sleepy/llama.cpp
1
0
Fork 0
Code Issues 11 Pull Requests Actions 108 Packages Projects Releases Wiki Activity

chat: fix parallel_tool_calls default setting based on model capabilities, add tests for parallel tool calls and structured outputs (#22217)

Browse Source 
* chat: fix parallel_tool_calls default setting based on model capabilities, add tests for parallel tool calls and structured outputs

* Fix ty errors.

* Fix flake8 err
This commit is contained in:
Piotr Wilkin (ilintar)
2026-04-22 18:10:56 +02:00
committed by GitHub
parent bcb5eeb645
commit 8bccdbbff9
4 changed files with 1977 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tools/cli/cli.cpp
+3 -1
View File
@@ -207,6 +207,8 @@ struct cli_context {
auto meta = ctx_server.get_meta();
auto & chat_params = meta.chat_params;
auto caps = common_chat_templates_get_caps(chat_params.tmpls.get());
common_chat_templates_inputs inputs;
inputs.messages = common_chat_msgs_parse_oaicompat(messages);
inputs.tools = {}; // TODO
@@ -214,7 +216,7 @@ struct cli_context {
inputs.json_schema = ""; // TODO
inputs.grammar = ""; // TODO
inputs.use_jinja = chat_params.use_jinja;
inputs.parallel_tool_calls = false;
inputs.parallel_tool_calls = caps["supports_parallel_tool_calls"];
inputs.add_generation_prompt = true;
inputs.reasoning_format = COMMON_REASONING_FORMAT_DEEPSEEK;
inputs.force_pure_content = chat_params.force_pure_content;
tools/server/server-common.cpp
+3 -1
View File
@@ -1027,6 +1027,8 @@ json oaicompat_chat_params_parse(
}
}
auto caps = common_chat_templates_get_caps(opt.tmpls.get());
common_chat_templates_inputs inputs;
inputs.messages = common_chat_msgs_parse_oaicompat(messages);
inputs.tools = common_chat_tools_parse_oaicompat(tools);
@@ -1034,7 +1036,7 @@ json oaicompat_chat_params_parse(
inputs.json_schema = json_schema.is_null() ? "" : json_schema.dump();
inputs.grammar = grammar;
inputs.use_jinja = opt.use_jinja;
inputs.parallel_tool_calls = json_value(body, "parallel_tool_calls", false);
inputs.parallel_tool_calls = json_value(body, "parallel_tool_calls", caps["supports_parallel_tool_calls"]);
inputs.add_generation_prompt = json_value(body, "add_generation_prompt", true);
inputs.reasoning_format = opt.reasoning_format;
if (body.contains("reasoning_format")) {