tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-06-28 04:15:21 +00:00
Code Issues Packages Projects Releases Wiki Activity

tool-call: refactor common chat / tool-call api (+ tests / fixes) (#11900)

Browse Source 
* tool-call refactoring: moved common_chat_* to chat.h, common_chat_templates_init return a unique_ptr to opaque type

* addressed clang-tidy lints in [test-]chat.*

* rm minja deps from util & common & move it to common/minja/

* add name & tool_call_id to common_chat_msg

* add common_chat_tool

* added json <-> tools, msgs conversions to chat.h

* fix double bos/eos jinja avoidance hack (was preventing inner bos/eos tokens)

* fix deepseek r1 slow test (no longer <think> opening w/ new template)

* allow empty tools w/ auto + grammar

* fix & test server grammar & json_schema params w/ & w/o --jinja
This commit is contained in:
Olivier Chafik
2025-02-18 18:03:23 +00:00
committed by GitHub
parent 63ac128563
commit 63e489c025
18 changed files with 1385 additions and 993 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
examples/server/tests/unit/test_tool_call.py
View File

@ -356,12 +356,12 @@ def test_weather(hf_repo: str, template_override: str | Tuple[str, str | None] |
(None, 128, "bartowski/functionary-small-v3.2-GGUF:Q8_0", ("meetkai/functionary-medium-v3.2", None)),
(None, 128, "bartowski/Meta-Llama-3.1-8B-Instruct-GGUF:Q4_K_M", None),
(None, 128, "bartowski/Mistral-Nemo-Instruct-2407-GGUF:Q4_K_M", None),
("^> 0.56$", 128, "bartowski/Mistral-Nemo-Instruct-2407-GGUF:Q4_K_M", "chatml"),
(None, 128, "bartowski/Mistral-Nemo-Instruct-2407-GGUF:Q4_K_M", "chatml"),
(None, 128, "bartowski/Phi-3.5-mini-instruct-GGUF:Q4_K_M", None),
# TODO: fix these (wrong results, either didn't respect decimal instruction or got wrong value)
("^The y-coordinate [\\s\\S]*?\\*\\*0.5\\*\\*", 8192, "bartowski/DeepSeek-R1-Distill-Qwen-7B-GGUF:Q4_K_M", None),
("[\\s\\S]*?\\*\\*0\\.5\\*\\*", 8192, "bartowski/DeepSeek-R1-Distill-Qwen-7B-GGUF:Q4_K_M", ("llama-cpp-deepseek-r1", None)),
("[\\s\\S]*?\\*\\*\\s*0.5($|\\*\\*)", 8192, "bartowski/DeepSeek-R1-Distill-Qwen-7B-GGUF:Q4_K_M", None),
# ("[\\s\\S]*?\\*\\*\\s*0.5($|\\*\\*)", 8192, "bartowski/DeepSeek-R1-Distill-Qwen-7B-GGUF:Q4_K_M", ("llama-cpp-deepseek-r1", None)),
])
def test_calc_result(result_override: str | None, n_predict: int, hf_repo: str, template_override: str | Tuple[str, str | None] | None):
global server
@ -401,7 +401,7 @@ def test_calc_result(result_override: str | None, n_predict: int, hf_repo: str,
{
"role": "tool",
"name": "calculate",
"content": 0.55644242476,
"content": "0.55644242476",
"tool_call_id": "call_6789"
}
],
@ -444,7 +444,7 @@ def test_calc_result(result_override: str | None, n_predict: int, hf_repo: str,
(128, None, "^The sum of 102 and 7 is 109.*", None, "bartowski/Phi-3.5-mini-instruct-GGUF:Q4_K_M", None),
(1024, 'deepseek', "To find the sum of.*", "I need to calculate the sum of 102 and 7.*", "bartowski/DeepSeek-R1-Distill-Qwen-7B-GGUF:Q4_K_M", None),
(1024, 'none', "<think>\n?I need[\\s\\S]*?</think>\n?To find.*", None, "bartowski/DeepSeek-R1-Distill-Qwen-7B-GGUF:Q4_K_M", None),
(1024, 'none', "^I need[\\s\\S]*?</think>\n?To find.*", None, "bartowski/DeepSeek-R1-Distill-Qwen-7B-GGUF:Q4_K_M", None),
(1024, 'deepseek', "To find the sum of.*", "First, I [\\s\\S]*", "bartowski/DeepSeek-R1-Distill-Qwen-7B-GGUF:Q4_K_M", ("llama-cpp-deepseek-r1", None)),
])