik_s = ggml_view_2d(ctx0, ik_kv, n_idx_dim, n_kv,
                                 ik_kv->nb[2], st*ik_kv->nb[3]);
-                        ggml_tensor * mf_s = ggml_view_3d(ctx0, msa_mf, n_kv, 1, n_tps,
-                                msa_mf->nb[1], msa_mf->nb[1], st*msa_mf->nb[3]);
-                        ggml_tensor * km_s = ggml_view_3d(ctx0, msa_kqm, n_kv, n_tps, 1,
-                                msa_kqm->nb[1], msa_kqm->nb[3], st*msa_kqm->nb[3]);
-                        ggml_tensor * bias_s = ggml_view_3d(ctx0, msa_loc->bias, nblk, 1, n_tps,
-                                msa_loc->bias->nb[1], msa_loc->bias->nb[1], st*n_tps*msa_loc->bias->nb[1]);
+                        ggml_tensor * psl_s = ggml_view_1d(ctx0, msa->pos_slot_i, n_ps,
+                                st*msa->pos_slot_i->nb[1]);
+                        ggml_tensor * pm_s = ggml_view_3d(ctx0, msa->pos_mask, n_ps, 1, n_tps,
+                                msa->pos_mask->nb[1], msa->pos_mask->nb[1], st*n_tps*msa->pos_mask->nb[1]);
+                        ggml_tensor * cb_s = ggml_view_1d(ctx0, msa->cell_blk, n_kv,
+                                st*msa->cell_blk->nb[1]);
+                        ggml_tensor * mf_s = ggml_view_3d(ctx0, msa_mf, n_kv, n_tps, 1,
+                                msa_mf->nb[1], msa_mf->nb[3], st*msa_mf->nb[3]);
+                        ggml_tensor * bias_s = ggml_view_3d(ctx0, msa->bias, nblk, 1, n_tps,
+                                msa->bias->nb[1], msa->bias->nb[1], st*n_tps*msa->bias->nb[1]);
                         ggml_tensor * q_s = ggml_view_3d(ctx0, Qcur, D, n_head, n_tps,
                                 Qcur->nb[1], Qcur->nb[2], st*n_tps*Qcur->nb[2]);
                         ggml_tensor * k_s = ggml_view_4d(ctx0, k, D, HKV, n_kv, 1,
@@ -385,14 +463,16 @@ llama_model_minimax_m3::graph::graph(const llama_model & model, const llm_graph_
                         ggml_tensor * v_s = ggml_view_4d(ctx0, v, D, HKV, n_kv, 1,
                                 v->nb[1], v->nb[2], v->nb[3], st*v->nb[3]);
 
-                        // block scores: bs = maxpool_blk(idx_q * idx_k^T + causal mask)
+                        // block scores: the indexer keys are gathered through the pos -> cell map first
                         // scores are unscaled, only the top-k ordering matters
-                        ggml_tensor * sc = ggml_mul_mat(ctx0, ik_s,
+                        ggml_tensor * ikp = ggml_get_rows(ctx0, ik_s, psl_s);   // [n_idx_dim, n_ps]
+                        ggml_tensor * sc = ggml_mul_mat(ctx0, ikp,
                                 ggml_reshape_2d(ctx0, iq_s, n_idx_dim, Hd*n_tps));
                         // indexer scores run in F32
                         ggml_mul_mat_set_prec(sc, GGML_PREC_F32);
-                        sc = ggml_reshape_3d(ctx0, sc, n_kv, Hd, n_tps);
-                        sc = ggml_add_inplace(ctx0, sc, mf_s);
+                        sc = ggml_reshape_3d(ctx0, sc, n_ps, Hd, n_tps);
+                        // unmapped positions (holes, padding, empty cells) come out -inf
+                        sc = ggml_add_inplace(ctx0, sc, pm_s);
                         ggml_tensor * bs = ggml_pool_2d(ctx0, sc, GGML_OP_POOL_MAX, blk, 1, blk, 1, 0, 0);
                         cb(bs, "msa_bs", il);
 
@@ -416,14 +496,16 @@ llama_model_minimax_m3::graph::graph(const llama_model & model, const llm_graph_
                         bm = ggml_cont(ctx0, ggml_permute(ctx0, bm, 0, 2, 1, 3)); // [nblk, n_tps, Hd]
                         cb(bm, "msa_block_mask", il);
 
-                        // expand block -> token granularity (j = bk*blk + t),
-                        // then combine with the causal mask in place
-                        ggml_tensor * bmx = ggml_repeat_4d(ctx0,
-                                ggml_reshape_3d(ctx0, bm, 1, nblk, n_tps*Hd),
-                                blk, nblk, n_tps*Hd, 1);
+                        // expand block -> cell granularity through the cell -> position block
+                        // map, then combine with the causal mask. empty cells are masked by the causal mask.
+                        ggml_tensor * bm2 = ggml_cont(ctx0, ggml_transpose(ctx0,
+                                ggml_reshape_2d(ctx0, bm, nblk, n_tps*Hd)));       // [n_tps*Hd, nblk]
+                        ggml_tensor * bmc = ggml_get_rows(ctx0, bm2, cb_s);        // [n_tps*Hd, n_kv] F32
+                        ggml_tensor * bmx = ggml_cont(ctx0, ggml_transpose(ctx0, bmc));
                         bmx = ggml_reshape_3d(ctx0, bmx, n_kv, n_tps, Hd);
-                        ggml_tensor * mask4 = ggml_add_inplace(ctx0, bmx, km_s);
-                        mask4 = ggml_reshape_4d(ctx0, mask4, n_kv, n_tps, 1, Hd);
+                        ggml_tensor * mask4 = ggml_add_inplace(ctx0, bmx, mf_s);
+                        mask4 = ggml_cast(ctx0,
+                                ggml_reshape_4d(ctx0, mask4, n_kv, n_tps, 1, Hd), GGML_TYPE_F16);
                         cb(mask4, "msa_mask4", il);
 
                         // cache views with groups on ne[3];

src/models/models.h

@@ -1089,6 +1089,10 @@ struct llama_model_deepseek2 : public llama_model_base {
         graph(const llama_model & model, const llm_graph_params & params);
     };
 
+    struct graph_mtp : public llm_graph_context {
+        graph_mtp(const llama_model & model, const llm_graph_params & params);
+    };
+
     std::unique_ptr<llm_graph_context> build_arch_graph(const llm_graph_params & params) const override;
 };
 
@@ -1102,6 +1106,10 @@ struct llama_model_deepseek32 : public llama_model_base {
         graph(const llama_model & model, const llm_graph_params & params);
     };
 
+    struct graph_mtp : public llm_graph_context {
+        graph_mtp(const llama_model & model, const llm_graph_params & params);
+    };
+
     std::unique_ptr<llm_graph_context> build_arch_graph(const llm_graph_params & params) const override;
 };
 
@@ -2038,6 +2046,10 @@ struct llama_model_qwen3next : public llama_model_base {
         const llama_model & model;
     };
 
+    struct graph_mtp : public llm_graph_context {
+        graph_mtp(const llama_model & model, const llm_graph_params & params);
+    };
+
     std::unique_ptr<llm_graph_context> build_arch_graph(const llm_graph_params & params) const override;
 };
 

src/models/qwen3next.cpp

@@ -13,7 +13,11 @@ void llama_model_qwen3next::load_arch_hparams(llama_model_loader & ml) {
     ml.get_key(LLM_KV_SSM_TIME_STEP_RANK, hparams.ssm_dt_rank);
     ml.get_key(LLM_KV_SSM_GROUP_COUNT,    hparams.ssm_n_group);
 
-    // Mark recurrent layers (linear attention layers)
+    // NextN/MTP: extra decoder block appended beyond the main stack
+    ml.get_key(LLM_KV_NEXTN_PREDICT_LAYERS, hparams.n_layer_nextn, false);
+    GGML_ASSERT(hparams.n_layer_nextn < hparams.n_layer_all && "n_layer_nextn must be < n_layer_all");
+
+    // Mark recurrent layers (linear attention layers).
     if (!ml.get_key_or_arr(LLM_KV_ATTENTION_RECURRENT_LAYERS, hparams.is_recr_impl, hparams.n_layer_all, false)) {
         uint32_t full_attn_interval = 4;
         ml.get_key(LLM_KV_FULL_ATTENTION_INTERVAL, full_attn_interval, false);
@@ -28,13 +32,17 @@ void llama_model_qwen3next::load_arch_hparams(llama_model_loader & ml) {
     }
 }
 
-void llama_model_qwen3next::load_arch_tensors(llama_model_loader &) {
+void llama_model_qwen3next::load_arch_tensors(llama_model_loader & ml) {
     LLAMA_LOAD_LOCALS;
 
     if (n_expert == 0) {
         throw std::runtime_error(arch_name() + " model cannot have zero experts");
     }
 
+    const bool mtp_only = (hparams.n_layer_nextn > 0) && (ml.get_weight("blk.0.attn_norm.weight") == nullptr);
+    const int trunk_flags = mtp_only ? TENSOR_NOT_REQUIRED : 0;
+    int mtp_flags = !ml.load_mtp ? TENSOR_SKIP : 0;
+
     tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), { n_embd, n_vocab }, 0);
 
     // output
@@ -61,49 +69,73 @@ void llama_model_qwen3next::load_arch_tensors(llama_model_loader &) {
     const int64_t qkvz_dim = key_dim * 2 + value_dim * 2;
     const int64_t ba_dim   = n_v_heads * 2;
 
-    for (int i = 0; i < n_layer; ++i) {
-        auto & layer = layers[i];
-        const uint32_t n_ff_shexp = hparams.n_ff_shexp > 0 ? hparams.n_ff_shexp : hparams.n_ff(i);
+    auto load_block_trunk = [&](int il, int flags) {
+        auto & layer = layers[il];
+        const uint32_t n_ff_shexp = hparams.n_ff_shexp > 0 ? hparams.n_ff_shexp : hparams.n_ff(il);
 
-        layer.attn_norm      = create_tensor(tn(LLM_TENSOR_ATTN_NORM,      "weight", i), { n_embd }, 0);
-        layer.attn_post_norm = create_tensor(tn(LLM_TENSOR_ATTN_POST_NORM, "weight", i), { n_embd }, 0);
+        layer.attn_norm      = create_tensor(tn(LLM_TENSOR_ATTN_NORM,      "weight", il), { n_embd }, flags);
+        layer.attn_post_norm = create_tensor(tn(LLM_TENSOR_ATTN_POST_NORM, "weight", il), { n_embd }, flags);
 
-        if (!hparams.is_recr(i)) {
+        if (!hparams.is_recr(il)) {
             // Attention layers
-            create_tensor_qkv(layer, i, n_embd, n_embd_head_k * n_head * 2, n_embd_k_gqa, n_embd_v_gqa, 0);
-            layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), { n_embd_head_k * n_head, n_embd }, 0);
-
+            create_tensor_qkv(layer, il, n_embd, n_embd_head_k * n_head * 2, n_embd_k_gqa, n_embd_v_gqa, flags);
+            layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", il), { n_embd_head_k * n_head, n_embd }, flags);
             // Q/K normalization for attention layers
-            layer.attn_q_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "weight", i), { n_embd_head_k }, 0);
-            layer.attn_k_norm = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "weight", i), { n_embd_head_k }, 0);
+            layer.attn_q_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "weight", il), { n_embd_head_k }, flags);
+            layer.attn_k_norm = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "weight", il), { n_embd_head_k }, flags);
         } else {
             // Linear attention (gated delta net) specific tensors
             // Create tensors with calculated dimensions
             // note: ssm_in is used by legacy GGUF
-            layer.ssm_in         = create_tensor(tn(LLM_TENSOR_SSM_IN,         "weight", i), { n_embd, qkvz_dim }, TENSOR_NOT_REQUIRED);
-            layer.wqkv           = create_tensor(tn(LLM_TENSOR_ATTN_QKV,       "weight", i), { n_embd, key_dim * 2 + value_dim }, TENSOR_NOT_REQUIRED);
-            layer.wqkv_gate      = create_tensor(tn(LLM_TENSOR_ATTN_GATE,      "weight", i), { n_embd, value_dim }, TENSOR_NOT_REQUIRED);
-            layer.ssm_conv1d     = create_tensor(tn(LLM_TENSOR_SSM_CONV1D,     "weight", i), { hparams.ssm_d_conv, conv_dim }, 0);
-            layer.ssm_dt         = create_tensor(tn(LLM_TENSOR_SSM_DT,         "bias",   i), { hparams.ssm_dt_rank }, 0);
-            layer.ssm_a          = create_tensor(tn(LLM_TENSOR_SSM_A_NOSCAN,             i), { hparams.ssm_dt_rank }, 0);
-            layer.ssm_beta_alpha = create_tensor(tn(LLM_TENSOR_SSM_BETA_ALPHA, "weight", i), { n_embd, ba_dim }, 0);
-            layer.ssm_norm       = create_tensor(tn(LLM_TENSOR_SSM_NORM,       "weight", i), { head_v_dim }, 0);
-            layer.ssm_out        = create_tensor(tn(LLM_TENSOR_SSM_OUT,        "weight", i), { value_dim, n_embd }, 0);
+            layer.ssm_in         = create_tensor(tn(LLM_TENSOR_SSM_IN,         "weight", il), { n_embd, qkvz_dim }, TENSOR_NOT_REQUIRED | flags);
+            layer.wqkv           = create_tensor(tn(LLM_TENSOR_ATTN_QKV,       "weight", il), { n_embd, key_dim * 2 + value_dim }, TENSOR_NOT_REQUIRED | flags);
+            layer.wqkv_gate      = create_tensor(tn(LLM_TENSOR_ATTN_GATE,      "weight", il), { n_embd, value_dim }, TENSOR_NOT_REQUIRED | flags);
+            layer.ssm_conv1d     = create_tensor(tn(LLM_TENSOR_SSM_CONV1D,     "weight", il), { hparams.ssm_d_conv, conv_dim }, flags);
+            layer.ssm_dt         = create_tensor(tn(LLM_TENSOR_SSM_DT,         "bias",   il), { hparams.ssm_dt_rank }, flags);
+            layer.ssm_a          = create_tensor(tn(LLM_TENSOR_SSM_A_NOSCAN,             il), { hparams.ssm_dt_rank }, flags);
+            layer.ssm_beta_alpha = create_tensor(tn(LLM_TENSOR_SSM_BETA_ALPHA, "weight", il), { n_embd, ba_dim }, flags);
+            layer.ssm_norm       = create_tensor(tn(LLM_TENSOR_SSM_NORM,       "weight", il), { head_v_dim }, flags);
+            layer.ssm_out        = create_tensor(tn(LLM_TENSOR_SSM_OUT,        "weight", il), { value_dim, n_embd }, flags);
         }
 
-        layer.ffn_gate_inp  = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP,  "weight", i), { n_embd, n_expert }, 0);
-        layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), { n_ff_exp, n_embd, n_expert }, 0);
-        create_tensor_gate_up_exps(layer, i, n_embd, n_ff_exp, n_expert, 0);
+        layer.ffn_gate_inp  = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP,  "weight", il), { n_embd, n_expert }, flags);
+        layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", il), { n_ff_exp, n_embd, n_expert }, flags);
+        create_tensor_gate_up_exps(layer, il, n_embd, n_ff_exp, n_expert, flags);
 
         // Shared experts
-        layer.ffn_gate_inp_shexp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP_SHEXP, "weight", i), { n_embd }, 0);
-        layer.ffn_gate_shexp     = create_tensor(tn(LLM_TENSOR_FFN_GATE_SHEXP,     "weight", i), { n_embd, n_ff_shexp }, 0);
-        layer.ffn_up_shexp       = create_tensor(tn(LLM_TENSOR_FFN_UP_SHEXP,       "weight", i), { n_embd, n_ff_shexp }, 0);
-        layer.ffn_down_shexp     = create_tensor(tn(LLM_TENSOR_FFN_DOWN_SHEXP,     "weight", i), { n_ff_shexp, n_embd }, 0);
+        layer.ffn_gate_inp_shexp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP_SHEXP, "weight", il), { n_embd }, flags);
+        layer.ffn_gate_shexp     = create_tensor(tn(LLM_TENSOR_FFN_GATE_SHEXP,     "weight", il), { n_embd, n_ff_shexp }, flags);
+        layer.ffn_up_shexp       = create_tensor(tn(LLM_TENSOR_FFN_UP_SHEXP,       "weight", il), { n_embd, n_ff_shexp }, flags);
+        layer.ffn_down_shexp     = create_tensor(tn(LLM_TENSOR_FFN_DOWN_SHEXP,     "weight", il), { n_ff_shexp, n_embd }, flags);
+    };
+
+    auto load_block_mtp = [&](int il) {
+        // MTP head is identical to the trunk block (full attention + FFN)
+        load_block_trunk(il, mtp_flags);
+
+        auto & layer = layers[il];
+
+        // NextN-specific tensors that define the MTP block.
+        layer.nextn.eh_proj          = create_tensor(tn(LLM_TENSOR_NEXTN_EH_PROJ,          "weight", il), { 2 * n_embd, n_embd }, mtp_flags);
+        layer.nextn.enorm            = create_tensor(tn(LLM_TENSOR_NEXTN_ENORM,            "weight", il), { n_embd },             mtp_flags);
+        layer.nextn.hnorm            = create_tensor(tn(LLM_TENSOR_NEXTN_HNORM,            "weight", il), { n_embd },             mtp_flags);
+        layer.nextn.embed_tokens     = create_tensor(tn(LLM_TENSOR_NEXTN_EMBED_TOKENS,     "weight", il), { n_embd, n_vocab },    mtp_flags | TENSOR_NOT_REQUIRED);
+        layer.nextn.shared_head_head = create_tensor(tn(LLM_TENSOR_NEXTN_SHARED_HEAD_HEAD, "weight", il), { n_embd, n_vocab },    mtp_flags | TENSOR_NOT_REQUIRED);
+        layer.nextn.shared_head_norm = create_tensor(tn(LLM_TENSOR_NEXTN_SHARED_HEAD_NORM, "weight", il), { n_embd },             mtp_flags | TENSOR_NOT_REQUIRED);
+    };
+
+    for (int i = 0; i < n_layer; i++) {
+        load_block_trunk(i, trunk_flags);
+    }
+    for (int i = n_layer; i < n_layer_all; i++) {
+        load_block_mtp(i);
     }
 }
 
 std::unique_ptr<llm_graph_context> llama_model_qwen3next::build_arch_graph(const llm_graph_params & params) const {
+    if (params.gtype == LLM_GRAPH_TYPE_DECODER_MTP) {
+        return std::make_unique<graph_mtp>(*this, params);
+    }
     return std::make_unique<graph>(*this, params);
 }
 
@@ -120,6 +152,7 @@ llama_model_qwen3next::graph::graph(const llama_model & model, const llm_graph_p
     ggml_tensor * inp_pos     = build_inp_pos();
     ggml_tensor * inp_out_ids = build_inp_out_ids();
 
+    // MTP/NextN layers are loaded as extra decoder blocks but not executed in the main pass.
     for (int il = 0; il < n_layer; ++il) {
         res->t_layer_inp[il] = inpL;
 
@@ -139,7 +172,7 @@ llama_model_qwen3next::graph::graph(const llama_model & model, const llm_graph_p
             cur = build_layer_attn(inp->get_attn(), cur, inp_pos, il);
         }
 
-        if (il == n_layer - 1 && inp_out_ids) {
+        if (il == n_layer - 1 && inp_out_ids && cparams.embeddings_nextn_masked) {
             cur   = ggml_get_rows(ctx0, cur, inp_out_ids);
             inpSA = ggml_get_rows(ctx0, inpSA, inp_out_ids);
         }
@@ -171,9 +204,16 @@ llama_model_qwen3next::graph::graph(const llama_model & model, const llm_graph_p
     }
     cur = inpL;
 
-    // Final norm
+    // post-norm hidden state is input to both the LM head and the MTP head
     cur = build_norm(cur, model.output_norm, nullptr, LLM_NORM_RMS, -1);
 
+    cb(cur, "h_nextn", -1);
+    res->t_h_nextn = cur;
+
+    if (!cparams.embeddings_nextn_masked && inp_out_ids) {
+        cur = ggml_get_rows(ctx0, cur, inp_out_ids);
+    }
+
     cb(cur, "result_norm", -1);
     res->t_embd = cur;
 
@@ -186,14 +226,6 @@ llama_model_qwen3next::graph::graph(const llama_model & model, const llm_graph_p
     ggml_build_forward_expand(gf, cur);
 }
 
-// utility to get one slice from the third dimension
-// input dim:  [x, y, c, b]
-// output dim: [x, y, 1, b]
-static ggml_tensor * get_slice_2d(ggml_context * ctx0, ggml_tensor * t, int64_t c) {
-    return ggml_view_4d(ctx0, t, t->ne[0], t->ne[1], 1, t->ne[3],
-        t->nb[1], t->nb[2], t->nb[3], t->nb[2] * c);
-}
-
 ggml_tensor * llama_model_qwen3next::graph::build_norm_gated(
         ggml_tensor * input,
         ggml_tensor * weights,
@@ -216,7 +248,7 @@ ggml_tensor * llama_model_qwen3next::graph::build_layer_attn(
     // Order: joint QG projection, QG split, Q norm, KV projection, K norm, RoPE, attention
 
     // Qwen3Next uses a single Q projection that outputs query + gate
-    ggml_tensor * Qcur_full = build_lora_mm(model.layers[il].wq, cur);
+    ggml_tensor * Qcur_full = build_lora_mm(model.layers[il].wq, cur, model.layers[il].wq_s);
     cb(Qcur_full, "Qcur_full", il);
 
     Qcur_full = ggml_reshape_4d(ctx0, Qcur_full, n_embd_head * 2, n_head, n_tokens, 1);
@@ -232,10 +264,10 @@ ggml_tensor * llama_model_qwen3next::graph::build_layer_attn(
                      Qcur_full->nb[1], Qcur_full->nb[2], Qcur_full->nb[3], n_embd_head * ggml_element_size(Qcur_full));
     cb(gate, "gate", il);
 
-    ggml_tensor * Kcur = build_lora_mm(model.layers[il].wk, cur);
+    ggml_tensor * Kcur = build_lora_mm(model.layers[il].wk, cur, model.layers[il].wk_s);
     cb(Kcur, "Kcur", il);
 
-    ggml_tensor * Vcur = build_lora_mm(model.layers[il].wv, cur);
+    ggml_tensor * Vcur = build_lora_mm(model.layers[il].wv, cur, model.layers[il].wv_s);
     cb(Vcur, "Vcur", il);
 
     Kcur = ggml_reshape_3d(ctx0, Kcur, n_embd_head, n_head_kv, n_tokens);
@@ -274,8 +306,6 @@ ggml_tensor * llama_model_qwen3next::graph::build_layer_attn(
     gate = ggml_sigmoid(ctx0, gate);
     cb(gate, "gate_sigmoid", il);
 
-    gate = ggml_reshape_2d(ctx0, gate, n_embd_head * n_head, n_tokens);
-
     cur = ggml_mul(ctx0, cur, gate);
     cb(cur, "attn_gated", il);
 
@@ -550,16 +580,19 @@ ggml_tensor * llama_model_qwen3next::graph::build_layer_ffn(ggml_tensor * cur, c
                 LLM_FFN_SILU, true,
                 hparams.expert_weights_scale,
                 LLAMA_EXPERT_GATING_FUNC_TYPE_SOFTMAX, il,
-                nullptr, model.layers[il].ffn_gate_up_exps);
+                nullptr, model.layers[il].ffn_gate_up_exps,
+                model.layers[il].ffn_up_exps_s,
+                model.layers[il].ffn_gate_exps_s,
+                model.layers[il].ffn_down_exps_s);
         cb(moe_out, "ffn_moe_out", il);
 
         // Add shared experts if present - following Qwen3Next reference implementation
         if (model.layers[il].ffn_up_shexp != nullptr) {
             ggml_tensor * ffn_shexp =
                 build_ffn(cur,
-                    model.layers[il].ffn_up_shexp,   NULL, NULL,
-                    model.layers[il].ffn_gate_shexp, NULL, NULL,
-                    model.layers[il].ffn_down_shexp, NULL, NULL,
+                    model.layers[il].ffn_up_shexp,   NULL, model.layers[il].ffn_up_shexp_s,
+                    model.layers[il].ffn_gate_shexp, NULL, model.layers[il].ffn_gate_shexp_s,
+                    model.layers[il].ffn_down_shexp, NULL, model.layers[il].ffn_down_shexp_s,
                     NULL,
                     LLM_FFN_SILU, LLM_FFN_PAR, il);
             cb(ffn_shexp, "ffn_shexp", il);
@@ -593,3 +626,198 @@ ggml_tensor * llama_model_qwen3next::graph::build_layer_ffn(ggml_tensor * cur, c
     }
     return cur;
 }
+
+// LLM_GRAPH_TYPE_DECODER_MTP draft head for Qwen3-Next
+llama_model_qwen3next::graph_mtp::graph_mtp(const llama_model & model, const llm_graph_params & params)
+    : llm_graph_context(params) {
+    GGML_ASSERT(hparams.n_layer_nextn > 0 && "QWEN3NEXT MTP requires n_layer_nextn > 0");
+    GGML_ASSERT(hparams.n_layer_nextn == 1 && "QWEN3NEXT MTP currently only supports a single MTP block");
+
+    const int64_t n_embd_head = hparams.n_embd_head_v();
+    GGML_ASSERT(n_embd_head == hparams.n_embd_head_k());
+
+    const int il = hparams.n_layer();
+    const auto & layer = model.layers[il];
+
+    GGML_ASSERT(layer.nextn.eh_proj    && "MTP block missing nextn.eh_proj");
+    GGML_ASSERT(layer.nextn.enorm      && "MTP block missing nextn.enorm");
+    GGML_ASSERT(layer.nextn.hnorm      && "MTP block missing nextn.hnorm");
+    GGML_ASSERT(layer.ffn_gate_inp     && "MTP block missing ffn_gate_inp");
+
+    // TODO: extract in a common llm_graph_context::build_inp_embd_h()
+    auto inp = std::make_unique<llm_graph_input_embd_h>(hparams.n_embd);
+
+    inp->tokens = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, n_tokens);
+    ggml_set_input(inp->tokens);
+
+    inp->embd = ggml_new_tensor_2d(ctx0, GGML_TYPE_F32, hparams.n_embd_inp(), n_tokens);
+    ggml_set_input(inp->embd);
+
+    // TODO: make static using `ggml_build_forward_select()`
+    //       see llm_graph_context::build_inp_embd() for reference
+    ggml_tensor * tok_embd;
+    if (ubatch.token) {
+        ggml_tensor * tok_embd_w = layer.nextn.embed_tokens ? layer.nextn.embed_tokens : model.tok_embd;
+
+        tok_embd = ggml_get_rows(ctx0, tok_embd_w, inp->tokens);
+    } else {
+        tok_embd = inp->embd;
+    }
+    cb(tok_embd, "mtp_tok_embd", il);
+
+    inp->h = ggml_new_tensor_2d(ctx0, GGML_TYPE_F32, hparams.n_embd, n_tokens);
+    ggml_set_input(inp->h);
+    ggml_set_name(inp->h, "mtp_h_input");
+
+    ggml_tensor * h_embd = inp->h;
+
+    res->add_input(std::move(inp));
+
+    ggml_tensor * inp_pos     = build_inp_pos();
+    ggml_tensor * inp_out_ids = build_inp_out_ids();
+
+    auto * inp_attn = build_attn_inp_kv();
+
+    ggml_tensor * h_norm = build_norm(h_embd, layer.nextn.hnorm, nullptr, LLM_NORM_RMS, il);
+    cb(h_norm, "mtp_hnorm", il);
+
+    ggml_tensor * e_norm = build_norm(tok_embd, layer.nextn.enorm, nullptr, LLM_NORM_RMS, il);
+    cb(e_norm, "mtp_enorm", il);
+
+    ggml_tensor * concat = ggml_concat(ctx0, e_norm, h_norm, /*dim=*/ 0);
+    cb(concat, "mtp_concat", il);
+
+    ggml_tensor * cur = build_lora_mm(layer.nextn.eh_proj, concat, layer.nextn.eh_proj_s);
+    cb(cur, "mtp_eh_proj", il);
+
+    ggml_tensor * inpSA = cur;
+
+    cur = build_norm(cur, layer.attn_norm, nullptr, LLM_NORM_RMS, il);
+    cb(cur, "mtp_attn_norm", il);
+
+    ggml_tensor * Qcur_full = build_lora_mm(layer.wq, cur, layer.wq_s);
+    cb(Qcur_full, "mtp_Qcur_full", il);
+
+    ggml_tensor * Qcur = ggml_view_3d(ctx0, Qcur_full,
+            n_embd_head, n_head, n_tokens,
+            ggml_element_size(Qcur_full) * n_embd_head * 2,
+            ggml_element_size(Qcur_full) * n_embd_head * 2 * n_head,
+            0);
+    Qcur = build_norm(Qcur, layer.attn_q_norm, nullptr, LLM_NORM_RMS, il);
+    cb(Qcur, "mtp_Qcur_normed", il);
+
+    ggml_tensor * Kcur = build_lora_mm(layer.wk, cur, layer.wk_s);
+    Kcur = ggml_reshape_3d(ctx0, Kcur, n_embd_head, n_head_kv, n_tokens);
+    Kcur = build_norm(Kcur, layer.attn_k_norm, nullptr, LLM_NORM_RMS, il);
+    cb(Kcur, "mtp_Kcur_normed", il);
+
+    ggml_tensor * Vcur = build_lora_mm(layer.wv, cur, layer.wv_s);
+    Vcur = ggml_reshape_3d(ctx0, Vcur, n_embd_head, n_head_kv, n_tokens);
+
+    Qcur = ggml_rope_ext(ctx0, Qcur, inp_pos, nullptr,
+            n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
+            ext_factor, attn_factor, beta_fast, beta_slow);
+    Kcur = ggml_rope_ext(ctx0, Kcur, inp_pos, nullptr,
+            n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
+            ext_factor, attn_factor, beta_fast, beta_slow);
+
+    cb(Qcur, "mtp_Qcur", il);
+    cb(Kcur, "mtp_Kcur", il);
+    cb(Vcur, "mtp_Vcur", il);
+
+    const float kq_scale = hparams.f_attention_scale == 0.0f
+            ? 1.0f / sqrtf(float(n_embd_head)) : hparams.f_attention_scale;
+
+    cur = build_attn(inp_attn,
+            nullptr, nullptr, nullptr,
+            Qcur, Kcur, Vcur, nullptr, nullptr, nullptr, kq_scale, il);
+    cb(cur, "mtp_attn_pregate", il);
+
+    ggml_tensor * gate = ggml_view_3d(ctx0, Qcur_full,
+            n_embd_head, n_head, n_tokens,
+            ggml_element_size(Qcur_full) * n_embd_head * 2,
+            ggml_element_size(Qcur_full) * n_embd_head * 2 * n_head,
+            ggml_element_size(Qcur_full) * n_embd_head);
+
+    // TODO: CUDA is missing non-contiguous unary ops. when implemented: remove this cont
+    gate = ggml_cont_2d(ctx0, gate, n_embd_head * n_head, n_tokens);
+    cb(gate, "mtp_gate", il);
+
+    cur = ggml_mul(ctx0, cur, ggml_sigmoid(ctx0, gate));
+    cur = build_lora_mm(layer.wo, cur, layer.wo_s);
+    cb(cur, "mtp_attn_out", il);
+
+    if (inp_out_ids) {
+        cur   = ggml_get_rows(ctx0, cur,   inp_out_ids);
+        inpSA = ggml_get_rows(ctx0, inpSA, inp_out_ids);
+    }
+
+    cur = ggml_add(ctx0, cur, inpSA);
+    cb(cur, "mtp_attn_residual", il);
+
+    ggml_tensor * ffn_residual = cur;
+    cur = build_norm(cur, layer.attn_post_norm, nullptr, LLM_NORM_RMS, il);
+    cb(cur, "mtp_attn_post_norm", il);
+
+    // MoE FFN — routed experts plus gated shared expert (mirrors the trunk).
+    ggml_tensor * moe_out =
+        build_moe_ffn(cur,
+            layer.ffn_gate_inp,
+            layer.ffn_up_exps,
+            layer.ffn_gate_exps,
+            layer.ffn_down_exps,
+            nullptr,
+            n_expert, n_expert_used,
+            LLM_FFN_SILU, true,
+            hparams.expert_weights_scale,
+            LLAMA_EXPERT_GATING_FUNC_TYPE_SOFTMAX, il,
+            nullptr, layer.ffn_gate_up_exps,
+            layer.ffn_up_exps_s,
+            layer.ffn_gate_exps_s,
+            layer.ffn_down_exps_s);
+    cb(moe_out, "mtp_ffn_moe_out", il);
+
+    if (layer.ffn_up_shexp != nullptr) {
+        ggml_tensor * ffn_shexp =
+            build_ffn(cur,
+                layer.ffn_up_shexp,   nullptr, layer.ffn_up_shexp_s,
+                layer.ffn_gate_shexp, nullptr, layer.ffn_gate_shexp_s,
+                layer.ffn_down_shexp, nullptr, layer.ffn_down_shexp_s,
+                nullptr,
+                LLM_FFN_SILU, LLM_FFN_PAR, il);
+        cb(ffn_shexp, "mtp_ffn_shexp", il);
+
+        ggml_tensor * shared_gate = build_lora_mm(layer.ffn_gate_inp_shexp, cur);
+        shared_gate = ggml_sigmoid(ctx0, shared_gate);
+        cb(shared_gate, "mtp_shared_expert_gate_sigmoid", il);
+
+        ffn_shexp = ggml_mul(ctx0, ffn_shexp, shared_gate);
+        cb(ffn_shexp, "mtp_ffn_shexp_gated", il);
+
+        cur = ggml_add(ctx0, moe_out, ffn_shexp);
+    } else {
+        cur = moe_out;
+    }
+    cb(cur, "mtp_ffn_out", il);
+
+    cur = ggml_add(ctx0, cur, ffn_residual);
+    cb(cur, "mtp_post_ffn", il);
+
+    ggml_tensor * head_norm_w = layer.nextn.shared_head_norm
+            ? layer.nextn.shared_head_norm
+            : model.output_norm;
+    GGML_ASSERT(head_norm_w && "QWEN3NEXT MTP: missing both nextn.shared_head_norm and output_norm");
+    cur = build_norm(cur, head_norm_w, nullptr, LLM_NORM_RMS, -1);
+
+    cb(cur, "h_nextn", -1);
+    res->t_h_nextn = cur;
+
+    ggml_tensor * head_w = layer.nextn.shared_head_head ? layer.nextn.shared_head_head : model.output;
+    ggml_tensor * head_s = layer.nextn.shared_head_head ? layer.nextn.shared_head_head_s : model.output_s;
+    GGML_ASSERT(head_w && "QWEN3NEXT MTP: missing LM head (nextn.shared_head_head or model.output)");
+    cur = build_lora_mm(head_w, cur, head_s);
+    cb(cur, "result_output", -1);
+
+    res->t_logits = cur;
+    ggml_build_forward_expand(gf, cur);
+}

tests/CMakeLists.txt

@@ -258,6 +258,9 @@ llama_build_and_test(test-thread-safety.cpp ARGS -m "${MODEL_DEST}" -ngl 99 -p "
 set_tests_properties(test-thread-safety PROPERTIES FIXTURES_REQUIRED test-download-model)
 
 llama_build_and_test(test-arg-parser.cpp)
+llama_build_and_test(test-model-resolution.cpp)
+# the test serves its repos from an httplib server, and the library links it privately
+target_link_libraries(test-model-resolution PRIVATE cpp-httplib)
 
 if (NOT LLAMA_SANITIZE_ADDRESS AND NOT GGML_SCHED_NO_REALLOC)
   # TODO: repair known memory leaks

tests/test-arg-parser.cpp

@@ -99,6 +99,34 @@ static void test(void) {
     argv = {"binary_name", "-sm", "hello"};
     assert(false == common_params_parse(argv.size(), list_str_to_char(argv).data(), params, LLAMA_EXAMPLE_COMMON));
 
+    {
+        common_params penalty_params;
+
+        argv = {"binary_name", "--repeat-penalty", "0"};
+        assert(false == common_params_parse(argv.size(), list_str_to_char(argv).data(), penalty_params, LLAMA_EXAMPLE_COMMON));
+
+        argv = {"binary_name", "--repeat-penalty", "-1"};
+        assert(false == common_params_parse(argv.size(), list_str_to_char(argv).data(), penalty_params, LLAMA_EXAMPLE_COMMON));
+
+        argv = {"binary_name", "--repeat-penalty", "nan"};
+        assert(false == common_params_parse(argv.size(), list_str_to_char(argv).data(), penalty_params, LLAMA_EXAMPLE_COMMON));
+
+        argv = {"binary_name", "--repeat-penalty", "inf"};
+        assert(false == common_params_parse(argv.size(), list_str_to_char(argv).data(), penalty_params, LLAMA_EXAMPLE_COMMON));
+
+        argv = {"binary_name", "--repeat-penalty", "-inf"};
+        assert(false == common_params_parse(argv.size(), list_str_to_char(argv).data(), penalty_params, LLAMA_EXAMPLE_COMMON));
+
+        const char * penalty_options[] = {"--frequency-penalty", "--presence-penalty"};
+        const char * nonfinite_values[] = {"nan", "inf", "-inf"};
+        for (const char * option : penalty_options) {
+            for (const char * value : nonfinite_values) {
+                argv = {"binary_name", option, value};
+                assert(false == common_params_parse(argv.size(), list_str_to_char(argv).data(), penalty_params, LLAMA_EXAMPLE_COMMON));
+            }
+        }
+    }
+
     // non-existence arg in specific example (--draft cannot be used outside llama-speculative)
     argv = {"binary_name", "--draft", "123"};
     assert(false == common_params_parse(argv.size(), list_str_to_char(argv).data(), params, LLAMA_EXAMPLE_EMBEDDING));

tests/test-backend-ops.cpp

@@ -9731,6 +9731,12 @@ static std::vector<std::unique_ptr<test_case>> make_test_cases_eval() {
         }
     }
 
+    for (int kv : { 1, 7, 8, 63, 64, 65 }) {
+        for (ggml_type type_K : {GGML_TYPE_F32, GGML_TYPE_F16, GGML_TYPE_BF16, GGML_TYPE_Q8_0, GGML_TYPE_Q5_1, GGML_TYPE_Q5_0, GGML_TYPE_Q4_1, GGML_TYPE_Q4_0}) {
+            test_cases.emplace_back(new test_lightning_indexer(128, 64, kv, 32, 4, 1, type_K));
+        }
+    }
+
     return test_cases;
 }
 #ifdef _MSC_VER

tests/test-backend-sampler.cpp

@@ -8,12 +8,15 @@
 #endif
 
 #include <algorithm>
+#include <cmath>
 #include <cstdlib>
 #include <cstring>
 #include <fstream>
+#include <functional>
 #include <map>
 #include <string>
 #include <unordered_map>
+#include <unordered_set>
 #include <vector>
 
 struct test_args {
@@ -761,6 +764,563 @@ static void test_backend_logit_bias_sampling(const test_params & params) {
     printf("backend logit bias sampling test PASSED\n");
 }
 
+static void accept_prompt(llama_sampler * smpl, const llama_vocab * vocab, const std::string & prompt) {
+    const llama_token bos = llama_vocab_bos(vocab);
+    if (bos != LLAMA_TOKEN_NULL) {
+        llama_sampler_accept(smpl, bos);
+    }
+
+    std::vector<llama_token> tokens(64);
+    int32_t n_tokens = llama_tokenize(vocab, prompt.c_str(), (int32_t) prompt.size(),
+            tokens.data(), (int32_t) tokens.size(), false, false);
+    if (n_tokens < 0) {
+        tokens.resize(-n_tokens);
+        n_tokens = llama_tokenize(vocab, prompt.c_str(), (int32_t) prompt.size(),
+                tokens.data(), (int32_t) tokens.size(), false, false);
+    }
+
+    for (int32_t i = 0; i < n_tokens; ++i) {
+        llama_sampler_accept(smpl, tokens[i]);
+    }
+}
+
+static std::vector<float> decode_raw_logits(const test_params & params, const std::string & prompt) {
+    const int seq_id = 0;
+    const int n_vocab = llama_vocab_n_tokens(llama_model_get_vocab(params.model.get()));
+    std::vector<llama_sampler_seq_config> empty_configs;
+    test_context ctx(params, empty_configs);
+
+    GGML_ASSERT(ctx.decode({{ seq_id, prompt }}));
+
+    float * logits = llama_get_logits_ith(ctx.ctx.get(), ctx.idx_for_seq(seq_id));
+    GGML_ASSERT(logits != nullptr);
+    return std::vector<float>(logits, logits + n_vocab);
+}
+
+static std::vector<llama_token_data> apply_cpu_sampler(
+        const std::vector<float> & raw_logits,
+        llama_sampler * sampler) {
+    std::vector<llama_token_data> data;
+    data.reserve(raw_logits.size());
+    for (llama_token token = 0; token < (llama_token) raw_logits.size(); ++token) {
+        data.push_back({ token, raw_logits[token], 0.0f });
+    }
+
+    llama_token_data_array cur_p = { data.data(), data.size(), -1, false };
+    llama_sampler_apply(sampler, &cur_p);
+    data.resize(cur_p.size);
+    return data;
+}
+
+using sampler_setup_fn = std::function<void(llama_sampler *)>;
+using sampler_init_fn = std::function<llama_sampler *()>;
+
+enum class penalties_position {
+    before_filter,
+    after_filter,
+};
+
+static void add_filter_and_penalties(
+        llama_sampler * chain,
+        const sampler_init_fn & init_filter,
+        int32_t penalty_last_n,
+        float penalty_repeat,
+        float penalty_freq,
+        float penalty_present,
+        penalties_position position) {
+    const auto add_penalties = [&]() {
+        llama_sampler_chain_add(chain, llama_sampler_init_penalties(
+                    penalty_last_n, penalty_repeat, penalty_freq, penalty_present));
+    };
+
+    if (position == penalties_position::before_filter) {
+        add_penalties();
+        llama_sampler_chain_add(chain, init_filter());
+    } else {
+        llama_sampler_chain_add(chain, init_filter());
+        add_penalties();
+    }
+}
+
+static llama_sampler_ptr make_sampler_chain(
+        const sampler_setup_fn & add_samplers,
+        const sampler_setup_fn & accept_history) {
+    llama_sampler_ptr chain(llama_sampler_chain_init(llama_sampler_chain_default_params()));
+    add_samplers(chain.get());
+    accept_history(chain.get());
+    return chain;
+}
+
+struct backend_sampler_output {
+    std::vector<float> logits;
+    std::vector<llama_token> candidates;
+};
+
+static backend_sampler_output run_backend_sampler(
+        const test_params & params,
+        const std::string & prompt,
+        llama_sampler * sampler) {
+    const int seq_id = 0;
+    std::vector<llama_sampler_seq_config> configs = {{ seq_id, sampler }};
+    test_context ctx(params, configs);
+
+    GGML_ASSERT(ctx.decode({{ seq_id, prompt }}));
+    llama_synchronize(ctx.ctx.get());
+
+    const int32_t idx = ctx.idx_for_seq(seq_id);
+    const uint32_t n_logits = llama_get_sampled_logits_count_ith(ctx.ctx.get(), idx);
+    const uint32_t n_candidates = llama_get_sampled_candidates_count_ith(ctx.ctx.get(), idx);
+    float * logits = llama_get_sampled_logits_ith(ctx.ctx.get(), idx);
+    llama_token * candidates = llama_get_sampled_candidates_ith(ctx.ctx.get(), idx);
+    GGML_ASSERT(logits != nullptr);
+
+    backend_sampler_output result;
+    result.logits.assign(logits, logits + n_logits);
+    result.candidates.resize(n_logits);
+
+    if (n_candidates == 0) {
+        for (uint32_t i = 0; i < n_logits; ++i) {
+            result.candidates[i] = (llama_token) i;
+        }
+    } else {
+        GGML_ASSERT(candidates != nullptr);
+        GGML_ASSERT(n_candidates == n_logits);
+        std::memcpy(result.candidates.data(), candidates, n_candidates * sizeof(llama_token));
+    }
+
+    return result;
+}
+
+struct sampler_comparison_output {
+    std::vector<llama_token_data> expected;
+    backend_sampler_output actual;
+};
+
+static sampler_comparison_output run_sampler_comparison(
+        const test_params & params,
+        const std::string & prompt,
+        const std::vector<float> & raw_logits,
+        const sampler_setup_fn & add_samplers,
+        const sampler_setup_fn & accept_history) {
+    llama_sampler_ptr cpu_chain = make_sampler_chain(add_samplers, accept_history);
+    llama_sampler_ptr backend_chain = make_sampler_chain(add_samplers, accept_history);
+    return {
+        apply_cpu_sampler(raw_logits, cpu_chain.get()),
+        run_backend_sampler(params, prompt, backend_chain.get()),
+    };
+}
+
+static std::unordered_map<llama_token, float> map_logits(const std::vector<llama_token_data> & data) {
+    std::unordered_map<llama_token, float> result;
+    result.reserve(data.size());
+    for (const auto & item : data) {
+        result[item.id] = item.logit;
+    }
+    return result;
+}
+
+struct sampler_comparison_stats {
+    int n_mismatch = 0;
+    int n_masked = 0;
+    float max_diff = 0.0f;
+};
+
+static sampler_comparison_stats compare_sampler_outputs(
+        const char * name,
+        const std::unordered_map<llama_token, float> & expected,
+        const backend_sampler_output & actual,
+        bool allow_extra_candidates = false) {
+    GGML_ASSERT(actual.logits.size() == actual.candidates.size());
+
+    sampler_comparison_stats result;
+    std::unordered_set<llama_token> seen;
+    seen.reserve(actual.candidates.size());
+
+    for (size_t i = 0; i < actual.logits.size(); ++i) {
+        const llama_token token = actual.candidates[i];
+        const float logit = actual.logits[i];
+        if (!seen.insert(token).second || std::isnan(logit)) {
+            if (result.n_mismatch < 5) {
+                printf("%s token %d has invalid backend output\n", name, token);
+            }
+            ++result.n_mismatch;
+            continue;
+        }
+
+        const auto it = expected.find(token);
+        if (it == expected.end()) {
+            if (std::isinf(logit) && logit < 0.0f) {
+                ++result.n_masked;
+            } else if (!allow_extra_candidates) {
+                if (result.n_mismatch < 5) {
+                    printf("%s token %d was not masked\n", name, token);
+                }
+                ++result.n_mismatch;
+            }
+            continue;
+        }
+
+        const float diff = fabsf(it->second - logit);
+        result.max_diff = std::max(result.max_diff, diff);
+        if (!std::isfinite(logit) || diff > 1e-3f) {
+            if (result.n_mismatch < 5) {
+                printf("%s mismatch token %d: cpu=%.6f backend=%.6f diff=%.6f\n",
+                        name, token, it->second, logit, diff);
+            }
+            ++result.n_mismatch;
+        }
+    }
+
+    for (const auto & item : expected) {
+        if (seen.find(item.first) == seen.end()) {
+            if (result.n_mismatch < 5) {
+                printf("%s missing backend token %d\n", name, item.first);
+            }
+            ++result.n_mismatch;
+        }
+    }
+
+    printf("%s logits: max_diff=%.6f n_masked=%d n_mismatch=%d\n",
+            name, result.max_diff, result.n_masked, result.n_mismatch);
+    return result;
+}
+
+static float find_backend_logit(const backend_sampler_output & output, llama_token token) {
+    for (size_t i = 0; i < output.candidates.size(); ++i) {
+        if (output.candidates[i] == token) {
+            return output.logits[i];
+        }
+    }
+    GGML_ABORT("backend token not found");
+}
+
+static sampler_comparison_output run_penalties_comparison(
+        const test_params & params,
+        int32_t penalty_last_n,
+        float penalty_repeat,
+        float penalty_freq,
+        float penalty_present,
+        const std::string & prompt,
+        const std::function<void(llama_sampler *)> & extra_accept = {}) {
+    const auto * vocab = llama_model_get_vocab(params.model.get());
+    const std::vector<float> raw_logits = decode_raw_logits(params, prompt);
+    const auto add_samplers = [&](llama_sampler * chain) {
+        llama_sampler_chain_add(chain, llama_sampler_init_penalties(
+                    penalty_last_n, penalty_repeat, penalty_freq, penalty_present));
+    };
+    const auto accept_history = [&](llama_sampler * chain) {
+        accept_prompt(chain, vocab, prompt);
+        if (extra_accept) {
+            extra_accept(chain);
+        }
+    };
+
+    return run_sampler_comparison(
+            params, prompt, raw_logits, add_samplers, accept_history);
+}
+
+static void compare_penalties_logits(
+        const test_params & params,
+        int32_t penalty_last_n,
+        float penalty_repeat,
+        float penalty_freq,
+        float penalty_present,
+        const std::string & prompt,
+        const std::function<void(llama_sampler *)> & extra_accept = {}) {
+    const sampler_comparison_output output = run_penalties_comparison(
+            params, penalty_last_n, penalty_repeat, penalty_freq, penalty_present, prompt, extra_accept);
+
+    GGML_ASSERT(output.expected.size() == output.actual.logits.size());
+
+    const sampler_comparison_stats stats = compare_sampler_outputs(
+            "penalties", map_logits(output.expected), output.actual);
+    GGML_ASSERT(stats.n_masked == 0);
+    GGML_ASSERT(stats.n_mismatch == 0);
+}
+
+static void test_penalty_parameter_values(const test_params & params) {
+    struct penalty_test_case {
+        const char * name;
+        float repeat;
+        float frequency;
+        float presence;
+    };
+
+    const penalty_test_case cases[] = {
+        { "frequency -1",   1.0f, -1.0f,     0.0f },
+        { "frequency 0",    1.0f,  0.0f,     0.0f },
+        { "frequency 1",    1.0f,  1.0f,     0.0f },
+        { "presence -1",    1.0f,  0.0f,    -1.0f },
+        { "presence 0",     1.0f,  0.0f,     0.0f },
+        { "presence 1",     1.0f,  0.0f,     1.0f },
+        { "repeat 1",       1.0f,  0.0f,     0.0f },
+    };
+
+    int n_failed = 0;
+    for (const auto & test : cases) {
+        const sampler_comparison_output output = run_penalties_comparison(
+                params, 64, test.repeat, test.frequency, test.presence, "Hello Hello world");
+        GGML_ASSERT(output.expected.size() == output.actual.logits.size());
+        const sampler_comparison_stats stats = compare_sampler_outputs(
+                test.name, map_logits(output.expected), output.actual);
+        n_failed += stats.n_mismatch != 0;
+    }
+
+    GGML_ASSERT(n_failed == 0);
+}
+
+static void compare_top_k_penalties_logits(
+        const test_params & params,
+        int32_t k,
+        int32_t penalty_last_n,
+        float penalty_repeat,
+        float penalty_freq,
+        float penalty_present,
+        const std::string & prompt,
+        penalties_position position) {
+    const auto * vocab = llama_model_get_vocab(params.model.get());
+    const std::vector<float> raw_logits = decode_raw_logits(params, prompt);
+    const int n_vocab = (int) raw_logits.size();
+
+    GGML_ASSERT(n_vocab > k);
+
+    const sampler_init_fn init_top_k = [k]() {
+        return llama_sampler_init_top_k(k);
+    };
+    llama_sampler_ptr top_k(init_top_k());
+    const std::vector<llama_token_data> top_k_data = apply_cpu_sampler(raw_logits, top_k.get());
+    GGML_ASSERT(top_k_data.size() == (size_t) k);
+    const llama_token retained_history_token = top_k_data[0].id;
+
+    llama_token excluded_history_token = LLAMA_TOKEN_NULL;
+    for (llama_token token = 0; token < n_vocab; ++token) {
+        const auto it = std::find_if(top_k_data.begin(), top_k_data.end(), [token](const llama_token_data & data) {
+            return data.id == token;
+        });
+        if (it == top_k_data.end()) {
+            excluded_history_token = token;
+            break;
+        }
+    }
+    GGML_ASSERT(excluded_history_token != LLAMA_TOKEN_NULL);
+
+    const auto add_samplers = [&](llama_sampler * chain) {
+        add_filter_and_penalties(chain, init_top_k,
+                penalty_last_n, penalty_repeat, penalty_freq, penalty_present, position);
+    };
+
+    auto accept_history = [&](llama_sampler * smpl) {
+        accept_prompt(smpl, vocab, prompt);
+        llama_sampler_accept(smpl, excluded_history_token);
+        llama_sampler_accept(smpl, excluded_history_token);
+        llama_sampler_accept(smpl, retained_history_token);
+        llama_sampler_accept(smpl, retained_history_token);
+    };
+
+    const sampler_comparison_output output = run_sampler_comparison(
+            params, prompt, raw_logits, add_samplers, accept_history);
+
+    GGML_ASSERT(output.expected.size() == (size_t) k);
+    GGML_ASSERT(output.actual.logits.size() == (size_t) k);
+
+    const std::unordered_map<llama_token, float> expected_logits = map_logits(output.expected);
+
+    if (position == penalties_position::after_filter) {
+        GGML_ASSERT(expected_logits.find(retained_history_token) != expected_logits.end());
+        GGML_ASSERT(fabsf(expected_logits.at(retained_history_token) - raw_logits[retained_history_token]) > 1e-6f);
+        GGML_ASSERT(expected_logits.find(excluded_history_token) == expected_logits.end());
+        GGML_ASSERT(std::find(output.actual.candidates.begin(), output.actual.candidates.end(),
+                    excluded_history_token) == output.actual.candidates.end());
+    } else {
+        const std::unordered_map<llama_token, float> unpenalized_logits = map_logits(top_k_data);
+        bool changed = false;
+        for (const auto & item : expected_logits) {
+            const auto it = unpenalized_logits.find(item.first);
+            if (it == unpenalized_logits.end() || fabsf(it->second - item.second) > 1e-6f) {
+                changed = true;
+                break;
+            }
+        }
+        GGML_ASSERT(changed);
+    }
+
+    const char * name = position == penalties_position::before_filter
+        ? "penalties top-k"
+        : "top-k penalties";
+    const sampler_comparison_stats stats = compare_sampler_outputs(
+            name, expected_logits, output.actual);
+    GGML_ASSERT(stats.n_masked == 0);
+    GGML_ASSERT(stats.n_mismatch == 0);
+}
+
+static void compare_masking_penalties_logits(
+        const test_params & params,
+        const char * filter_name,
+        const sampler_init_fn & init_filter,
+        int32_t penalty_last_n,
+        float penalty_repeat,
+        float penalty_freq,
+        float penalty_present,
+        const std::string & prompt,
+        penalties_position position,
+        bool allow_extra_candidates,
+        bool add_history = true) {
+    const auto * vocab = llama_model_get_vocab(params.model.get());
+    const std::vector<float> raw_logits = decode_raw_logits(params, prompt);
+    const int n_vocab = (int) raw_logits.size();
+    llama_sampler_ptr filter(init_filter());
+    const std::vector<llama_token_data> filtered_data = apply_cpu_sampler(raw_logits, filter.get());
+    GGML_ASSERT(!filtered_data.empty());
+    GGML_ASSERT(filtered_data.size() < (size_t) n_vocab);
+
+    const llama_token penalized_token = filtered_data[0].id;
+    std::unordered_set<llama_token> retained_tokens;
+    retained_tokens.reserve(filtered_data.size());
+    for (const auto & data : filtered_data) {
+        retained_tokens.insert(data.id);
+    }
+
+    llama_token masked_token = LLAMA_TOKEN_NULL;
+    for (llama_token token = 0; token < n_vocab; ++token) {
+        if (retained_tokens.find(token) == retained_tokens.end()) {
+            masked_token = token;
+            break;
+        }
+    }
+    GGML_ASSERT(masked_token != LLAMA_TOKEN_NULL);
+
+    const auto add_samplers = [&](llama_sampler * chain) {
+        add_filter_and_penalties(chain, init_filter,
+                penalty_last_n, penalty_repeat, penalty_freq, penalty_present, position);
+    };
+    auto accept_history = [&](llama_sampler * smpl) {
+        if (!add_history) {
+            return;
+        }
+        accept_prompt(smpl, vocab, prompt);
+        llama_sampler_accept(smpl, penalized_token);
+        llama_sampler_accept(smpl, penalized_token);
+        llama_sampler_accept(smpl, masked_token);
+        llama_sampler_accept(smpl, masked_token);
+    };
+
+    const sampler_comparison_output output = run_sampler_comparison(
+            params, prompt, raw_logits, add_samplers, accept_history);
+
+    GGML_ASSERT(output.actual.logits.size() == (size_t) n_vocab);
+
+    const std::unordered_map<llama_token, float> expected_logits = map_logits(output.expected);
+
+    GGML_ASSERT(expected_logits.find(masked_token) == expected_logits.end());
+    if (add_history) {
+        if (position == penalties_position::after_filter) {
+            GGML_ASSERT(expected_logits.find(penalized_token) != expected_logits.end());
+            GGML_ASSERT(fabsf(expected_logits.at(penalized_token) - raw_logits[penalized_token]) > 1e-6f);
+        } else {
+            llama_sampler_ptr penalties(llama_sampler_init_penalties(
+                        penalty_last_n, penalty_repeat, penalty_freq, penalty_present));
+            accept_history(penalties.get());
+            const std::unordered_map<llama_token, float> penalized_logits =
+                map_logits(apply_cpu_sampler(raw_logits, penalties.get()));
+            GGML_ASSERT(fabsf(penalized_logits.at(penalized_token) - raw_logits[penalized_token]) > 1e-6f);
+        }
+    }
+
+    const std::string name = position == penalties_position::before_filter
+        ? "penalties " + std::string(filter_name)
+        : std::string(filter_name) + " penalties";
+    const sampler_comparison_stats stats = compare_sampler_outputs(
+            name.c_str(), expected_logits, output.actual, allow_extra_candidates);
+    const float masked_logit = find_backend_logit(output.actual, masked_token);
+    GGML_ASSERT(stats.n_masked > 0);
+    GGML_ASSERT(std::isinf(masked_logit) && masked_logit < 0.0f);
+    GGML_ASSERT(stats.n_mismatch == 0);
+}
+
+static void test_backend_penalties_sampling(const test_params & params) {
+    printf("Testing backend penalties (repeat + freq + presence)\n");
+    compare_penalties_logits(params, 64, 1.1f, 0.5f, 0.25f, "Hello Hello world");
+
+    printf("Testing backend penalties with penalty_last_n > 64\n");
+    const auto * vocab = llama_model_get_vocab(params.model.get());
+    std::vector<llama_token> tokens(8);
+    int32_t n_tok = llama_tokenize(vocab, "a", 1, tokens.data(), (int32_t) tokens.size(), false, false);
+    if (n_tok < 0) {
+        tokens.resize(-n_tok);
+        n_tok = llama_tokenize(vocab, "a", 1, tokens.data(), (int32_t) tokens.size(), false, false);
+    }
+    GGML_ASSERT(n_tok > 0);
+    const llama_token tok = tokens[0];
+
+    compare_penalties_logits(params, 80, 1.15f, 0.1f, 0.05f, "a", [tok](llama_sampler * smpl) {
+        // accept_prompt already accepted BOS + one 'a'; fill the ring to n=80
+        for (int i = 0; i < 78; ++i) {
+            llama_sampler_accept(smpl, tok);
+        }
+    });
+
+    printf("Testing backend penalties without filler entries\n");
+    compare_penalties_logits(params, 64, 1.1f, 0.5f, 0.25f, "Hello", [](llama_sampler * smpl) {
+        for (llama_token token = 0; token < 64; ++token) {
+            llama_sampler_accept(smpl, token);
+        }
+    });
+
+    printf("Testing backend top-k followed by penalties\n");
+    compare_top_k_penalties_logits(params, 8, 64, 1.1f, 0.5f, 0.25f, "Hello",
+            penalties_position::after_filter);
+
+    printf("Testing backend penalties followed by top-k\n");
+    compare_top_k_penalties_logits(params, 8, 64, 1.1f, 0.5f, 0.25f, "Hello",
+            penalties_position::before_filter);
+
+    printf("Testing backend top-p followed by penalties\n");
+    compare_masking_penalties_logits(params, "top-p", []() {
+        return llama_sampler_init_top_p(0.9f, 0);
+    }, 64, 1.1f, 0.5f, 0.25f, "Hello", penalties_position::after_filter, true);
+
+    printf("Testing backend top-p followed by penalties with a large history window\n");
+    compare_masking_penalties_logits(params, "top-p large-window", []() {
+        return llama_sampler_init_top_p(0.9f, 0);
+    }, 4096, 1.1f, 0.5f, 0.25f, "Hello", penalties_position::after_filter, true);
+
+    printf("Testing backend penalties followed by top-p\n");
+    compare_masking_penalties_logits(params, "top-p", []() {
+        return llama_sampler_init_top_p(0.9f, 0);
+    }, 64, 1.1f, 0.5f, 0.25f, "Hello", penalties_position::before_filter, true);
+
+    printf("Testing backend min-p followed by penalties\n");
+    compare_masking_penalties_logits(params, "min-p", []() {
+        return llama_sampler_init_min_p(0.1f, 0);
+    }, 64, 1.1f, 0.5f, 0.25f, "Hello", penalties_position::after_filter, false);
+
+    printf("Testing backend penalties followed by min-p\n");
+    compare_masking_penalties_logits(params, "min-p", []() {
+        return llama_sampler_init_min_p(0.1f, 0);
+    }, 64, 1.1f, 0.5f, 0.25f, "Hello", penalties_position::before_filter, false);
+
+    printf("Testing backend top-p followed by penalties with empty history\n");
+    compare_masking_penalties_logits(params, "top-p empty", []() {
+        return llama_sampler_init_top_p(0.9f, 0);
+    }, 64, 1.1f, 0.5f, 0.25f, "Hello", penalties_position::after_filter, true, false);
+
+    printf("Testing backend top-p followed by individual penalties\n");
+    compare_masking_penalties_logits(params, "top-p repeat", []() {
+        return llama_sampler_init_top_p(0.9f, 0);
+    }, 64, 1.1f, 0.0f, 0.0f, "Hello", penalties_position::after_filter, true);
+    compare_masking_penalties_logits(params, "top-p frequency", []() {
+        return llama_sampler_init_top_p(0.9f, 0);
+    }, 64, 1.0f, 0.5f, 0.0f, "Hello", penalties_position::after_filter, true);
+    compare_masking_penalties_logits(params, "top-p presence", []() {
+        return llama_sampler_init_top_p(0.9f, 0);
+    }, 64, 1.0f, 0.0f, 0.25f, "Hello", penalties_position::after_filter, true);
+
+    printf("Testing backend penalty parameter values\n");
+    test_penalty_parameter_values(params);
+
+    printf("backend penalties sampling test PASSED\n");
+}
+
 // This test verifies that it is possible to have two different backend samplers,
 // one that uses the backend dist sampler, and another that uses CPU dist sampler.
 static void test_backend_mixed_sampling(const test_params & params) {
@@ -1014,6 +1574,7 @@ struct backend_test_case {
 static const backend_test_case BACKEND_TESTS[] = {
     { "greedy",          test_backend_greedy_sampling,         true  },
     { "logit_bias",      test_backend_logit_bias_sampling,     true  },
+    { "penalties",       test_backend_penalties_sampling,      true  },
     { "temp",            test_backend_temp_sampling,           true  },
     { "temp_ext",        test_backend_temp_ext_sampling,       true  },
     { "top_k",           test_backend_top_k_sampling,          true  },

tests/test-model-resolution.cpp

@@ -0,0 +1,506 @@
+// tests the HF model resolution and the model handler assembly end-to-end on
+// synthetic repo listings: a local httplib server bound to the loopback
+// serves hardcoded HF API responses, so the real client, hf_cache, resolution
+// and CLI parsing run against them without external network access
+
+#include "arg.h"
+#include "common.h"
+#include "download.h"
+#include "http.h"
+#include "log.h"
+
+#include <nlohmann/json.hpp>
+
+#include <algorithm>
+#include <cstdio>
+#include <cstdlib>
+#include <filesystem>
+#include <map>
+#include <thread>
+#include <string>
+#include <vector>
+
+// the case and reordering being checked, printed with every failure
+static std::string g_context;
+
+// independent of NDEBUG, so the checks stay alive in Release builds
+#define REQUIRE(x) do {                                                         \
+    if (!(x)) {                                                                 \
+        fprintf(stderr, "%s:%d: [%s] REQUIRE(%s) failed\n",                     \
+                __FILE__, __LINE__, g_context.c_str(), #x);                     \
+        std::abort();                                                           \
+    }                                                                           \
+} while (0)
+
+#define REQUIRE_EQ(actual, expected) do {                                       \
+    if (!((actual) == (expected))) {                                            \
+        fprintf(stderr, "%s:%d: [%s] REQUIRE_EQ(%s, %s) failed\n  actual:   '%s'\n  expected: '%s'\n", \
+                __FILE__, __LINE__, g_context.c_str(), #actual, #expected,      \
+                std::string(actual).c_str(), std::string(expected).c_str());    \
+        std::abort();                                                           \
+    }                                                                           \
+} while (0)
+
+//
+// synthetic repos keyed by repo id, served over the loopback by a real
+// httplib server, so the tested code runs its own client and transport
+//
+
+static std::map<std::string, std::vector<std::string>> g_repos;
+
+static const char * COMMIT = "aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa";
+
+// the server lives in main, so its destructor runs before the static teardown
+// tears down the winsock state httplib brings in
+static void serve_repos(httplib::Server & server) {
+    server.Get(R"(/api/models/(.+)/refs)", [](const httplib::Request & req, httplib::Response & res) {
+        if (g_repos.count(req.matches[1])) {
+            res.set_content(nlohmann::json{{"branches", {{{"name", "main"}, {"targetCommit", COMMIT}}}}}.dump(),
+                            "application/json");
+        } else {
+            res.status = 404;
+        }
+    });
+    server.Get(R"(/api/models/(.+)/tree/.+)", [](const httplib::Request & req, httplib::Response & res) {
+        if (!g_repos.count(req.matches[1])) {
+            res.status = 404;
+            return;
+        }
+        auto files = nlohmann::json::array();
+        size_t i = 0;
+        for (const auto & p : g_repos[req.matches[1]]) {
+            char oid[41];
+            snprintf(oid, sizeof(oid), "%040lx", (unsigned long) ++i);
+            files.push_back({{"type", "file"}, {"path", p}, {"size", 1}, {"oid", oid}});
+        }
+        res.set_content(files.dump(), "application/json");
+    });
+}
+
+static common_params_model model_ref(const std::string & hf_repo, const std::string & hf_file = "") {
+    common_params_model m;
+    m.hf_repo = hf_repo;
+    m.hf_file = hf_file;
+    return m;
+}
+
+// the model cache is isolated under a temporary directory named after the
+// loopback port, so concurrent runs on a shared machine keep their own, and
+// the local path the handler wires for a file is snapshots/<commit>/<path>
+static std::filesystem::path cache_dir;
+
+static std::string cached(std::string repo_id, const std::string & path) {
+    string_replace_all(repo_id, "/", "--");
+    return (cache_dir / ("models--" + repo_id) / "snapshots" / COMMIT / path).string();
+}
+
+//
+// fixtures mimicking real repo layouts
+//
+
+// flat layout in the style of ggml-org/gemma-4-31B-it-GGUF
+static const std::vector<std::string> flat = {
+    "README.md",
+    "model-BF16.gguf",
+    "model-Q4_K_M.gguf",
+    "model-Q8_0.gguf",
+    "mmproj-model-BF16.gguf",
+    "mmproj-model-Q8_0.gguf",
+    "mtp-model-BF16.gguf",
+    "mtp-model-Q4_0.gguf",
+    "mtp-model-Q8_0.gguf",
+    "dflash-model-BF16.gguf",
+    "dflash-model-Q8_0.gguf",
+};
+
+// quants in subdirectories with sharded files and root sidecars,
+// in the style of stepfun-ai/Step-3.7-Flash-GGUF
+static const std::vector<std::string> subdir = {
+    "mmproj-model-f16.gguf",
+    "model-mtp-BF16.gguf",
+    "model-mtp-Q8_0.gguf",
+    "Q3_K_M/model-Q3_K_M-00001-of-00003.gguf",
+    "Q3_K_M/model-Q3_K_M-00002-of-00003.gguf",
+    "Q3_K_M/model-Q3_K_M-00003-of-00003.gguf",
+    "Q8_0/model-Q8_0-00001-of-00002.gguf",
+    "Q8_0/model-Q8_0-00002-of-00002.gguf",
+};
+
+// sidecar quants exist where the full model quant does not,
+// in the style of ggml-org/Qwen3.6-27B-GGUF
+static const std::vector<std::string> hole = {
+    "model-BF16.gguf",
+    "model-Q4_K_M.gguf",
+    "model-Q8_0.gguf",
+    "mtp-model-BF16.gguf",
+    "mtp-model-Q4_0.gguf",
+    "mtp-model-Q8_0.gguf",
+    "dflash-model-BF16.gguf",
+    "dflash-model-Q8_0.gguf",
+};
+
+// unsloth-style naming with UD quants and a suffix MTP file
+static const std::vector<std::string> unsloth = {
+    "model-UD-Q8_K_XL.gguf",
+    "mmproj-BF16.gguf",
+    "model-MTP-BF16.gguf",
+};
+
+// bartowski-style vendor prefix and mradermacher-style dot quant
+static const std::vector<std::string> vendors = {
+    "TheDrummer_Model-24B-v4.1-Q8_0.gguf",
+    "BlackSheep-24B.Q8_0.gguf",
+};
+
+// every speculative sidecar type at the same quant
+static const std::vector<std::string> quad = {
+    "model-Q8_0.gguf",
+    "mtp-model-Q8_0.gguf",
+    "dflash-model-Q8_0.gguf",
+    "eagle3-model-Q8_0.gguf",
+    "dspark-model-Q8_0.gguf",
+};
+
+static const std::vector<std::string> dflash_only = {
+    "model-Q8_0.gguf",
+    "dflash-model-Q8_0.gguf",
+};
+
+static const std::vector<std::string> eagle3_only = {
+    "model-Q8_0.gguf",
+    "eagle3-model-Q8_0.gguf",
+};
+
+// a single full quant with dspark sidecars at other quants,
+// in the style of ggml-org/DeepSeek-V4-Flash-0731-GGUF
+static const std::vector<std::string> spark = {
+    "README.md",
+    "model-MXFP4.gguf",
+    "dspark-model-BF16.gguf",
+    "dspark-model-MXFP4.gguf",
+};
+
+// dspark outranks dflash in the type auto-selection
+static const std::vector<std::string> dspark_dflash = {
+    "model-Q8_0.gguf",
+    "dflash-model-Q8_0.gguf",
+    "dspark-model-Q8_0.gguf",
+};
+
+//
+// table-driven plan resolution through the real entry point,
+// each case replayed on multiple deterministic reorderings of the listing,
+// except the cases whose pick legitimately depends on the listing order
+//
+
+struct plan_case {
+    const char * name;
+    const std::vector<std::string> & files;
+    const char * hf_repo;
+    const char * hf_file;
+    bool sidecars;        // request mmproj + mtp + dflash + eagle3 + dspark
+    bool order_dependent; // the expected pick depends on the listing order
+    const char * primary;
+    std::vector<std::string> model_files;
+    const char * mmproj;
+    const char * mtp;
+    const char * dflash;
+    const char * eagle3;
+    const char * dspark;
+};
+
+static const plan_case plan_cases[] = {
+    // exact tag picks the matching primary, sidecars follow the tag
+    {"flat exact tag", flat, "test/repo:Q8_0", "", true, false,
+     "model-Q8_0.gguf", {"model-Q8_0.gguf"},
+     "mmproj-model-Q8_0.gguf", "mtp-model-Q8_0.gguf", "dflash-model-Q8_0.gguf", "", ""},
+
+    // no tag falls back to the default quant preference
+    {"flat default", flat, "test/repo", "", false, false,
+     "model-Q4_K_M.gguf", {"model-Q4_K_M.gguf"},
+     "", "", "", "", ""},
+
+    // no tag and no default match falls back to the first model in the listing
+    {"unsloth fallback", unsloth, "test/repo", "", true, true,
+     "model-UD-Q8_K_XL.gguf", {"model-UD-Q8_K_XL.gguf"},
+     "mmproj-BF16.gguf", "", "", "", ""},
+
+    // explicit hf_file picks that exact file
+    {"flat hf_file", flat, "test/repo", "model-BF16.gguf", false, false,
+     "model-BF16.gguf", {"model-BF16.gguf"},
+     "", "", "", "", ""},
+
+    // missing hf_file resolves nothing
+    {"flat missing hf_file", flat, "test/repo", "nope.gguf", false, false,
+     "", {},
+     "", "", "", "", ""},
+
+    // a sharded primary brings all its parts, a subdir primary finds the root sidecar
+    {"subdir shards", subdir, "test/repo:Q3_K_M", "", true, false,
+     "Q3_K_M/model-Q3_K_M-00001-of-00003.gguf",
+     {"Q3_K_M/model-Q3_K_M-00001-of-00003.gguf",
+      "Q3_K_M/model-Q3_K_M-00002-of-00003.gguf",
+      "Q3_K_M/model-Q3_K_M-00003-of-00003.gguf"},
+     "mmproj-model-f16.gguf", "model-mtp-Q8_0.gguf", "", "", ""},
+
+    // a tag with no matching full model still resolves the requested sidecars
+    {"hole tag sidecar", hole, "test/repo:Q4_0", "", true, false,
+     "", {},
+     "", "mtp-model-Q4_0.gguf", "dflash-model-Q8_0.gguf", "", ""},
+
+    // the same tag without a requested sidecar resolves nothing
+    {"hole tag alone", hole, "test/repo:Q4_0", "", false, false,
+     "", {},
+     "", "", "", "", ""},
+
+    // no tag anchors the sidecars on the primary quant
+    {"hole default anchor", hole, "test/repo", "", true, false,
+     "model-Q4_K_M.gguf", {"model-Q4_K_M.gguf"},
+     "", "mtp-model-Q4_0.gguf", "dflash-model-Q8_0.gguf", "", ""},
+
+    // the mtp- keyword is case sensitive, a suffix -MTP file is not discovered
+    {"unsloth suffix mtp", unsloth, "test/repo:Q8_K_XL", "", true, false,
+     "model-UD-Q8_K_XL.gguf", {"model-UD-Q8_K_XL.gguf"},
+     "mmproj-BF16.gguf", "", "", "", ""},
+
+    // vendor prefixes and the dot quant convention both match the tag,
+    // first match wins between two files at the same quant
+    {"vendor prefix", vendors, "test/repo:Q8_0", "", false, true,
+     "TheDrummer_Model-24B-v4.1-Q8_0.gguf", {"TheDrummer_Model-24B-v4.1-Q8_0.gguf"},
+     "", "", "", "", ""},
+
+    // every sidecar type resolves at the tag
+    {"quad exact tag", quad, "test/repo:Q8_0", "", true, false,
+     "model-Q8_0.gguf", {"model-Q8_0.gguf"},
+     "", "mtp-model-Q8_0.gguf", "dflash-model-Q8_0.gguf", "eagle3-model-Q8_0.gguf", "dspark-model-Q8_0.gguf"},
+
+    // no tag anchors the dspark sidecar on the only full quant
+    {"spark default anchor", spark, "test/repo", "", true, false,
+     "model-MXFP4.gguf", {"model-MXFP4.gguf"},
+     "", "", "", "", "dspark-model-MXFP4.gguf"},
+
+    // a tag with no matching full model still resolves the exact dspark sidecar
+    {"spark tag sidecar", spark, "test/repo:BF16", "", true, false,
+     "", {},
+     "", "", "", "", "dspark-model-BF16.gguf"},
+};
+
+static void check_plan(const plan_case & c) {
+    common_download_opts opts;
+    opts.download_mmproj = c.sidecars;
+    opts.download_mtp    = c.sidecars;
+    opts.download_dflash = c.sidecars;
+    opts.download_eagle3 = c.sidecars;
+    opts.download_dspark = c.sidecars;
+
+    auto plan = common_download_get_hf_plan(model_ref(c.hf_repo, c.hf_file), opts);
+
+    REQUIRE_EQ(plan.primary.path, c.primary);
+    REQUIRE_EQ(plan.mmproj.path,  c.mmproj);
+    REQUIRE_EQ(plan.mtp.path,     c.mtp);
+    REQUIRE_EQ(plan.dflash.path,  c.dflash);
+    REQUIRE_EQ(plan.eagle3.path,  c.eagle3);
+    REQUIRE_EQ(plan.dspark.path,  c.dspark);
+
+    // exact shard set, order insensitive; the primary must be the first split
+    std::vector<std::string> actual;
+    for (const auto & f : plan.model_files) {
+        actual.push_back(f.path);
+    }
+    std::sort(actual.begin(), actual.end());
+    auto expected = c.model_files;
+    std::sort(expected.begin(), expected.end());
+    REQUIRE(actual == expected);
+    if (!expected.empty()) {
+        REQUIRE(plan.primary.path == expected.front());
+    }
+}
+
+static void test_plan_resolution() {
+    printf("test-model-resolution: plan resolution on %zu cases\n", sizeof(plan_cases) / sizeof(plan_cases[0]));
+
+    for (const auto & c : plan_cases) {
+        printf("  %s\n", c.name);
+        // invariant: the resolution is insensitive to the listing order
+        for (size_t rot = 0; rot < c.files.size(); ++rot) {
+            if (c.order_dependent && rot > 0) {
+                continue;
+            }
+            g_context = std::string(c.name) + ", reordering " + std::to_string(rot);
+            auto files = c.files;
+            std::rotate(files.begin(), files.begin() + rot, files.end());
+            if (rot % 2 == 1) {
+                std::reverse(files.begin(), files.end());
+            }
+            g_repos["test/repo"] = files;
+            check_plan(c);
+        }
+    }
+    g_repos.clear();
+}
+
+//
+// end-to-end assembly: real CLI parsing, real handler init resolving over the
+// loopback, downloads skipped by flipping offline before apply
+//
+
+static void assemble(std::vector<std::string> argv, common_params & params) {
+    std::vector<char *> cargv;
+    g_context.clear();
+    for (auto & a : argv) {
+        g_context += g_context.empty() ? a : " " + a;
+        cargv.push_back(a.data());
+    }
+    bool ok = common_params_parse((int) cargv.size(), cargv.data(), params, LLAMA_EXAMPLE_SERVER);
+    REQUIRE(ok);
+
+    auto handler = common_models_handler_init(params, LLAMA_EXAMPLE_SERVER);
+
+    // skip the network execution, on_done still wires the params
+    params.offline = true;
+    common_models_handler_apply(handler, params);
+}
+
+static void test_task_assembly() {
+    printf("test-model-resolution: end-to-end assembly\n");
+
+    g_repos["test/main"]   = flat;
+    g_repos["test/hole"]   = hole;
+    g_repos["test/quad"]   = quad;
+    g_repos["test/dflash"] = dflash_only;
+    g_repos["test/eagle3"] = eagle3_only;
+    g_repos["test/spark"]  = spark;
+    g_repos["test/pair"]   = dspark_dflash;
+    g_repos["test/small"]  = {"draft-model-Q4_K_M.gguf"};
+    g_repos["test/preset"] = {"preset.ini", "model-Q8_0.gguf"};
+
+    {
+        // plain -hf wires the model and its mmproj, nothing speculative
+        common_params params;
+        assemble({"server", "-hf", "test/main:Q8_0"}, params);
+        REQUIRE_EQ(params.model.path,  cached("test/main", "model-Q8_0.gguf"));
+        REQUIRE_EQ(params.mmproj.path, cached("test/main", "mmproj-model-Q8_0.gguf"));
+        REQUIRE(params.speculative.draft.mparams.path.empty());
+    }
+    {
+        // --no-mmproj disables the mmproj discovery
+        common_params params;
+        assemble({"server", "-hf", "test/main:Q8_0", "--no-mmproj"}, params);
+        REQUIRE(params.mmproj.path.empty());
+    }
+    {
+        // an explicit --mmproj wins over the discovery
+        common_params params;
+        assemble({"server", "-hf", "test/main:Q8_0", "--mmproj", "/local/mmproj.gguf"}, params);
+        REQUIRE(params.mmproj.path == "/local/mmproj.gguf");
+    }
+    {
+        // -hf with a spec type wires the sidecar of the main repo as fallback draft
+        common_params params;
+        assemble({"server", "-hf", "test/main:Q8_0", "--spec-type", "draft-mtp"}, params);
+        REQUIRE_EQ(params.speculative.draft.mparams.path, cached("test/main", "mtp-model-Q8_0.gguf"));
+    }
+    {
+        // -hfd with a spec type wires the draft repo sidecar at its tag,
+        // not its full model, and suppresses the main repo fallback
+        common_params params;
+        assemble({"server", "-hf", "test/hole:Q8_0", "-hfd", "test/hole:Q4_0", "--spec-type", "draft-mtp"}, params);
+        REQUIRE_EQ(params.speculative.draft.mparams.path, cached("test/hole", "mtp-model-Q4_0.gguf"));
+    }
+    {
+        // an explicit -md file wins over the sidecar resolution
+        common_params params;
+        assemble({"server", "-hf", "test/main:Q8_0", "-hfd", "test/main", "-md", "mtp-model-BF16.gguf", "--spec-type", "draft-mtp"}, params);
+        REQUIRE_EQ(params.speculative.draft.mparams.path, cached("test/main", "mtp-model-BF16.gguf"));
+    }
+    {
+        // -hfd without a spec type auto-selects the type, mtp first when all ship
+        common_params params;
+        assemble({"server", "-hf", "test/main:Q8_0", "-hfd", "test/quad:Q8_0"}, params);
+        REQUIRE(params.speculative.types == std::vector<enum common_speculative_type>{COMMON_SPECULATIVE_TYPE_DRAFT_MTP});
+        REQUIRE_EQ(params.speculative.draft.mparams.path, cached("test/quad", "mtp-model-Q8_0.gguf"));
+    }
+    {
+        // auto-selection with only a dflash sidecar
+        common_params params;
+        assemble({"server", "-hf", "test/main:Q8_0", "-hfd", "test/dflash:Q8_0"}, params);
+        REQUIRE(params.speculative.types == std::vector<enum common_speculative_type>{COMMON_SPECULATIVE_TYPE_DRAFT_DFLASH});
+        REQUIRE_EQ(params.speculative.draft.mparams.path, cached("test/dflash", "dflash-model-Q8_0.gguf"));
+    }
+    {
+        // auto-selection with only an eagle3 sidecar
+        common_params params;
+        assemble({"server", "-hf", "test/main:Q8_0", "-hfd", "test/eagle3:Q8_0"}, params);
+        REQUIRE(params.speculative.types == std::vector<enum common_speculative_type>{COMMON_SPECULATIVE_TYPE_DRAFT_EAGLE3});
+        REQUIRE_EQ(params.speculative.draft.mparams.path, cached("test/eagle3", "eagle3-model-Q8_0.gguf"));
+    }
+    {
+        // auto-selection prefers dspark over dflash when both ship
+        common_params params;
+        assemble({"server", "-hf", "test/main:Q8_0", "-hfd", "test/pair:Q8_0"}, params);
+        REQUIRE(params.speculative.types == std::vector<enum common_speculative_type>{COMMON_SPECULATIVE_TYPE_DRAFT_DSPARK});
+        REQUIRE_EQ(params.speculative.draft.mparams.path, cached("test/pair", "dspark-model-Q8_0.gguf"));
+    }
+    {
+        // -hf with the dspark spec type wires the sidecar of the main repo,
+        // anchored on the only full quant
+        common_params params;
+        assemble({"server", "-hf", "test/spark", "--spec-type", "draft-dspark"}, params);
+        REQUIRE_EQ(params.model.path, cached("test/spark", "model-MXFP4.gguf"));
+        REQUIRE_EQ(params.speculative.draft.mparams.path, cached("test/spark", "dspark-model-MXFP4.gguf"));
+    }
+    {
+        // -hfd on a repo without sidecars keeps resolving a full model as draft
+        common_params params;
+        assemble({"server", "-hf", "test/main:Q8_0", "-hfd", "test/small"}, params);
+        REQUIRE(params.speculative.types == std::vector<enum common_speculative_type>{COMMON_SPECULATIVE_TYPE_NONE});
+        REQUIRE_EQ(params.speculative.draft.mparams.path, cached("test/small", "draft-model-Q4_K_M.gguf"));
+    }
+    {
+        // a preset repo wires the preset and clears the model for router mode
+        common_params params;
+        assemble({"server", "-hf", "test/preset"}, params);
+        REQUIRE_EQ(params.models_preset, cached("test/preset", "preset.ini"));
+        REQUIRE(params.model.path.empty());
+        REQUIRE(params.model.hf_repo.empty());
+    }
+
+    g_repos.clear();
+}
+
+int main(void) {
+    // unbuffered, so a crash cannot swallow the reports already printed
+    setvbuf(stdout, nullptr, _IONBF, 0);
+    setvbuf(stderr, nullptr, _IONBF, 0);
+
+    // the negative cases legitimately log errors on every reordering,
+    // keep the output down to the reports
+    common_log_pause(common_log_main());
+
+    // the loopback endpoint also keeps the client init from rejecting
+    // https on the builds without TLS support
+    httplib::Server server;
+    serve_repos(server);
+    int port = server.bind_to_any_port("127.0.0.1");
+
+    // isolate the cache, its location is read once so it is set
+    // before anything else
+    cache_dir = std::filesystem::temp_directory_path() /
+                ("test-model-resolution-cache-" + std::to_string(port));
+    std::filesystem::remove_all(cache_dir);
+    common_set_env("LLAMA_CACHE", cache_dir.string());
+
+    std::thread server_thread([&server] { server.listen_after_bind(); });
+    server.wait_until_ready();
+    common_set_env("MODEL_ENDPOINT", "http://127.0.0.1:" + std::to_string(port) + "/");
+
+    test_plan_resolution();
+    test_task_assembly();
+
+    server.stop();
+    server_thread.join();
+
+    std::filesystem::remove_all(cache_dir);
+    printf("test-model-resolution: all tests OK\n");
+    return 0;
+}

tools/server/README-dev.md

@@ -199,6 +199,9 @@ Invoke a tool call, request body is a JSON object with:
 - `tool` (string): the name of the tool
 - `params` (object): a mapping from argument name (string) to argument value
 
+Headers:
+- `x-tool-cwd`: optional; if set, use as the CWD for tool; this is not part of tool's params because it's meant to be set by the runtime, not the LLM itself
+
 Returns JSON object. There are two response formats (MCP tools use the same two formats: their result content is concatenated into `plain_text_response`, and RPC or tool errors are surfaced as the `error` string):
 
 Format 1: Plain text. The text will be placed into a field called `plain_text_response`, example:

tools/server/README.md

@@ -198,7 +198,9 @@ For the full list of features, please refer to [server's changelog](https://gith
 | `--ui-config, --webui-config JSON` | JSON that provides default UI settings (overrides UI defaults)<br/>(env: LLAMA_ARG_UI_CONFIG) |
 | `--ui-config-file, --webui-config-file PATH` | JSON file that provides default UI settings (overrides UI defaults)<br/>(env: LLAMA_ARG_UI_CONFIG_FILE) |
 | `--ui-mcp-proxy, --webui-mcp-proxy, --no-ui-mcp-proxy, --no-webui-mcp-proxy` | experimental: whether to enable MCP CORS proxy - do not enable in untrusted environments (default: disabled)<br/>(env: LLAMA_ARG_UI_MCP_PROXY) |
-| `--tools TOOL1,TOOL2,...` | experimental: whether to enable built-in tools for AI agents - do not enable in untrusted environments (default: no tools)<br/>specify "all" to enable all tools<br/>available tools: read_file, file_glob_search, grep_search, exec_shell_command, write_file, edit_file, get_datetime<br/>note: for security reasons, this will limit --cors-origins to localhost by default<br/>(env: LLAMA_ARG_TOOLS) |
+| `--tools TOOL1,TOOL2,...` | experimental: whether to enable built-in tools for AI agents - do not enable in untrusted environments (default: no tools)<br/>specify "all" to enable all tools<br/>available tools: read_file, file_glob_search, grep_search, exec_shell_command, write_file, edit_file, get_datetime, get_info<br/>note: for security reasons, this will limit --cors-origins to localhost by default<br/>(env: LLAMA_ARG_TOOLS) |
+| `--mcp-servers-config PATH` | experimental: path to JSON file with MCP server definitions (Cursor-compatible format) - do not enable in untrusted environments (default: none)<br/>note: for security reasons, this will limit --cors-origins to localhost by default<br/>(env: LLAMA_ARG_MCP_SERVERS_CONFIG) |
+| `--mcp-servers-json JSON` | experimental: inline JSON with MCP server definitions (Cursor-compatible format) - do not enable in untrusted environments (default: none)<br/>note: for security reasons, this will limit --cors-origins to localhost by default<br/>(env: LLAMA_ARG_MCP_SERVERS_JSON) |
 | `-ag, --agent, -no-ag, --no-agent` | whether to enable CORS proxy and all built-in tools - do not enable in untrusted environments (default: disabled)<br/>note: for security reasons, this will limit --cors-origins to localhost by default<br/>(env: LLAMA_ARG_AGENT) |
 | `--ui, --webui, --no-ui, --no-webui` | whether to enable the Web UI (default: enabled)<br/>(env: LLAMA_ARG_UI) |
 | `--embedding, --embeddings` | restrict to only support embedding use case; use only with dedicated embedding models (default: disabled)<br/>(env: LLAMA_ARG_EMBEDDINGS) |

tools/server/server-context.cpp

@@ -1807,7 +1807,8 @@ private:
         // initialize samplers
         if (task.need_sampling()) {
             try {
-                slot.smpl.reset(common_sampler_init(model_tgt, task.params.sampling));
+                slot.smpl.reset(common_sampler_init(
+                        model_tgt, task.params.sampling, (int32_t) llama_n_ctx(ctx_tgt)));
             } catch (std::exception & e) {
                 std::string err_msg = std::string("Failed to initialize samplers: ") + e.what();
                 send_error(task, err_msg, ERROR_TYPE_INVALID_REQUEST);

tools/server/server-tools.cpp

@@ -64,24 +64,27 @@ public:
 
 class tools_io_basic : public tools_io {
 public:
+    // cwd, if non-empty, is used to resolve relative paths and as the working directory for run()
+    explicit tools_io_basic(std::string cwd = "") : cwd(std::move(cwd)) {}
+
     bool is_directory(const std::string & path) const override {
         std::error_code ec;
-        return fs::is_directory(path, ec) && !ec;
+        return fs::is_directory(resolve(path), ec) && !ec;
     }
 
     bool is_regular_file(const std::string & path) const override {
         std::error_code ec;
-        return fs::is_regular_file(path, ec) && !ec;
+        return fs::is_regular_file(resolve(path), ec) && !ec;
     }
 
     bool file_size(const std::string & path, uintmax_t & out_size) const override {
         std::error_code ec;
-        out_size = fs::file_size(path, ec);
+        out_size = fs::file_size(resolve(path), ec);
         return !ec;
     }
 
     bool read_file(const std::string & path, std::string & out) const override {
-        std::ifstream f(path, std::ios::binary);
+        std::ifstream f(resolve(path), std::ios::binary);
         if (!f) return false;
         std::ostringstream ss;
         ss << f.rdbuf();
@@ -91,12 +94,12 @@ public:
 
     bool write_file(const std::string & path, const std::string & content) const override {
         std::error_code ec;
-        fs::path fpath(path);
+        fs::path fpath(resolve(path));
         if (fpath.has_parent_path()) {
             fs::create_directories(fpath.parent_path(), ec);
             if (ec) return false;
         }
-        std::ofstream f(path, std::ios::binary);
+        std::ofstream f(fpath, std::ios::binary);
         if (!f) return false;
         f << content;
         return (bool) f;
@@ -104,13 +107,14 @@ public:
 
     std::vector<std::string> list_files(const std::string & base, std::string & err) const override {
         err.clear();
+        std::string abs_base = resolve(base);
         if (!is_directory(base)) {
             err = "path does not exist or is not a directory: " + base;
             return {};
         }
 
         auto res = run(
-            {"git", "-C", base, "ls-files", "--cached", "--others", "--exclude-standard"},
+            {"git", "-C", abs_base, "ls-files", "--cached", "--others", "--exclude-standard"},
             SERVER_TOOL_GIT_LS_FILES_MAX_OUTPUT, SERVER_TOOL_GIT_LS_FILES_TIMEOUT);
 
         if (res.exit_code == 0 && !res.timed_out) {
@@ -128,7 +132,7 @@ public:
             return result;
         }
 
-        return list_files_fallback(base);
+        return list_files_fallback(abs_base);
     }
 
     exec_result run(
@@ -145,7 +149,7 @@ public:
                     | subprocess_option_inherit_environment
                     | subprocess_option_search_user_path;
 
-        if (!proc.create(args, options)) {
+        if (!proc.create(args, options, {}, cwd.empty() ? nullptr : cwd.c_str())) {
             res.output = "failed to spawn process";
             return res;
         }
@@ -205,6 +209,16 @@ public:
     }
 
 private:
+    std::string cwd;
+
+    // resolves `path` against `cwd` if `path` is relative and `cwd` is set; otherwise returns `path` unchanged
+    std::string resolve(const std::string & path) const {
+        if (cwd.empty() || fs::path(path).is_absolute()) {
+            return path;
+        }
+        return (fs::path(cwd) / path).string();
+    }
+
     static const std::unordered_set<std::string> & junk_dir_names() {
         static const std::unordered_set<std::string> names = {
             ".git", ".svn", ".hg", "node_modules", "__pycache__",
@@ -244,8 +258,8 @@ private:
 };
 
 static std::unique_ptr<tools_io> make_tools_io(const json & params) {
-    GGML_UNUSED(params); // TODO in follow-up PR
-    return std::make_unique<tools_io_basic>();
+    std::string cwd = json_value(params, "cwd", std::string());
+    return std::make_unique<tools_io_basic>(cwd);
 }
 
 // no '/' in pattern -> match basename at any depth; else match full relative path
@@ -1076,6 +1090,56 @@ struct server_tool_get_datetime : server_tool {
     }
 };
 
+//
+// get_info: returns runtime info (OS name/version and cwd)
+//
+
+struct server_tool_get_info : server_tool {
+    server_tool_get_info() {
+        name = "get_info";
+        display_name = "Get Runtime Info";
+        permission_write = false;
+    }
+
+    json get_definition() const override {
+        return {
+            {"type", "function"},
+            {"function", {
+                {"name", name},
+                {"description", "Returns runtime info: the OS name/version and the current working directory"},
+                {"parameters", {
+                    {"type", "object"},
+                    {"properties", json::object()},
+                }},
+            }},
+        };
+    }
+
+    json invoke(json params, server_tool::stream *) const override {
+        auto io = make_tools_io(params);
+
+#ifdef _WIN32
+        auto res = io->run({"cmd", "/c", "ver"}, 4096, 5);
+#else
+        auto res = io->run({"uname", "-a"}, 4096, 5);
+#endif
+        // "ver" prints a blank line before the version, so the output is stripped on both ends;
+        // a failed spawn or a timeout leaves a diagnostic in res.output, which is not an OS name
+        std::string os_info = res.exit_code == 0 && !res.timed_out ? string_strip(res.output) : "unknown";
+
+        std::string cwd = json_value(params, "cwd", std::string());
+        if (cwd.empty()) {
+            std::error_code ec;
+            cwd = fs::current_path(ec).string();
+        }
+
+        return {
+            {"os",  os_info},
+            {"cwd", cwd},
+        };
+    }
+};
+
 struct server_tool_stream_result : server_task_result {
     std::string chunk;
     bool done = false;
@@ -1185,9 +1249,26 @@ static std::vector<std::unique_ptr<server_tool>> build_tools() {
     tools.push_back(std::make_unique<server_tool_write_file>());
     tools.push_back(std::make_unique<server_tool_edit_file>());
     tools.push_back(std::make_unique<server_tool_get_datetime>());
+    tools.push_back(std::make_unique<server_tool_get_info>());
     return tools;
 }
 
+static std::string str_to_lower(const std::string & value) {
+    std::string lowered(value.size(), '\0');
+    std::transform(value.begin(), value.end(), lowered.begin(), [](unsigned char c) { return std::tolower(c); });
+    return lowered;
+}
+
+static std::string get_header(const std::map<std::string, std::string> & headers, const std::string & key, std::string default_value = "") {
+    const auto lowered_key = str_to_lower(key);
+    for (const auto & h : headers) {
+        if (str_to_lower(h.first) == lowered_key) {
+            return h.second;
+        }
+    }
+    return default_value;
+}
+
 void server_tools::setup(const std::vector<std::string> & enabled_tools,
                          server_mcp & mcp_mgr) {
     if (!enabled_tools.empty()) {
@@ -1271,6 +1352,12 @@ void server_tools::setup(const std::vector<std::string> & enabled_tools,
             json params = body.value("params", json::object());
             bool stream = body.value("stream", false);
 
+            // accept x-tool-cwd header to override of the process
+            auto cwd = get_header(req.headers, "x-tool-cwd");
+            if (!cwd.empty()) {
+                params["cwd"] = cwd;
+            }
+
             server_tool & tool = find_tool(tools, tool_name, stream);
 
             if (stream) {

tools/server/server.cpp

@@ -486,6 +486,13 @@ int llama_server(common_params & params, int argc, char ** argv) {
 
     SRV_INF("listening on %s\n", ctx_http.listening_address.c_str());
 
+    // TODO: remove this in the future
+    // check the string to also handle the .sock case
+    if (string_ends_with(ctx_http.listening_address, ":8080")) {
+        SRV_WRN("%s", "NOTICE: server default port will be changed to :9931 in a future release\n");
+        SRV_WRN("%s", "        ref: https://github.com/ggml-org/llama.cpp/pull/26508\n");
+    }
+
     if (is_router_server) {
         if (!params.models_preset_hf.empty()) {
             SRV_WRN(      "NOTE: using preset.ini from HF repo '%s'\n", params.models_preset_hf.c_str());

tools/server/tests/unit/test_tools_builtin.py

@@ -19,8 +19,8 @@ def create_server():
     server.server_tools = "all"
 
 
-def call_tool(name: str, params: dict) -> dict:
-    res = server.make_request("POST", "/tools", data={"tool": name, "params": params})
+def call_tool(name: str, params: dict, headers: dict | None = None) -> dict:
+    res = server.make_request("POST", "/tools", data={"tool": name, "params": params}, headers=headers)
     assert res.status_code == 200, res.body
     assert "error" not in res.body, res.body
     return res.body
@@ -123,6 +123,29 @@ def test_tools_builtin_exec_shell_command_stream():
     assert "[exit code: 0]" in chunks
 
 
+def test_tools_builtin_cwd_header():
+    global server
+    server.start()
+
+    cwd_dir = os.path.join(PROJECT_ROOT, "tools", "server", "tests", "unit")
+    headers = {"x-tool-cwd": cwd_dir}
+
+    res = call_tool("read_file", {"path": "test_tools_builtin.py"}, headers=headers)
+    assert GREP_MARKER in res["plain_text_response"]
+
+    # exec_shell_command should also run with that directory as its working directory:
+    # writing to a relative filename must land inside cwd_dir
+    marker_name = "llama_cpp_test_tools_builtin_cwd_marker.txt"
+    marker_path = os.path.join(cwd_dir, marker_name)
+    try:
+        command = f"echo hello > {marker_name}"
+        call_tool("exec_shell_command", {"command": command}, headers=headers)
+        assert os.path.exists(marker_path)
+    finally:
+        if os.path.exists(marker_path):
+            os.remove(marker_path)
+
+
 def test_tools_builtin_edit_file_rejects_overlapping_edits():
     global server
     server.start()

vendor/cpp-httplib/CMakeLists.txt

@@ -41,7 +41,7 @@ if (LLAMA_BUILD_BORINGSSL)
     set(FIPS OFF CACHE BOOL "Enable FIPS (BoringSSL)")
 
     set(BORINGSSL_GIT "https://boringssl.googlesource.com/boringssl" CACHE STRING "BoringSSL git repository")
-    set(BORINGSSL_VERSION "0.20260730.0" CACHE STRING "BoringSSL version")
+    set(BORINGSSL_VERSION "0.20260803.0" CACHE STRING "BoringSSL version")
 
     message(STATUS "Fetching BoringSSL version ${BORINGSSL_VERSION}")