EmergenceAI / Agent-E

Agent driven automation starting with the web. Discord: https://discord.gg/wgNfmFuqJF
MIT License
815 stars 114 forks source link

[QUERY] I can't execute any commands #69

Open xktm2 opened 4 months ago

xktm2 commented 4 months ago

After configuring Agent-E, I can successfully open Google, but I am unable to execute any commands. What could be the reason for this issue? (Agent-E) D:\AGENTE\Agent-E>python -m ae.main

Creating AutogenWrapper with ['user', 'planner_agent', 'browser_nav_agent', 'browser_nav_executor'] and 1000 rounds. INFO:ae.utils.logger:Using model gpt-3.5-turbo-1106 for AutoGen from the environment variables. Created browser_nav_executor_agent: <ae.utils.autogen_sequential_function_call.UserProxyAgent_SequentialFunctionExecution object at 0x00000182ADF07A10> INFO:ae.utils.logger:User preferences loaded from: D:\AGENTE\Agent-E\ae\user_preferences\user_preferences.txt INFO:ae.utils.logger:User preferences loaded from: D:\AGENTE\Agent-E\ae\user_preferences\user_preferences.txt INFO:ae.utils.logger:User dir: C:\Users\Administrator\AppData\Local\Google\Chrome\User Data\Default\Default INFO:ae.utils.logger:Received command: 打开微博 INFO:ae.utils.logger:Processing command: 打开微博 INFO:ae.utils.logger:Prompt for command: Execute the user task "打开微博" Current Page: https://www.google.com/ user (to planner_agent):

Execute the user task "打开微博" Current Page: https://www.google.com/


INFO:openai._base_client:Retrying request to /chat/completions in 0.773081 seconds INFO:openai._base_client:Retrying request to /chat/completions in 1.584054 seconds

deepak-akkil commented 4 months ago

Hi @xktm2 , It appears to be an issue related to openAI usage. Could you confirm the following: a) You have the latest set of dependencies installed from the pyproject.toml. Specifically the nest-asyncio package. b) Can you double check that the OpenAI key you are using is valid.

Once you confirm the above, i would suggest that you change the log levels to Debug in ae>utils>logger.py https://github.com/EmergenceAI/Agent-E/blob/master/ae/utils/logger.py#L8C1-L8C40

The debug logs should give us some clues on whats going on. Please post the logs here.

On a side note, once the issue is fixed, for best results, I would recommend to use gpt-4-turbo or gpt-4o.

luaonze commented 4 months ago

我用我自己的模型 AUTOGEN_MODEL_NAME=NotRequired AUTOGEN_MODEL_API_KEY=NotRequired AUTOGEN_MODEL_BASE_URL=http://127.0.0.1:11434

BROWSER_STORAGE_DIR=/Users/tom/Library/Application Support/Google/Chrome/Default 在ollama上使用llama3 也是会出现这个问题

teaxio commented 4 months ago

@luaonze are you sure your comment above is supposed to go in this issue? The logs from @xktm2 show the use of gpt-3.5

luaonze commented 4 months ago

im sure

verbraucher commented 1 week ago

I am experiencing a similar issue with Agent-E and the Default subdirectory in the User Data folder on Windows 10.

After configuring Agent-E to use my Chrome profile directory (C:\Users\danyo\AppData\Local\Google\Chrome\User Data\Profile 1), I noticed that Agent-E automatically appends an additional Default subdirectory, resulting in a path like C:\Users\danyo\AppData\Local\Google\Chrome\User Data\Profile 1\Default. This causes Chrome to create a new profile instead of using my existing one, which affects login status and other saved session data.

Any advice on how to prevent Agent-E from adding the Default subdirectory would be nice :)