Cant change the model of custom-llm from gpt-3.5-...
# supporth
healthnow
10/08/2024, 8:22 PMAny reason I can't change the model of custom-llm from gpt-3.5-turbo?
https://cdn.discordapp.com/attachments/1293307512237326346/1293307512476536904/Screenshot_2024-10-08_at_1.16.03_PM.png?ex=6706e5fb&is=6705947b&hm=44ecb94e2126481f27245ab98ae7af8cc2ec4759bb305fe6bac995183bfe952b&
v
Vapi
10/08/2024, 8:24 PM@healthnow
The ability to change the model of a custom LLM (Large Language Model) from gpt-3.5-turbo to another model depends on the specific configurations and options provided by the API you're using. If you're working with VAPI, you would typically have the option to specify which model you want to use when making API calls, as long as the model you want to switch to is supported and available within the API's framework.
If you're encountering issues or limitations in changing the model, it could be due to several reasons:
1. API Restrictions: The API might have restrictions on changing models based on the plan you're subscribed to or the endpoint you're using.
2. Model Availability: The model you want to switch to might not be available or released for public use yet.
3. Configuration Constraints: There might be constraints in the API configuration that prevent changing the model without proper authorization or additional setup.
4. Technical Limitations: There could be technical limitations or compatibility issues with the new model that need to be addressed first.
If you provide more context or specify the error message you're receiving, I can offer more targeted assistance. Otherwise, if this is a technical issue with the API, it would be best to reach out to the API's support team for more detailed information.
Vapi
10/08/2024, 8:24 PMYou can continue this conversation by mentioning me in the message.
h
healthnow
10/08/2024, 8:33 PM@Vapi how do i call a different model if it wont let me switch the name from the vapi interface
healthnow
10/08/2024, 8:34 PM@User
n
naaviii
10/09/2024, 7:10 AM!urgent , please help us ASAP @User
naaviii
10/09/2024, 8:57 AM@Sahil
s
Shubham Bajaj
10/09/2024, 8:21 PM@naaviii with custom-llm model the request will go to your model server url so it won't make a difference.
Shubham Bajaj
10/09/2024, 8:21 PMAre your facing any issues?
n
naaviii
10/10/2024, 8:15 AMyes facing the same issue .
naaviii
10/10/2024, 8:15 AMI guess it is making difference (please check with you tech team)
s
Shubham Bajaj
10/10/2024, 6:41 PMcan you share call id along with expected vs actual response.
Shubham Bajaj
10/10/2024, 7:08 PM@SP can you please look into this.
c
confunius
10/12/2024, 1:13 PMi had faced the same issue, until i discovered that i was mislead by the documentation a little bit
confunius
10/12/2024, 1:13 PMconfunius
10/12/2024, 1:16 PM> Using Fine-Tuned OpenAI Models
> To set up your OpenAI Fine-Tuned model, you need to follow these steps:
> Set the custom llm URL to https://api.openai.com/v1.
> Assign the custom llm key to the OpenAI key.
> Update the model to their model.
> Execute a PATCH request to the /assistant endpoint and ensure that model.metadataSendMode is set to off.
i thought that we needed to do this all in the vapi assistants dashboard, but instead i had success using Postman or the vapi API reference: https://docs.vapi.ai/api-reference/assistants/update-assistant and executing a patch request to update my assistant with the fine tuned model of my choice
confunius
10/12/2024, 1:16 PMhopefully this can be of help!
s
Shubham Bajaj
10/15/2024, 4:22 PM@healthnow please check this.