LLM模型除了需要Function Call，还需要Format Tool Call吗？

[Chanpoe]

我们测试了很多模型都不支持format tool call，openai的gpt-4o倒是比较稳定，大家有没有推荐的可用模型？

[apoet]

https://ollama.com/hhao/qwen2.5-coder-tools

[z-x-x136]

Global LLM configuration

[llm]
model = "hhao/qwen2.5-coder-tools:7b"
base_url = "http://localhost:11435/v1"
api_key = "ollama"
max_tokens = 2046
temperature = 0.6

Optional configuration for specific LLM models

[llm.vision]
model = "claude-3-5-sonnet"
base_url ="https://api.openai.com/v1"
api_key="sk-..."

(E:\conda_env\open_manus) H:\git_model_demo\OpenManus-main>python main.py
INFO [browser_use] BrowserUse logging setup complete with level info
INFO [root] Anonymized telemetry enabled. See https://docs.browser-use.com/development/telemetry for more information.
Enter your prompt (or 'exit' to quit): 你好
2025-03-07 21:54:52.775 | WARNING | main:main:15 - Processing your request...
2025-03-07 21:54:52.791 | INFO | app.agent.base:run:137 - Executing step 1/30
2025-03-07 21:54:54.853 | ERROR | app.llm:ask_tool:250 - API error: Error code: 502
2025-03-07 21:54:57.797 | ERROR | app.llm:ask_tool:250 - API error: Error code: 502
2025-03-07 21:55:01.112 | ERROR | app.llm:ask_tool:250 - API error: Error code: 502

[apoet]

Global LLM configuration

[llm] model = "hhao/qwen2.5-coder-tools:7b" base_url = "http://localhost:11435/v1" api_key = "ollama" max_tokens = 2046 temperature = 0.6

Optional configuration for specific LLM models

[llm.vision] model = "claude-3-5-sonnet" base_url ="https://api.openai.com/v1" api_key="sk-..."

(E:\conda_env\open_manus) H:\git_model_demo\OpenManus-main>python main.py INFO [browser_use] BrowserUse logging setup complete with level info INFO [root] Anonymized telemetry enabled. See https://docs.browser-use.com/development/telemetry for more information. Enter your prompt (or 'exit' to quit): 你好 2025-03-07 21:54:52.775 | WARNING | main:main:15 - Processing your request... 2025-03-07 21:54:52.791 | INFO | app.agent.base:run:137 - Executing step 1/30 2025-03-07 21:54:54.853 | ERROR | app.llm:ask_tool:250 - API error: Error code: 502 2025-03-07 21:54:57.797 | ERROR | app.llm:ask_tool:250 - API error: Error code: 502 2025-03-07 21:55:01.112 | ERROR | app.llm:ask_tool:250 - API error: Error code: 502

502是网关错误，确定是11435端口？

[z-x-x136]

Global LLM configuration  全局LLM配置

[llm] model = "hhao/qwen2.5-coder-tools:7b" base_url = "http://localhost:11435/v1" api_key = "ollama" max_tokens = 2046 temperature = 0.6[llm] model = “hhao/qwen2.5-coder-tools：7b” base_url = “http://localhost:11435/v1” api_key = “ollama” max_tokens = 2046 温度 = 0.6

Optional configuration for specific LLM models特定LLM型号的可选配置

[llm.vision] model = "claude-3-5-sonnet" base_url ="https://api.openai.com/v1" api_key="sk-..."[llm.vision] model = “claude-3-5-sonnet” base_url =“https://api.openai.com/v1” api_key=“sk-...”
(E:\conda_env\open_manus) H:\git_model_demo\OpenManus-main>python main.py INFO [browser_use] BrowserUse logging setup complete with level info INFO [root] Anonymized telemetry enabled. See https://docs.browser-use.com/development/telemetry for more information. Enter your prompt (or 'exit' to quit): 你好 2025-03-07 21:54:52.775 | WARNING | main:main:15 - Processing your request... 2025-03-07 21:54:52.791 | INFO | app.agent.base:run:137 - Executing step 1/30 2025-03-07 21:54:54.853 | ERROR | app.llm:ask_tool:250 - API error: Error code: 502 2025-03-07 21:54:57.797 | ERROR | app.llm:ask_tool:250 - API error: Error code: 502 2025-03-07 21:55:01.112 | ERROR | app.llm:ask_tool:250 - API error: Error code: 502（E：\conda_env\open_manus）H：\git_model_demo\OpenManus-main>python main.py INFO [browser_use] 浏览器使用带有级别信息 INFO 的日志记录设置[root] 启用匿名遥测。有关更多信息，请参阅 https://docs.browser-use.com/development/telemetry。输入你的提示（或'退出'退出）： 你好 2025-03-07 21：54：52.775 |警告 |main：main：15 - 正在处理您的请求...2025-03-07 21：54：52.791 |信息 |app.agent.base：run：137 - 执行步骤 1/30 2025-03-07 21：54：54.853 |错误 |app.llm：ask_tool：250 - API 错误：错误代码：502 2025-03-07 21：54：57.797 |错误 |app.llm：ask_tool：250 - API 错误：错误代码：502 2025-03-07 21：55：01.112 |错误 |app.llm：ask_tool：250 - API 错误：错误代码：502

502是网关错误，确定是11435端口？

http://localhost:11435/ Ollama is running 确定，我改为11435了

[Chanpoe]

Global LLM configuration

[llm] model = "hhao/qwen2.5-coder-tools:7b" base_url = "http://localhost:11435/v1" api_key = "ollama" max_tokens = 2046 temperature = 0.6

Optional configuration for specific LLM models

[llm.vision] model = "claude-3-5-sonnet" base_url ="https://api.openai.com/v1" api_key="sk-..."

(E:\conda_env\open_manus) H:\git_model_demo\OpenManus-main>python main.py INFO [browser_use] BrowserUse logging setup complete with level info INFO [root] Anonymized telemetry enabled. See https://docs.browser-use.com/development/telemetry for more information. Enter your prompt (or 'exit' to quit): 你好 2025-03-07 21:54:52.775 | WARNING | main:main:15 - Processing your request... 2025-03-07 21:54:52.791 | INFO | app.agent.base:run:137 - Executing step 1/30 2025-03-07 21:54:54.853 | ERROR | app.llm:ask_tool:250 - API error: Error code: 502 2025-03-07 21:54:57.797 | ERROR | app.llm:ask_tool:250 - API error: Error code: 502 2025-03-07 21:55:01.112 | ERROR | app.llm:ask_tool:250 - API error: Error code: 502

你的claude模型用的是openai的接口

# 你的配置
model = "claude-3-5-sonnet"
base_url ="https://api.openai.com/v1"

可能是这个原因导致的502，当然也不排除ollama配置有误