Not respecting OpenAI model selection

[Kevin-Mok]

Describe the bug

The bot is responding twice, one with the correct model (small) and another with an unselected model (large) despite setting it the character file as such from #845, reverting the hardcoded value from #853 (Kevin-Mok@466782c) and setting [SMALL|MEDIUM|LARGE]_OPENAI_MODEL=gpt-4o-mini in .env.

To Reproduce

1. git clone https://github.com/Kevin-Mok/eliza.git
   

2. Set .env like so.

   SMALL_OPENAI_MODEL=gpt-4o-mini
   MEDIUM_OPENAI_MODEL=gpt-4o-mini
   LARGE_OPENAI_MODEL=gpt-4o-mini
   

3. Set the character config like so.

   {
   "modelProvider": "openai",
     "settings": {
       "model": "small"
     }
   }
   

4. pnpm start --character="path/to/character.json"

Expected behavior

Only the small model should be used.

Logs

Pre-Large Usage

 ◎ LOGS
   Creating Memory 
   2584cfba-64a2-0306-8b33-0b5958bcc298 
   youre stuck 

Message: {
  content: {
    text: 'youre stuck',
    attachments: [],
    source: 'discord',
    url: 'https://discord.com/channels/401422081025245184/1315441204715851826/1317575064295833640',
    inReplyTo: '96bb78ce-e84e-031f-9946-c0566ef94bb4'
  },
  userId: '6b35833a-b50c-0d49-8449-0941f15cf0ee',
  agentId: '03419646-efb1-0b57-a0ec-0027c41d97ae',
  roomId: 'a3464273-b491-037e-992c-a1e51f03ed03'
}
Validating transfer from user: 6b35833a-b50c-0d49-8449-0941f15cf0ee
Message: {
  content: {
    text: 'youre stuck',
    attachments: [],
    source: 'discord',
    url: 'https://discord.com/channels/401422081025245184/1315441204715851826/1317575064295833640',
    inReplyTo: '96bb78ce-e84e-031f-9946-c0566ef94bb4'
  },
  userId: '6b35833a-b50c-0d49-8449-0941f15cf0ee',
  agentId: '03419646-efb1-0b57-a0ec-0027c41d97ae',
  roomId: 'a3464273-b491-037e-992c-a1e51f03ed03'
}

 ["◎ Generating message response.."] 

 ["◎ Generating text..."] 

 ℹ INFORMATIONS
   Generating text with options: 
   {"modelProvider":"openai","model":"small"} 

 ℹ INFORMATIONS
   Selected model: 
   gpt-4o-mini 

 ◎ LOGS
   Creating Memory 
   3ae62ff6-70a5-0972-9d38-8cb2ed3aedcc 
   stuck? nah, just recharging my creative circuits! what’s got you feeling that way? 

 ["✓ Normalized action: continue"] 

 ["ℹ Executing handler for action: CONTINUE"] 

 ["◎ Generating text..."] 

 ℹ INFORMATIONS
   Generating text with options: 
   {"modelProvider":"openai","model":"small"} 

 ℹ INFORMATIONS
   Selected model: 
   gpt-4o-mini 

 ["◎ Generating message response.."] 

Large Model Usage (cont. from above)

 ["◎ Generating text..."] 

 ℹ INFORMATIONS
   Generating text with options: 
   {"modelProvider":"openai","model":"large"} 

 ℹ INFORMATIONS
   Selected model: 
   gpt-4o 

 ◎ LOGS
   Creating Memory 
   975c6fda-78f3-0bf0-b6f9-ee7780e86797 
   nah, just recharging my creative circuits! what's got you feeling that way? maybe we can troubleshoot it together. 

 ◎ LOGS
   Evaluating 
   GET_FACTS 

 ◎ LOGS
   Evaluating 
   UPDATE_GOAL 

 ◎ LOGS
   Evaluating 
   EXTRACT_RECOMMENDATIONS 

 ["◎ Generating text..."] 

 ℹ INFORMATIONS
   Generating text with options: 
   {"modelProvider":"openai","model":"small"} 

 ℹ INFORMATIONS
   Selected model: 
   gpt-4o-mini 

 ["◎ Generating text..."] 

 ℹ INFORMATIONS
   Generating text with options: 
   {"modelProvider":"openai","model":"small"} 

 ℹ INFORMATIONS
   Selected model: 
   gpt-4o-mini 

Skipping process

[oxSaturn]

Try to sync your main branch to the latest commit and try again, I belive configuring in .env should work fine. See PR here #999.

[Kevin-Mok]

This worked, thanks. I'm assuming it respects my LARGE_OPENAI_MODEL=gpt-4o-mini in .env due to #999 now. Thanks for the PR!

 ["◎ Generating text..."] 

 ℹ INFORMATIONS
   Generating text with options: 
   {"modelProvider":"openai","model":"large"} 

 ℹ INFORMATIONS
   Selected model: 
   gpt-4o-mini