vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

tool-call: functionary-small-v3.2 test now green

Browse source
This commit is contained in:
ochafik 2024-10-31 14:11:34 +00:00
parent b35aa4ae1c
commit 9477c54676
2 changed files with 13 additions and 14 deletions
Download patch file Download diff file Expand all files Collapse all files

23
examples/agent/README.md
View file

@ -7,38 +7,37 @@
```bash ```bash
make -j LLAMA_CURL=1 llama-server make -j LLAMA_CURL=1 llama-server
# Nous Hermes 2 Pro Llama 3 8B
./llama-server --jinja -fa --verbose \ ./llama-server --jinja -fa --verbose \
-hfr NousResearch/Hermes-2-Pro-Llama-3-8B-GGUF -hff Hermes-2-Pro-Llama-3-8B-Q8_0.gguf \ -hfr bartowski/Qwen2.5-7B-Instruct-GGUF -hff Qwen2.5-7B-Instruct-Q4_K_M.gguf
--chat-template "$( python scripts/get_hf_chat_template.py NousResearch/Hermes-2-Pro-Llama-3-8B tool_use )"
# Llama 3.1 8B # Nous Hermes 3 Pro Llama 3.1 8B
./llama-server --jinja -fa --verbose \ ./llama-server --jinja -fa --verbose \
-hfr lmstudio-community/Meta-Llama-3.1-8B-Instruct-GGUF -hff Meta-Llama-3.1-8B-Instruct-Q5_K_M.gguf -hfr NousResearch/Hermes-3-Llama-3.1-8B-GGUF -hff Hermes-3-Llama-3.1-8B.Q4_K_M.gguf \
--chat-template-file <( python scripts/get_hf_chat_template.py NousResearch/Hermes-3-Llama-3.1-8B tool_use )
# Llama 3.1 70B # Phi-3.5 mini (generic support)
./llama-server --jinja -fa --verbose \ ./llama-server --jinja -fa --verbose \
-hfr lmstudio-community/Meta-Llama-3.1-70B-Instruct-GGUF -hff Meta-Llama-3.1-70B-Instruct-Q4_K_M.gguf -hfr bartowski/Phi-3.5-mini-instruct-GGUF -hff Phi-3.5-mini-instruct-Q4_K_M.gguf
# functionary-small-v3 # functionary-small-v3
./llama-server --jinja -fa --verbose \ ./llama-server --jinja -fa --verbose \
-hfr meetkai/functionary-small-v3.2-GGUF -hff functionary-small-v3.2.Q4_0.gguf \ -hfr meetkai/functionary-small-v3.2-GGUF -hff functionary-small-v3.2.Q8_0.gguf \
--chat-template "$( python scripts/get_hf_chat_template.py meetkai/functionary-medium-v3.2 )" --chat-template-file <( python scripts/get_hf_chat_template.py meetkai/functionary-medium-v3.2 )
# Llama 3.2 3B (poor adherence) # Llama 3.2 3B (poor adherence)
./llama-server --jinja -fa --verbose \ ./llama-server --jinja -fa --verbose \
-hfr lmstudio-community/Llama-3.2-3B-Instruct-GGUF -hff Llama-3.2-3B-Instruct-Q6_K.gguf \ -hfr lmstudio-community/Llama-3.2-3B-Instruct-GGUF -hff Llama-3.2-3B-Instruct-Q6_K.gguf \
--chat-template "$( python scripts/get_hf_chat_template.py meta-llama/Llama-3.2-3B-Instruct )" --chat-template-file <( python scripts/get_hf_chat_template.py meta-llama/Llama-3.2-3B-Instruct )
# Llama 3.2 1B (very poor adherence) # Llama 3.2 1B (very poor adherence)
./llama-server --jinja -fa --verbose \ ./llama-server --jinja -fa --verbose \
-hfr lmstudio-community/Llama-3.2-1B-Instruct-GGUF -hff Llama-3.2-1B-Instruct-Q4_K_M.gguf \ -hfr lmstudio-community/Llama-3.2-1B-Instruct-GGUF -hff Llama-3.2-1B-Instruct-Q4_K_M.gguf \
--chat-template "$( python scripts/get_hf_chat_template.py meta-llama/Llama-3.2-3B-Instruct )" --chat-template-file <( python scripts/get_hf_chat_template.py meta-llama/Llama-3.2-3B-Instruct )
# Mistral NeMo # Mistral NeMo
./llama-server --jinja -fa --verbose \ ./llama-server --jinja -fa --verbose \
-hfr bartowski/Mistral-Nemo-Instruct-2407-GGUF -hff Mistral-Nemo-Instruct-2407-Q8_0.gguf \ -hfr bartowski/Mistral-Nemo-Instruct-2407-GGUF -hff Mistral-Nemo-Instruct-2407-Q8_0.gguf \
--chat-template "$( python scripts/get_hf_chat_template.py mistralai/Mistral-Nemo-Instruct-2407 )" --chat-template-file <( python scripts/get_hf_chat_template.py mistralai/Mistral-Nemo-Instruct-2407 )
``` ```
- Run the tools in [examples/agent/tools](./examples/agent/tools) inside a docker container for *some* level of isolation (+ sneaky logging of outgoing http and https traffic: you wanna watch over those agents' shoulders for the time being 🧐). Check http://localhost:8088/docs to see the tools exposed. - Run the tools in [examples/agent/tools](./examples/agent/tools) inside a docker container for *some* level of isolation (+ sneaky logging of outgoing http and https traffic: you wanna watch over those agents' shoulders for the time being 🧐). Check http://localhost:8088/docs to see the tools exposed.

4
examples/server/tests/features/tool_call.feature
View file

@ -104,6 +104,7 @@ Feature: llama.cpp server
| python | {"code": "print('Hello, World!'}"} | bartowski/Llama-3.2-1B-Instruct-GGUF | Llama-3.2-1B-Instruct-Q4_K_M.gguf | meta-llama-Llama-3.2-3B-Instruct | | python | {"code": "print('Hello, World!'}"} | bartowski/Llama-3.2-1B-Instruct-GGUF | Llama-3.2-1B-Instruct-Q4_K_M.gguf | meta-llama-Llama-3.2-3B-Instruct |
| python | {"code": "print("} | bartowski/Llama-3.2-3B-Instruct-GGUF | Llama-3.2-3B-Instruct-Q4_K_M.gguf | meta-llama-Llama-3.2-3B-Instruct | | python | {"code": "print("} | bartowski/Llama-3.2-3B-Instruct-GGUF | Llama-3.2-3B-Instruct-Q4_K_M.gguf | meta-llama-Llama-3.2-3B-Instruct |
| python | {"code": "print("} | lmstudio-community/Meta-Llama-3.1-8B-Instruct-GGUF | Meta-Llama-3.1-8B-Instruct-Q4_K_M.gguf | | | python | {"code": "print("} | lmstudio-community/Meta-Llama-3.1-8B-Instruct-GGUF | Meta-Llama-3.1-8B-Instruct-Q4_K_M.gguf | |
| python | {"code": "print('Hello, World!')"} | bartowski/functionary-small-v3.2-GGUF | functionary-small-v3.2-Q8_0.gguf | meetkai-functionary-medium-v3.2 |
| code_interpreter | {"code": "print('Hello, world!')"} | bartowski/gemma-2-2b-it-GGUF | gemma-2-2b-it-Q4_K_M.gguf | | | code_interpreter | {"code": "print('Hello, world!')"} | bartowski/gemma-2-2b-it-GGUF | gemma-2-2b-it-Q4_K_M.gguf | |
| code_interpreter | {"code": "print('Hello, World!')"} | bartowski/Mistral-Nemo-Instruct-2407-GGUF | Mistral-Nemo-Instruct-2407-Q4_K_M.gguf | mistralai-Mistral-Nemo-Instruct-2407 | | code_interpreter | {"code": "print('Hello, World!')"} | bartowski/Mistral-Nemo-Instruct-2407-GGUF | Mistral-Nemo-Instruct-2407-Q4_K_M.gguf | mistralai-Mistral-Nemo-Instruct-2407 |
| code_interpreter | {"code": "print(\"Hello World\")"} | bartowski/Qwen2.5-7B-Instruct-GGUF | Qwen2.5-7B-Instruct-Q4_K_M.gguf | | | code_interpreter | {"code": "print(\"Hello World\")"} | bartowski/Qwen2.5-7B-Instruct-GGUF | Qwen2.5-7B-Instruct-Q4_K_M.gguf | |
@ -113,8 +114,7 @@ Feature: llama.cpp server
| code_interpreter | {"code": "print('Hello, World!'}"} | lmstudio-community/Llama-3.2-1B-Instruct-GGUF | Llama-3.2-1B-Instruct-Q4_K_M.gguf | meta-llama-Llama-3.2-3B-Instruct | | code_interpreter | {"code": "print('Hello, World!'}"} | lmstudio-community/Llama-3.2-1B-Instruct-GGUF | Llama-3.2-1B-Instruct-Q4_K_M.gguf | meta-llama-Llama-3.2-3B-Instruct |
| code_interpreter | {"code": "print("} | lmstudio-community/Llama-3.2-3B-Instruct-GGUF | Llama-3.2-3B-Instruct-Q4_K_M.gguf | meta-llama-Llama-3.2-3B-Instruct | | code_interpreter | {"code": "print("} | lmstudio-community/Llama-3.2-3B-Instruct-GGUF | Llama-3.2-3B-Instruct-Q4_K_M.gguf | meta-llama-Llama-3.2-3B-Instruct |
| code_interpreter | {"code": "print("} | lmstudio-community/Meta-Llama-3.1-8B-Instruct-GGUF | Meta-Llama-3.1-8B-Instruct-Q4_K_M.gguf | | | code_interpreter | {"code": "print("} | lmstudio-community/Meta-Llama-3.1-8B-Instruct-GGUF | Meta-Llama-3.1-8B-Instruct-Q4_K_M.gguf | |
# | python | {"code": "print('Hello, World!')"} | bartowski/functionary-small-v3.2-GGUF | functionary-small-v3.2-Q8_0.gguf | meetkai-functionary-medium-v3.2 | | code_interpreter | {"code": "print('Hello, World!')"} | bartowski/functionary-small-v3.2-GGUF | functionary-small-v3.2-Q8_0.gguf | meetkai-functionary-medium-v3.2 |
# | code_interpreter | {"code": "print('Hello, World!')"} | bartowski/functionary-small-v3.2-GGUF | functionary-small-v3.2-Q8_0.gguf | meetkai-functionary-medium-v3.2 |
@slow @slow