Public/koboldcpp
1
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2026-09-06 13:01:21 +02:00
Code Issues Packages Projects Releases Wiki Activity

Merge commit '5a69c974392020e514c3b2b2910bb92f847cb4c9' into concedo_experimental

Browse Source 
# Conflicts:
#	ggml/src/ggml-opencl/ggml-opencl.cpp
#	ggml/src/ggml-opencl/kernels/concat.cl
#	ggml/src/ggml-opencl/kernels/cpy.cl
#	ggml/src/ggml-opencl/kernels/get_rows.cl
#	ggml/src/ggml-opencl/kernels/mul_mv_q6_k_f32_flat.cl
#	tests/test-chat.cpp
#	tools/mtmd/CMakeLists.txt
#	tools/mtmd/clip.cpp
#	tools/mtmd/clip.h
This commit is contained in:
Concedo
2026-06-07 17:46:09 +08:00
parent 202aaef40b 5a69c97439
commit e40e5c660e
30 changed files with 1301 additions and 280 deletions
Download Patch File Download Diff File Expand all files Collapse all files
common/chat.cpp
+26 -115
View File
@@ -1623,42 +1623,40 @@ static common_chat_params common_chat_params_init_kimi_k2(const common_chat_temp
return data;
}
// LFM2 format: uses <|tool_list_start|>[...]<|tool_list_end|> in system prompt
// and <|tool_call_start|>[name(arg="val")]<|tool_call_end|> for tool calls.
// - Reasoning: <think>{reasoning}</think> (optional)
// - Content: text before a tool call (optional)
// - Tool calls: Python-style, e.g. [function_name(arg1="value1", arg2="value2")]
// Tool calls can appear multiple times (parallel tool calls supported)
static common_chat_params common_chat_params_init_lfm2(const common_chat_template & tmpl,
const autoparser::generation_params & inputs) {
// LFM2/LFM2.5 parser. Tool calls are almost Python-style and parallel-capable
// (except dotted names and JSON literals true/false/null).
// Always wrapped in <|tool_call_start|>[name(args)]<|tool_call_end|> with optional <think> reasoning.
// tool_list_tokens preserves LFM2 system tool-list markers.
static common_chat_params common_chat_params_init_lfm2(const common_chat_template & tmpl,
const autoparser::generation_params & inputs,
bool tool_list_tokens) {
common_chat_params data;
const std::string TOOL_CALL_START = "<|tool_call_start|>";
const std::string TOOL_CALL_END = "<|tool_call_end|>";
const std::string TOOL_LIST_START = "<|tool_list_start|>";
const std::string TOOL_LIST_END = "<|tool_list_end|>";
const std::string THINK_START = "<think>";
const std::string THINK_END = "</think>";
const std::string GEN_PROMPT = "<|im_start|>assistant\n";
data.prompt = common_chat_template_direct_apply_impl(tmpl, inputs);
data.generation_prompt = common_chat_template_generation_prompt_impl(tmpl, inputs);
data.format = COMMON_CHAT_FORMAT_PEG_NATIVE;
data.supports_thinking = true;
data.preserved_tokens = {
"<|tool_list_start|>",
"<|tool_list_end|>",
"<|tool_call_start|>",
"<|tool_call_end|>",
"<think>",
"</think>",
};
data.preserved_tokens = { TOOL_CALL_START, TOOL_CALL_END, THINK_START, THINK_END };
if (tool_list_tokens) {
data.preserved_tokens.push_back(TOOL_LIST_START);
data.preserved_tokens.push_back(TOOL_LIST_END);
}
data.thinking_start_tag = THINK_START;
data.thinking_end_tag = THINK_END;
auto has_tools = inputs.tools.is_array() && !inputs.tools.empty();
auto extract_reasoning = inputs.reasoning_format != COMMON_REASONING_FORMAT_NONE;
auto include_grammar = has_tools && inputs.tool_choice != COMMON_CHAT_TOOL_CHOICE_NONE;
const std::string TOOL_CALL_START = "<|tool_call_start|>";
const std::string TOOL_CALL_END = "<|tool_call_end|>";
const std::string THINK_START = "<think>";
const std::string THINK_END = "</think>";
const std::string GEN_PROMPT = "<|im_start|>assistant\n";
data.thinking_start_tag = THINK_START;
data.thinking_end_tag = THINK_END;
if (inputs.has_continuation()) {
const auto & msg = inputs.continue_msg;
@@ -1685,7 +1683,7 @@ static common_chat_params common_chat_params_init_lfm2(const common_chat_templat
auto tool_calls = p.rule("tool-calls",
p.trigger_rule("tool-call",
p.literal(TOOL_CALL_START) +
p.python_style_tool_calls(inputs.tools, inputs.parallel_tool_calls) +
p.python_style_tool_calls(inputs.tools, inputs.parallel_tool_calls, /* allow_json_literals = */ true) +
p.literal(TOOL_CALL_END)
)
);
@@ -1712,93 +1710,6 @@ static common_chat_params common_chat_params_init_lfm2(const common_chat_templat
{ COMMON_GRAMMAR_TRIGGER_TYPE_WORD, TOOL_CALL_START }
};
}
return data;
}
// LFM2.5 format: uses plain "List of tools: [...]" in system prompt, no wrapper tokens.
// Tool calls are bare [name(arg="val")], though model may optionally emit <|tool_call_start|>.
// - Reasoning: <think>{reasoning}</think> (optional)
// - Content: text before a tool call (optional)
// - Tool calls: Python-style, e.g. [function_name(arg1="value1", arg2="value2")]
// Tool calls can appear multiple times (parallel tool calls supported)
static common_chat_params common_chat_params_init_lfm2_5(const common_chat_template & tmpl,
const autoparser::generation_params & inputs) {
common_chat_params data;
data.prompt = common_chat_template_direct_apply_impl(tmpl, inputs);
data.generation_prompt = common_chat_template_generation_prompt_impl(tmpl, inputs);
data.format = COMMON_CHAT_FORMAT_PEG_NATIVE;
data.supports_thinking = true;
data.preserved_tokens = {
"<|tool_call_start|>",
"<|tool_call_end|>",
"<think>",
"</think>",
};
auto has_tools = inputs.tools.is_array() && !inputs.tools.empty();
auto extract_reasoning = inputs.reasoning_format != COMMON_REASONING_FORMAT_NONE;
auto include_grammar = has_tools && inputs.tool_choice != COMMON_CHAT_TOOL_CHOICE_NONE;
const std::string THINK_START = "<think>";
const std::string THINK_END = "</think>";
const std::string GEN_PROMPT = "<|im_start|>assistant\n";
data.thinking_start_tag = THINK_START;
data.thinking_end_tag = THINK_END;
if (inputs.has_continuation()) {
const auto & msg = inputs.continue_msg;
data.generation_prompt = GEN_PROMPT + THINK_START + msg.reasoning_content;
if (inputs.continue_final_message == COMMON_CHAT_CONTINUATION_CONTENT) {
data.generation_prompt += THINK_END + msg.render_content();
}
data.prompt += data.generation_prompt;
}
auto parser = build_chat_peg_parser([&](common_chat_peg_builder & p) {
auto generation_prompt = p.literal(GEN_PROMPT);
auto end = p.end();
auto reasoning = p.eps();
if (extract_reasoning && inputs.enable_thinking) {
reasoning = p.optional(THINK_START + p.reasoning(p.until(THINK_END)) + THINK_END);
}
if (!has_tools || inputs.tool_choice == COMMON_CHAT_TOOL_CHOICE_NONE) {
return generation_prompt + reasoning + p.content(p.rest()) + end;
}
auto tool_calls = p.rule("tool-calls",
p.trigger_rule("tool-call",
p.python_style_tool_calls(inputs.tools, inputs.parallel_tool_calls)
)
);
auto content = p.content(p.until_one_of({"<|tool_call_start|>", "["}));
auto maybe_start = p.optional(p.literal("<|tool_call_start|>"));
return generation_prompt + reasoning + content + maybe_start + tool_calls + end;
});
data.parser = parser.save();
if (include_grammar) {
data.grammar_lazy = inputs.tool_choice == COMMON_CHAT_TOOL_CHOICE_AUTO;
data.grammar = build_grammar([&](const common_grammar_builder & builder) {
foreach_function(inputs.tools, [&](const json & tool) {
const auto & function = tool.at("function");
auto schema = function.at("parameters");
builder.resolve_refs(schema);
});
parser.build_grammar(builder, data.grammar_lazy);
});
foreach_function(inputs.tools, [&](const json & tool) {
const std::string name = tool.at("function").at("name");
data.grammar_triggers.push_back({ COMMON_GRAMMAR_TRIGGER_TYPE_WORD, "[" + name + "(" });
});
}
return data;
}
@@ -2313,14 +2224,14 @@ std::optional<common_chat_params> common_chat_try_specialized_template(
if (is_lfm2_template(src)) {
LOG_DBG("Using specialized template: LFM2\n");
return common_chat_params_init_lfm2(tmpl, params);
return common_chat_params_init_lfm2(tmpl, params, /* tool_list_tokens = */ true);
}
// LFM2.5 format detection: template uses plain "List of tools: [...]" with no special tokens
if (src.find("List of tools: [") != std::string::npos &&
src.find("<|tool_list_start|>") == std::string::npos) {
LOG_DBG("Using specialized template: LFM2.5\n");
return common_chat_params_init_lfm2_5(tmpl, params);
return common_chat_params_init_lfm2(tmpl, params, /* tool_list_tokens = */ false);
}
// GigaChatV3 format detection