control temperature in chatgpt via chatgpt WEB (pandora)

[jaodei]

Hi,

I see in v3.py, we are able to change the temperature of chatgpt, which you default to 0.5. Does it also apply when we use chatgpt via chatgpt WEB (pandora)?

Also is text-davinci-002-render-sha-mobile the only or the best model in pandora_api_model? Any other options?

Thank you.

[hibobmaster]

We can't change temperature since it is controlled by ChatGPT Web. v3.py use official chatgpt api endpoint which can specific temperature. I will delete v3.py in v1.2.0 to achieve session isolation(Use node-chatgpt-api instead like bing). By default, temperature uses in node-chatgpt-api is 0.8, i can provide an option to change it.

For free account we can use text-davinci-002-render-sha and text-davinci-002-render-sha-mobile :

{
    "models": [
        {
            "slug": "text-davinci-002-render-sha",
            "max_tokens": 8191,
            "title": "Turbo (Default for free users)",
            "description": "Our fastest model, great for most everyday tasks.",
            "tags": [
                "gpt3.5"
            ],
            "qualitative_properties": {}
        },
        {
            "slug": "text-davinci-002-render-sha-mobile",
            "max_tokens": 8191,
            "title": "Turbo (Default for free users)",
            "description": "Our fastest model, great for most everyday tasks.",
            "tags": [
                "mobile"
            ],
            "qualitative_properties": {}
        }
    ],
    "categories": [
        {
            "category": "gpt_3.5",
            "human_category_name": "GPT-3.5",
            "subscription_level": "free",
            "default_model": "text-davinci-002-render-sha",
            "browsing_model": null,
            "code_interpreter_model": null,
            "plugins_model": null
        }
    ]
}

If you have chatgpt plus, you can also get models following the steps below:

1. install pandora from pip pip install pandora-chatgpt
2. launch http service and login pandora -s
3. access http://127.0.0.1:8008/api/models to find which models you can use

[jaodei]

Well explained. Thank you!