sleepy/llama.cpp
1
0
Fork 0
Code Issues 11 Pull Requests Actions 109 Packages Projects Releases Wiki Activity

common : enable reasoning budget sampler for gemma4 (#21697)

Browse Source 
* fix: enable reasoning budget sampler for gemma4

Add thinking_start_tag and thinking_end_tag to
common_chat_params_init_gemma4(). Without these, the reasoning
budget sampler never activates for gemma4.

Make the newline after "thought" optional in the PEG parser to
handle budget=0 (sampler forces end tag before the newline).

Add test case for empty thinking block.

Fixes #21487

* use p.space() instead of p.optional(p.literal("\n")) in gemma4 thought parser
This commit is contained in:
Berk Idem
2026-04-10 05:49:14 -04:00
committed by GitHub
parent 3f8752b559
commit d7ff074c87
2 changed files with 12 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
common/chat.cpp
+5 -3
View File
@@ -1083,7 +1083,9 @@ static common_chat_params common_chat_params_init_gemma4(const common_chat_templ
data.prompt = common_chat_template_direct_apply_impl(tmpl, inputs); data.prompt = common_chat_template_direct_apply_impl(tmpl, inputs);
data.format = COMMON_CHAT_FORMAT_PEG_GEMMA4; data.format = COMMON_CHAT_FORMAT_PEG_GEMMA4;
data.supports_thinking = true; data.supports_thinking = true;
data.thinking_start_tag = "<|channel>thought";
data.thinking_end_tag = "<channel|>";
data.preserved_tokens = { data.preserved_tokens = {
"<|channel>", "<|channel>",
@@ -1102,9 +1104,9 @@ static common_chat_params common_chat_params_init_gemma4(const common_chat_templ
auto start = p.rule("start", p.prefix(inputs.generation_prompt, "<|channel>")); auto start = p.rule("start", p.prefix(inputs.generation_prompt, "<|channel>"));
if (extract_reasoning) { if (extract_reasoning) {
p.rule("thought", p.literal("<|channel>thought\n") + p.reasoning(p.until("<channel|>")) + p.literal("<channel|>")); p.rule("thought", p.literal("<|channel>thought") + p.space() + p.reasoning(p.until("<channel|>")) + p.literal("<channel|>"));
} else { } else {
p.rule("thought", p.content(p.literal("<|channel>thought\n") + p.until("<channel|>") + p.literal("<channel|>"))); p.rule("thought", p.content(p.literal("<|channel>thought") + p.space() + p.until("<channel|>") + p.literal("<channel|>")));
} }
auto thought = (p.peek(p.literal("<|channel>")) + p.ref("thought")) | p.negate(p.literal("<|channel>")); auto thought = (p.peek(p.literal("<|channel>")) + p.ref("thought")) | p.negate(p.literal("<|channel>"));
tests/test-chat.cpp
+7
View File
@@ -1988,6 +1988,13 @@ static void test_template_output_peg_parsers(bool detailed_debug) {
.expect(message_assist_thoughts) .expect(message_assist_thoughts)
.run(); .run();
// Empty reasoning (budget=0: sampler forces end tag before newline)
tst.test(
"<|channel>thought<channel|>Hello, world!\nWhat's up?")
.reasoning_format(COMMON_REASONING_FORMAT_AUTO)
.expect(simple_assist_msg("Hello, world!\nWhat's up?", ""))
.run();
// Reasoning and content with reasoning_format = none // Reasoning and content with reasoning_format = none
tst.test( tst.test(
"<|channel>thought\nI'm\nthinking<channel|>Hello, world!\nWhat's up?") "<|channel>thought\nI'm\nthinking<channel|>Hello, world!\nWhat's up?")