silva96 / ai_stories Goto Github PK

We want to explore how are the stories generated using the new chatgpt api model

Here's the documentation of the new model

https://platform.openai.com/docs/guides/chat

It should be as easy as changing the usage of the old endpoint for the new one and configure parameters properly.

But first we need to update the open ai gem.

We are using this gem:

https://github.com/alexrudall/ruby-openai

Otherwise you wont be able to use the new endpoint.

Please make sure you update every usage of the gem to the new endpoint.

They are mostly used in the jobs (app/jobs directory)

Take as an example this:

@client.completions(
      parameters: {
        model: 'text-davinci-003',
        prompt:, max_tokens: (DAVINCI_MAX_TOKENS - prompt.size)
      }
    )

That's using davinci 003

You need to update those calls to use gpt3.5-turbo

Please refer to the gem's documentation to do it accordingly.

You would also want to update the DAVINCI_MAX_TOKENS constant to 4096

The maximum number of tokens allowed for the generated answer. By default, the number of tokens the model can return will be (4096 - prompt tokens).

I am unable to connect to OpenAI when I enter my token without enclosing in quotations:
i.e. OPENAI_ACCESS_TOKEN=sk-randomcharacters

I have no issues connecting to OpenAI if I enter my token enclosed with quotes:
i.e. OPENAI_ACCESS_TOKEN="sk-randomcharacters"

Hi, I've done all the steps. I can also access the web ui, but after entering the prompt, it just keeps loading, any idea what should be done, please note I've also updated the open ai api.