* adapt the api
* text model ok
* working impl, need verify and clean up
* mtmd: build the pocket-tts transposed convolutions as GEMM + col2im
ggml_conv_transpose_1d has no grouped mode, so the depthwise upsample
was built as one convolution and one concat per channel, which floods
the graph with small nodes and makes kernel launches dominate the
decoder.
Fold both cases into the column form the seanet decoder already needs:
the general case reshapes the kernel to [IC, K * OC] and matmuls it
with the input, the depthwise case batches a matmul over the channels
so a step scales its own kernel. A single col2im_1d then scatter-adds
the columns back to the signal, with the same shape as before, so the
overlap-add tail, the streaming state and the bias are untouched.
Generation time per frame drops by 80% on CUDA and by 50% on CPU. The
output matches the previous implementation sample for sample, with a
correlation of 0.999994 and identical frame counts.
* flow_temp + frames_after_eos
* chunking
* mtmd: carry the remaining pocket-tts per-pack settings
The language packs also tune the end-of-speech padding and the padding
of short prompts, next to the temperature already carried in the
mmproj: french_24l asks for 8 tail frames instead of the guessed 3,
english_2026-01 asks for short prompts to be padded with spaces.
Write both in the mmproj as clip.gen.audio.frames_after_eos and
clip.gen.audio.pad_short_text, keyed on the pack in the conversion
script like the temperature. The loader keeps them optional, so a
mmproj without them behaves as before. Map semicolons to commas for
every pack instead, the reference only asks for it on three of them and
it costs nothing elsewhere.
Existing mmproj files must be converted again to carry the two keys.
On a long french text the port now lands within 2% of the reference:
22.96s against 23.44s, with the same peak level and the same amount of
silence.
* clip.gen.audio.model_variant
* clean up code comments
* nit: drop the dead flow_temp hparam, the pack table holds the default
* update docs
* address security problems
* less invasive base.py
* lint
* add mtmd_gen_inp_default
* add docs
* rm gen_flow_temp
---------
Co-authored-by: Pascal <admin@serveurperso.com>
* llama: add new default load-mode auto which picks mmap unless a non-Metal iGPU is used
* Update ggml/src/ggml-hexagon/ggml-hexagon.cpp
Co-authored-by: Max Krasnyansky <maxk@qti.qualcomm.com>
* set mmap_support to false on OpenCL backend
* fix order of load modes
* use -1 for auto
* resolve load mode auto earlier to correctly pick gpu host or cpu memory
* add load mode auto to llama-bench
* bump virtgpu api version, regenerate docs
---------
Co-authored-by: Piotr Wilkin (ilintar) <piotr.wilkin@syndatis.com>
Co-authored-by: Max Krasnyansky <maxk@qti.qualcomm.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* model : fix SWA not being enabled for EXAONE 4.5
load_arch_hparams tests `hparams.n_layer() == 64` before
LLM_KV_NEXTN_PREDICT_LAYERS has been read. n_layer() returns
n_layer_all - n_layer_nextn and n_layer_nextn defaults to 0, so a GGUF
carrying the MTP head (block_count=65, nextn=1) evaluates to 65 and the
whole SWA block is skipped. The model type switch further down in the
same function reads 64, because by then the key has been loaded.
n_swa is still filled in by the unconditional get_key below the block, so
llama_model_n_swa() reports 4096 and the logs look correct while only
swa_type stays LLAMA_SWA_TYPE_NONE.
This affects the official LGAI-EXAONE GGUF release as well. EXAONE 4.0 has
no MTP head, so block_count is 64 there and the check matches.
* model-loader : skip TENSOR_SKIP tensors in the metadata-only path
create_tensor asserts on a null buffer type when building from metadata
alone, but buft_for_tensor returns null by design for tensors marked
TENSOR_SKIP, which is how architectures with nextn/MTP layers mark theirs.
Those models cannot be constructed by llama_model_init_from_user at all.
The file-backed path below already returns nullptr for the same tensors, so
callers see the same thing either way.
* tests : cover exaone4 hparams ordering
Builds a synthetic exaone4 model with the layout the shipped EXAONE 4.5
GGUFs use (block_count 65 + nextn 1). The swa_type check is the one that
catches the ordering bug; the n_layer_nextn and n_layer() checks only tell
a broken fixture apart from a real regression.
Fails before the ordering fix with "swa_type is not STANDARD", passes after.
* Revert "tests : cover exaone4 hparams ordering"
This reverts commit d2f3bafeee.
* Revert "model-loader : skip TENSOR_SKIP tensors in the metadata-only path"
This reverts commit aecb9bc0c7.
* Enable backend sampling with token speculation
* Clamp the mask sum before converting it into the sampled index
* Add a numeric context parameter declaring the maximum outputs one sequence
* More fixes
* Don't reuse memory for output views.
* Match dist between CPU and GPU
* Fix CPU and backend sampling mismatches
* Simpify some of the changes
* Fix tests on Vulkan
* More test fixes
* Rebase changes
* Rebase and address review comments
* Address review comments
* Address review comments
* Update src/llama-sampler.cpp
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Get started with Onyx
* Add architecture
* Skip keys handled in super()
* Loading tensors
* Shorten
* Graph
* Apply suggestion from @pcuenca
* Remove norm now embedding in transformers weights
* Add eot
* Explicit output_multiplier
* Handle post_norm_eps
* No super call; unhardcode eot.
The pattern `self._set_vocab_gpt2()` seems preferred throughout the
codebase, and it allows `set_vocab()` to be called from a different part
of the Python class hierarchy: the drafter model converter that we may
need eventually.
* Register for drafting
* DFlash: inherit rope type from the linked target.
Another option would be to store it in the gguf file itself.
* mmproj conversion
Note: some fields to be renamed after the implementation works. We are
keeping compatibility with the reference Meta gguf for testing purposes.
* "clip" header declarations
* Load mmproj
* Pre-processing
* Graph
* Go back to using delimiters.
Otherwise our generations are worse.
Transformers does not use them. We need to trace inputs to verify
whether they are equivalent.
* downsample_factor -> merge_size
* Add vision graph
lol, forgot from a previous commit
* Additional renames, align with llama.cpp / transformers
* Prefer _size instead of independent _h and _w
* Fix token layout
Co-authored-by: Young Han <younghan@fb.com>
* onyx: bring the chat parser onto the onyx branch
common/chat.cpp on this branch has no Onyx handling, so a converted model
serves malformed chat: the assistant preamble leaks into content
("to=self<|message|>...") and tool calls fail with
HTTP 500 "The model produced output that does not match the expected
peg-native format"
common_chat_params_init_onyx exists on onyx-fair-patch, added there by
8bb73dd3d. It was never on this branch, so this is not a regression --
the two lines developed independently.
The code here is taken verbatim from that commit. It is the clean side of
`git merge origin/onyx-fair-patch`: chat.cpp is one of the files that
merges without conflict. The full merge is not viable -- it produces 13
conflicts, including add/add on conversion/onyx.py and src/models/onyx.cpp
where the q_norm-folding and metadata-scale approaches contradict each
other, and #4/#7 are stacked on this branch's side of that.
Verified on this branch: builds with 0 errors, converts an Onyx checkpoint,
and serving it gives "4" for "What is 2+2?" plus a correct
get_weather {"city":"Paris"} tool call, where the unported branch gives the
two failures above.
No converter or runtime changes are included, so this should not interact
with the q_norm work.
Co-authored-by: Beto de Paola <betodepaola@meta.com>
* Less params, bilinear pos-emb interpolation as a graph op instead of CPU
* Map to symbolic V_MMPROJ instead of strings
* Make a couple params explicit
* Patchify via build_inp()
* No param for rope_theta
* Small cleanup
* Restore blank line
* Unpermute, to adapt to the latest transformers checkpoint
* Apply norm after token embeddings
This follows the latest transformers approach.
* Remove duplicated function
* build_vit
* onyx: use the model rope theta on sliding-window layers
* DFlash: conversion from transformers drafter
* Revert rope_type derivation from target
NOTE: this breaks compatibility with Meta's distributed DFlash GGUFs, as
the Q/K are stored in "NEOX" (rotated half) format, like in
transformers.
* Apply suggestion from @pcuenca
* Set model type
* Remove comment that will become obsolete
* Hardcode post_norm_rms_eps instead of new param
* Derive SWA+RoPE pattern from gguf array or scalar
* Fix model type <-> number of layers
* Reorder
* Rename
* Fix typo
* DFlash: seed the draft KV cache from multimodal embedding batches
`common_speculative_impl_draft_dflash::process()` returned early on any batch carrying embeddings, so an image prefill never had its target-layer features fused through the DFlash encoder and injected into the draft's KV cache. That left a hole spanning the image's positions, and the next injection at a post-image position failed to initialize its batch:
```
decoding image batch 1/1, n_tokens_batch = 256
decode: failed to initialize batch
llama_decode: failed to decode, ret = -1
process: llama_decode(ctx_dft) failed rc=-1 (n_tokens=17, offset=0)
srv decode: failed to process speculative batch
```
Every image request with `--spec-type draft-dflash` failed with HTTP 500. Text-only was unaffected, since those batches carry token ids and were let through.
Restore the earlier condition, which admits a batch that is either tokens or embeddings and skips only the degenerate neither/both cases. The rest of `process()` is already layout-agnostic -- it gathers features via `llama_get_embeddings_layer_inp()` and indexes `batch_in.pos[]` / `batch_in.seq_id[]`, none of which assume token ids -- so this is the whole fix.
Validated against `muse-glimmer-30B-bf16.gguf` + `mmproj-muse-glimmer-30B-bf16.gguf` + a DFlash draft head, on an image describe-the-shapes request:
- before: HTTP 500, `failed to process speculative batch`
- after: HTTP 200, draft acceptance 0.34012 (167 accepted / 491 generated), mean len 3.04
Output equivalence holds, which is the property that matters: at temperature 0 the drafted response is byte-identical to the same request served with no draft attached (1213/1213 chars), so the draft is drafting correctly through the image context rather than merely not crashing.
* Conversion: prefer rewrite to mapping
* Revert "Conversion: prefer rewrite to mapping"
This reverts commit a92d0ac584.
* fix lint
* sliding_window metadata is not optional
* disable state save/load
* Apply suggestion from @pcuenca
---------
Co-authored-by: Young Han <younghan@fb.com>
Co-authored-by: Beto de Paola <betodepaola@meta.com>
Co-authored-by: Daniel Han <michaelhan2050@gmail.com>
Co-authored-by: ruanrms <ruanslv@gmail.com>
Co-authored-by: Xuan Son Nguyen <son@huggingface.co>
Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@huggingface.co>
* Restore quantization of mmprojs
This was lost in the refactor undertaken in #22004.
* add noreturn
---------
Co-authored-by: Xuan Son Nguyen <son@huggingface.co>
* granite-switch: add llama.cpp backend (POC, CPU)
New "granite-switch" architecture: a dense, all-attention Granite-4.1
model with N embedded LoRA adapters selected per-token by control tokens.
- gguf-py schema (arch, KV keys, stacked LoRA tensor names) + writer helpers
- conversion/granite.py: GraniteSwitchModel converter (stacks N adapters +
zero base slot into per-projection A/B tensors; emits switch metadata)
- C++ arch registration (llama-arch.{h,cpp}, llama-model.{h,cpp})
- src/models/granite_switch.cpp: load + per-token switched-LoRA graph via
ggml_mul_mat_id over stacked tensors; sticky per-token index + control-token
substitution in llm_graph_input_switch::set_input
- llm_graph_input_switch in src/models/models.h
Runs end-to-end on CPU: convert 3b checkpoint (842 tensors, stacked dim 13)
and generate on both base and control-token paths. Sticky switch state is
single-sequence (POC); full multi-sequence machinery is a follow-up.
* granite-switch: add Mac (Metal) build + mid-sequence switch demo script
Self-contained script to build llama.cpp on Apple Silicon (Metal),
convert the composed 3b checkpoint, and run the crisp mid-sequence
adapter-switch demos verified on Vela:
- answerability: <|answerability|> mid-seq -> "unanswerable"
- query_rewrite: <|query_rewrite|> mid-seq -> {"rewritten_question": ...}
Each demo runs the same prompt twice, differing only by a control token
placed before the assistant turn, so the per-token switch is visible.
* granite-switch mac demo: add -no-cnv so each run is one-shot
The composed model ships a chat template, so llama-completion auto-enables
interactive conversation mode and halts at a `>` prompt after generating,
stalling the script. -no-cnv disables conversation mode: generate once from
the raw prompt and exit (also prints special tokens, making the switch visible).
* granite-switch: replace global sticky index with in-graph router attention
The POC computed the per-token adapter index on the CPU and carried it
across ubatches in ONE global `mutable int32_t poc_sticky_index`, reset
only when a ubatch contained sequence position 0. That global had two
problems:
1. Concurrency: with multiple sequences in a batch it was last-writer-
wins — one sequence's adapter leaked into the others.
2. Multi-turn: an interactive `ollama run` chat continues one KV cache,
so turn 2 never saw position 0 and the index never reset — the
adapter stayed stuck on across turns.
Port the vLLM/HF backend mechanism faithfully: a single-head causal
"router" attention recovers the adapter index in-graph. Per token, only
dim 0 carries signal — Q[0]=1, K[0]=+gain for a control token / -gain
otherwise, V[0]=adapter slot / 0 — and the causal softmax over the single
visible control token recovers that adapter's slot (readback =
clamp(round(V[0]), 0, n_adapters)). gain=15 matches config.py and is
F16-safe (no F32 cache).
The router's K/V live in the model KV cache at an extra layer
R == hparams.router_layer (== n_layer). We bump n_layer_all to n_real+1
so the cache allocator gives the router its own per-sequence slot, and
set n_layer_nextn=1 so n_layer() stays n_real — the decoder loop and
tensor loading are untouched and never reference layer R. The router K is
exempted from the k-shift RoPE loop (its dim-0 value is a literal
magnitude, not a rotation).
Because the selection now lives in the per-sequence KV cache, CONCURRENT
requests are isolated for free (problem 1 fixed; verified by
scratch/concurrent_switch_test.cpp). set_input becomes stateless pure
per-token maps; the global is gone.
Single-switch contract / known limitation, identical to vLLM & HF: the
gain is flat (no recency), so within one sequence there is no mechanism to
revert to base mid-sequence — once an adapter fires it stays on until that
sequence ends (problem 2 is therefore NOT fixed by a faithful copy; vLLM/HF
avoid it only because each served request is a fresh sequence). A client
continuing one KV cache across turns must start a fresh sequence per turn,
or opt into a recency-biased router (a deliberate divergence, not done
here). Documented in granite_switch.cpp and asserted by
scratch/multiturn_leak_test.cpp.
Verified (CPU): both demos unchanged (answerability -> "unanswerable",
query_rewrite -> rewritten query); concurrent two-sequence isolation
passes; multi-turn carry-over matches the vLLM/HF contract.
* granite-switch: drop scratch tests and mac demo for upstream PR
Remove the local-only development artifacts that should not ship in the
upstream PR:
- granite-switch-mac-demo.sh (local Metal build + demo driver)
- scratch/concurrent_switch_test.cpp
- scratch/multiturn_leak_test.cpp
Also drop the now-dangling reference to the scratch tests from the
granite_switch.cpp header comment. Leaves only the core architecture
support (conversion, gguf constants, llama-arch/model/kv-cache, and the
granite_switch graph).
* granite-switch: trim comments to match native llama.cpp style
* granite-switch: trim conversion comments to match native style
* granite-switch: drop unused adapter_ranks metadata
* granite-switch: rename arch to graniteswitch and drop obid alias
* granite-switch: fix non-ASCII comments and document router gain assumption
* granite-switch: drop section comments from constants.py to match native style
* granite-switch: add functional tensor block comments matching Granite4 Vision style
* granite-switch: clarify n_expert_used comment
State the actual constraint: mul_mat_id needs n_expert_used == 1, and
since the GGUF carries expert_count = 0 the generic loader's
n_expert == 0 => n_expert_used == 0 assertion has already passed by the
time load_arch_hparams runs, so it is forced to 1 here.
* granite-switch: note n_layer_nextn reuse has no MTP
The router carving reuses n_layer_nextn, normally the MTP/next-token
count. Clarify in the comment that it is borrowed here purely as the
trailing-layers lever and that there is no MTP head, to spare readers
the double-take.
* granite-switch: rename source file and apply review nits
* granite-switch: don't force LoRA tensors to F16, follow --outtype instead
* granite-switch: drop redundant _permute_qk wrapper, call LlamaModel.permute directly
* granite-switch: read router gain from GGUF (control_token_gain) instead of hardcoding 15.0
* granite-switch: derive n_slots()
* granite-switch: move llm_graph_input_switch into granite-switch.cpp
* granite-switch: cut AI-style narration comments
* granite-switch: collapse multi-line comments
* granite-switch: rename control_token_* maps to adapter_token_*
* granite-switch: cut noise comments
* granite-switch: rename embedded LoRA tensors to <base>.lora_a/lora_b
* granite-switch: GGML_ASSERT token input to avoid UB on embeddings
* granite-switch: TODO for raw embedding input support
* granite-switch: collapse LoRA tensor constants to .lora_a/.lora_b suffix
* granite-switch: drop n_expert_used hack, guard mul_mat_id buft probe
* granite-switch: stop forcing dense expert counts, read from config
* granite-switch: renamed control_token_gain metadata key to router_gain
* granite-switch: trim header comments to match native style
* granite-switch: collapse LoRA tensors to base name + suffix
* granite-switch: inline suffix checks in tensor op resolution
* granite-switch: drop switch-lora struct comment
* granite-switch: guard router layer index and inline n_slots
* granite-switch: group adapter metadata under {arch}.adapters.* namespace
* granite-switch: add hparams.has_rope(il) for KV-shift rope skipping
* granite-switch: skip arch in test-llama-archs (adapter fixture missing, TODO)
* granite-switch: Keys.Adapters namespace + simplify n_slots
* granite-switch: validate substitute token ids against n_vocab
* granite-switch: bound adapter count and lora rank from GGUF
* granite-switch: reject MTP context type when router_layer is set
* granite-switch: throw on bad adapter metadata instead of GGML_ASSERT
* granite-switch: use ASCII +/- in router K signal comment
* granite-switch: document n_layer_nextn repurpose and its leak points
* granite-switch: gate lora_a/lora_b op mapping on router_layer
* granite-switch: label all three preview model sizes
The saver called add_kv with LLM_KV_EXPERT_SHARED_FEED_FORWARD_LENGTH twice, the
second time passing n_ff_chexp. gguf_set_val_u32 removes-then-appends, so the second
call clobbers the first: the saved shared_feed_forward_length ends up as n_ff_chexp
(0 for every arch except GroveMoE), and expert_chunk_feed_forward_length is never
written at all.
So a save->load roundtrip of any MoE model with a shared expert loses n_ff_shexp. On
reload the arch falls back to n_ff for the shexp tensor shape, that no longer matches
the saved tensor, and the model FAILS to load. Hits qwen2moe, qwen3-next, granite-moe,
hunyuan-moe, ernie4.5, bailingmoe2, nemotron-h, and the other shared-expert MoEs.
Fix: the second call writes LLM_KV_EXPERT_CHUNK_FEED_FORWARD_LENGTH.
test-llama-archs: set expert_shared_feed_forward_length to a value distinct from n_ff
in the MoE setup so the roundtrip exercises it. Without the fix the reload fails on a
shexp tensor-shape mismatch; with it, every arch roundtrips clean.
* Resolve -1 to 1024 instead of ctx-len for samplers
Because of backend-sampling we initialize samplers before the complete
llama_context is there. Therefore, we cannot infer the resolved context
length yet at the time we construct the samplers.
* Shared default of 64 for history-based samplers, remove context_size
* sampling: enhance penalty handling in common_sampler_init
- Set default value for penalty_last_n based on model context if not specified.
- Ensure penalty_last_n and n_prev are non-negative.
- Update llama_sampler_penalties structure to inherit from llama_sampler_backend and add backend input handling for penalties.
- Implement backend initialization and application logic for penalties, including frequency and presence adjustments.
* tests: add backend penalties sampling tests and utility functions
- Introduced `accept_prompt` and `unique_prompt_tokens` functions to handle prompt acceptance and token uniqueness.
- Implemented `compare_penalties_logits` to compare logits from backend and CPU samplers with penalties.
- Added `test_backend_penalties_sampling` to validate backend penalties with various configurations.
- Enhanced the test suite for better coverage of penalty handling in sampling.
* sampling: add support for top-k penalties in backend sampling
* sampling: add fix to ensure stable numerical results. Preserve masked logits as -Inf and no longer generate NaN.
* sampling: enhance penalty comparison tests with masking penalties logic
* add comments on padding
* sampling: add comments on modifications
* add the unit test to cover masked-out token as -INF
* validate repeat penalty to ensure it is finite and greater than 0; add tests for invalid values
* refactor: test functions to share logic and be less verbose
* add test to cover case where previously penalized token is not part of candidates
* remove comments
* remove redundant penalty_last_n initialization and validation in common_sampler_init
* add support for penalties in sampler chain with configurable positions
* add validation for penalty parameters and enhance tests for non-finite values
* add context parameter to common_sampler_init and set default for penalty_last_n
* add llama_n_ctx parameter to common_sampler_init for improved sampler initialization
* replace penalty_last_n x n_candidates comparison matrix with a vocabulary-sized count tensor
* add tests for backend penalties sampling without filler entries , token_count.size() == n_active == n_max == 64
* add test for backend penalties sampling after top-p with large history window
* remove as unused
* add is_disabled method, tensor logits reshape, add rest review suggestions
* clarify comment
* llama : MTP support for DeepSeek V3.2
* model : no need to include MTP layers during DeepSeek V3.2 model type discovery
---------
Co-authored-by: Stanisław Szymczyk <sszymczy@gmail.com>
* llama : load MTP tensors only if they are really used
* llama : skip loading MTP (if not used) in remaining models that support MTP
---------
Co-authored-by: Stanisław Szymczyk <sszymczy@gmail.com>
* llama : enforce the same K and V cache types for DeepSeek V4; enable FA if V cache is quantized
* llama : enforce the same K and V cache types for MLA models
---------
Co-authored-by: Stanisław Szymczyk <sszymczy@gmail.com>
* model: add NextN/MTP speculative decoding support for GLM_DSA (GLM-5.2)
Adds GLM-5.2 NextN/MTP as a --spec-type draft-mtp target: nextn tensor
loading via the qwen35moe/step35-style presence probe, a graph_mtp
builder (enorm/hnorm/eh_proj + dense MLA + sigmoid-gated MoE with
shared expert + shared head with fallbacks, _s scale tensors passed
for NVFP4), t_h_nextn extraction in the trunk graph, and MTP-context
KV setup: the draft head runs dense MLA, so the MTP context uses a
plain attention KV cache holding only the nextn layer(s) (same
pattern as the hybrid Qwen3.5 MTP context) while the main context
keeps the DSA cache, now filtered to trunk layers only.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* convert : support --mtp/--no-mtp export for GlmMoeDsaForCausalLM (GLM-5.2)
Opt GLM-5.2 into the supports_mtp_export contract (post-#25641 shape,
mirroring HYV3Model/Step35Model): --no-mtp drops the appended NextN
block (blk.78) and its nextn_predict_layers KV; --mtp keeps only the
NextN block plus shared embeddings/norm/lm_head. Default (bundled)
output is unchanged.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
* spec: add DSpark speculative decoding
DSpark (DeepSpec, 2026) on top of the merged DFlash drafter. It reuses the
DFlash encoder/decoder graph, target feature extraction and KV-cache injection,
and the verify/accept path unchanged; the draft model is a new "dspark" arch
adding a low-rank Markov head (markov_w1/w2) and an optional (unused here)
confidence head. No new public APIs.
The proposal is the only change: the block is anchor-first (position 0 already
predicts the first draft) and the decoder graph applies a semi-autoregressive,
previous-token conditioned logit bias in-graph, chained per block position:
logits'(i) = logits(i) + markov_w2 . markov_w1[prev(i)]
prev(0) = the block's anchor token, prev(i>0) = argmax(logits'(i-1))
vectorized across all blocks in the batch; the anchors are fed through a
dedicated graph input (token 0 of every block). Greedy stays lossless
(verify unchanged, same as DFlash).
- new arch "dspark" (llama_model_dspark : llama_model_dflash, reuses the graph,
loads the markov/confidence tensors; shares the target's embed/lm_head).
- Qwen3DSparkModel converter.
- new spec type "draft-dspark" (common_speculative_impl_draft_dspark :
common_speculative_impl_draft_dflash, overrides draft() only: submits whole
anchor-first blocks and greedily reads back the biased logits).
* spec: read draft block size in the dflash impl
* docs: add DSpark section to speculative.md
* spec: keep dspark block size read in the dspark impl
* dspark : add TODOs for incomplete parts
- confidence head is loaded but not used yet
- confidence-scheduled prefix pruning is not implemented
- the in-graph Markov chain is greedy-only
- only Qwen3 backbones are supported for now (also noted in docs)
* spec: fold DSpark into the DFlash arch
Address review: drop LLM_ARCH_DSPARK and the dspark.block_size /
markov_rank GGUF keys. A DSpark draft now converts to a DFlash GGUF;
the Markov head tensors are detected by presence (like eagle3 d2t),
block_size is read from the existing dflash.block_size key, and the
block anchors are taken as a strided view of the decoder's token
input instead of a separate graph input.
* spec: add confidence-based draft pruning for DSpark
The DSpark confidence head predicts per-position acceptance of the
drafted block. --spec-draft-conf-min truncates the block at the first
position below the threshold (default 0 = disabled).
* fold the dspark impl into dflash, selected by spec type
* address review comments
* dspark: clean up and improve naming
* update readme
* remove trailing whitespace
* dflash: draft full n_max blocks, defer dp.n_max to the central truncation
The DSpark markov head views the draft batch as a uniform [n_seqs x block]
grid, but the per-seq dp.n_max clamp could produce blocks of different
sizes, silently corrupting the strided views and the resulting logits.
Drop the clamp and always draft the full n_max block for every sequence:
dp.n_max is already enforced by the central truncation in
common_speculative_draft(), the same way eagle3 handles it.
Co-authored-by: Zaire404 <3147879462@qq.com>
* dflash: assert the markov head block-uniformity invariant, require the conf head
With the draft batch always submitting equal-size n_max blocks, a
non-divisible token count can only mean the batch was split across
ubatches or a caller broke the layout - fail loudly instead of silently
dropping the markov bias. The block_drafts > block_size early return
stays: worst-case graph reserve passes legitimately build with
n_seq_tokens > block_size.
Also make conf_proj required when the markov head is present: the
confidence head is part of the DSpark checkpoint format, and a missing
head would otherwise leave --spec-draft-conf-min silently reading stale
embeddings instead of confidences.
Co-authored-by: Zaire404 <3147879462@qq.com>
* dspark: fold conf_min into p_min
p_min and conf_min express the same thing - the minimum predicted
survival probability for a drafted position - differing only in how the
estimate is obtained: token probability for regular drafters, the
trained confidence head for DSpark. The DSpark readback never used
p_min, so reuse it for the confidence threshold and drop the separate
--spec-draft-conf-min flag. Both defaulted to 0 (disabled), so behavior
is unchanged.
Co-authored-by: Zaire404 <3147879462@qq.com>
* dflash: note the confidence broadcast workaround
Requested in review: the ggml_repeat only adapts the [1, n_tok]
confidences to the n_embd-wide embd_nextn transport so that
llama_get_embeddings_nextn can be reused - not a placeholder.
Co-authored-by: Zaire404 <3147879462@qq.com>
* cont : clarify
[no ci]
---------
Co-authored-by: Ruixiang Wang <wangruixiang07@outlook.com>
Co-authored-by: Zaire404 <3147879462@qq.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Add preliminary MiniMax-M3 support
Text-only port that re-uses existing components: MiniMax-M2 style GQA with
per-head QK-norm and partial rotary, DeepSeek-V3 style leading-dense and
routed/shared experts, and swigluoai activation. Sparse attention is not
yet supported (dense fallback); vision tower and MTP heads are dropped.
* MiniMax-M3 vision tower (mmproj + clip graph)
* Delete m3_vision_ref.py
* Update clip.cpp
* MSA
* Update constants.py
* Update minimax.py
* Cache creation. Working withotu flash attention
* Added flash attention for sparse layers
* Decomposed slow cpu OP into GPU + CPU ops. Massive speedup over long ctx
* Rewrote indexer op to be cuda native. Modified flash attention to match per group block picking
* Implement sparse attention calc out of stock ops.
* Fix a cache allocation and cont issue
* Fixed -fa auto crash, flagged debug spots
* Delete vocab.json
* Delete model.safetensors.index.json
* Delete generation_config.json
* Delete Minimax directory
* Handled multi stream case to fall back on Dense Attention
* Development scaffolding cleanup. No functional change to the decode or
4-way paths. Full debug harness remains at <8136a9c68ed7a5eb009aa67bba3fda8062f4648f> for reproducing the
selection-parity validation.
* Remove redundant comment from minimax-m3.cpp
* Changed 3 Gelu Ops for vision into Gelu_erf ops
* Assert that n_kv is multiple of 128
* Rename MSA index tensors to indexer convention
Note: All GGUFs generated before this change will need to be regenerated.
* Fix incorrect Assert
* Review driven changes (#3)
* Remove comment from conversion minimax.py
Co-authored-by: Sigbjørn Skjæret <1629204+CISC@users.noreply.github.com>
* Remove whitespaces from constants.py
Co-authored-by: Sigbjørn Skjæret <1629204+CISC@users.noreply.github.com>
* Tighten comment in minimax.py
Co-authored-by: Sigbjørn Skjæret <1629204+CISC@users.noreply.github.com>
* inherit MiniMax-M3 from MiniMax-M2
* drop dead text_config fallbacks
* Add indexer writer methods
* Reuse LLM_FFN_SWIGLU_OAI_MOE
* Remove duplicate indexer setters, add only block_size/local_blocks, follow value naming convention
* Fix conversion error /gguf_writer.py
Co-authored-by: Sigbjørn Skjæret <1629204+CISC@users.noreply.github.com>
* Update gguf-py/gguf/gguf_writer.py
Co-authored-by: Sigbjørn Skjæret <1629204+CISC@users.noreply.github.com>
* Update gguf-py/gguf/tensor_mapping.py
Co-authored-by: Sigbjørn Skjæret <1629204+CISC@users.noreply.github.com>
* Update conversion/minimax.py
Co-authored-by: Sigbjørn Skjæret <1629204+CISC@users.noreply.github.com>
* Update conversion/minimax.py
Co-authored-by: Sigbjørn Skjæret <1629204+CISC@users.noreply.github.com>
* Remove whitespace in src/llama-kv-cache.cpp
Co-authored-by: Sigbjørn Skjæret <1629204+CISC@users.noreply.github.com>
* Remove Whitespace in Update src/llama-model.h
Co-authored-by: Sigbjørn Skjæret <1629204+CISC@users.noreply.github.com>
* Remove whitespace in src/llama-hparams.h
Co-authored-by: Sigbjørn Skjæret <1629204+CISC@users.noreply.github.com>
* remove multimodal code upon maintainer request. Will be made as a separate PR
* Whitespace clean in tensor_mapping.py
* Log cache size on launch, block ctx shift, support prompt caching
Log indexer cache size on launch
Disallow ctx shift
Support prompt caching
* Update minimax-m3.cpp
* Optimize implementation, add multi stream support.
Fully rewrote minimax-m3.cpp for speed and buffer size gains:
Unified the 4-way + decode, 1 FA call per layer instead of 4, with the groups mapped onto ne[3]
Custom CPU op now emits block-level mask, expanded on GPU, which causes CPU to GPU transfer to shrinks at prefill
Decode: ~25 nodes/layer vs ~50, no per-group concats/conts
Unified selection semantics, so both regimes rank bs + local bias (position-anchored local force), which means prefill/decode can no longer disagree on selection
can_reuse on the MSA bias input. Graph reuse at decode restored (was rebuilding the full graph every token)
In-place mask adds, shrinking compute buffer ~6.8 to ~4.2 GiB at ub2048/62k
Multi-stream: MSA now runs with -np N when kv_unified=false. Decode stays batched across streams (still 1 FA call), prefill loops per stream. dense fallback only for --kv-unified + multi-seq
Measured effect on expert offload bound setup: decode 6.2(4WAY)–7.15(MSA_decode) -> 7.7~7.8 t/s, flat from 5k to 60k+. prefill around 10% faster. buffer about 20% smaller, multi-user support.
* set default cache type to F32
* Fix potential DSA double indexer cache allocation bug, only allocate in-cache k_idx for archs that opt in
* remove F16 downcasts in MSA attention, force F32 indexer score accum
* Add Minimax eos to llama vocab
* Guard edge case where idx cache can become stale after a tail trim
* Update llama-kv-cache.h
* Update llama-kv-cache.cpp
* Update llama-kv-cache.cpp
* Update llama-kv-cache.h
* Update llama-kv-cache.cpp
* Review driven changes
* style fix
* indexer hparams are required
* fix tests
* fix lint
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
Co-authored-by: Sigbjørn Skjæret <1629204+CISC@users.noreply.github.com>
Co-authored-by: Xuan Son Nguyen <son@huggingface.co>
* Start building graph - reuse deepseek32
* Enable kv cache and rotation for glm_dsa architecture
Just follow Deepseek 3.2 for now.
* Reuse prev_top_k for "shared" indexer layers
* GLM 5.2 uses LLAMA_ROPE_TYPE_NORM for the indexer.
This is transformers' `apply_rotary_pos_emb_interleave`
* Default indexer types to GLM pattern
Previous converted GGUFs like https://huggingface.co/unsloth/GLM-5.2-GGUF write indexer weights to _all_ layers, even if they are only required for "full" types. This PR relies on a new key "%s.attention.indexer.types"; if absent, it will use the default GLM 5.2 schedule as defined in https://huggingface.co/zai-org/GLM-5.2/blob/main/config.json#L26.
Note that conversion is not saving this key yet.
* Save indexer types to gguf, restore on load
* Use ggml_lightning_indexer when cparams.fused_lid
Co-authored-by: fairydreaming <166155368+fairydreaming@users.noreply.github.com>
* GLM 5 and 5.1 use full indexers
Co-authored-by: fairydreaming <166155368+fairydreaming@users.noreply.github.com>
* Fix indentation
* Ensure array is zero-filled
* Prefer explicit std::fill
* Assert prev_top_k exists for shared indexer
---------
Co-authored-by: fairydreaming <166155368+fairydreaming@users.noreply.github.com>
lnigam
Xuan-Son Nguyen
Ruben Ortlam
Junmo Kim
Gaurav Garg
Pedro Cuenca
Ruixiang Wang
Bar Haim
Caleb DeLeeuw
Georgi Gerganov
Aldehir Rojas
Sergey Malinin
Sigbjørn Skjæret
jacekpoplawski
timkhronos
fairydreaming
Konrad Moren
Masashi Yoshimura
Aman Gupta
Sigbjørn Skjæret
o7si
Tunahan
Satinder Grewal
Guido Imperiale
王金旭
zql
Aaron Teo
Bartowski