[Help] [Python] [API] Params Default

API
1

For the gpt-3.5-turbo, text-davinci-003 and text-davinci-002 models, what are the default values for each parameter?

In python is it indifferent to make a request by JSON or dict?

2

Welcome @tiago.feitoza

You’ll find all of them in the API Reference

4

I want to complete the tables and I can’t find all the data in the documentation.

image
image560×746 26.2 KB


Python v1/chat/completions
Attributo Tipo Status Default Limit
frequency_penalty number Optional 0 -2,0 ~ 2,0
logit_bias map Optional null {?:?}
max_tokens integer Optional Defaults to inf ?
messages array Required ? ?
model string Required ? ?
n integer Optional 1 ?
presence_penalty number Optional 0 -2,0 ~ 2,0
stop string or array Optional null and [,]
stream boolean Optional false false or true
temperature number Optional 1 0 ~ 2
top_p number Optional 1 0 ~ 100
user string Optional ? ?

Python v1/completions
Attributo Tipo Status Default Limit
best_of integer Optional 1 1 ~ ?
echo boolean Optional false
frequency_penalty number Optional 0 -2,0 ~ 2.0
logit_bias map Optional null {“”: -100 ~ 100}
logprobs integer Optional null 1 ~ 5
max_tokens integer Optional 16 0 ~ ?
model string Required ? *
n integer Optional 1 1~
presence_penalty number Optional 0 -2,0 ~ 2.0
prompt string or array Optional ? ?
stop string or array Optional null ~ [,]
stream boolean Optional false false or true
suffix string Optional null
temperature number Optional 1 0 ~ 2
top_p number Optional 1 0 ~ 100
user string Optional ? ?

I would like this information to check before sending a request, overwriting the default parameters with the wrong type or limit value

I have built “default” requests as follows:

"api_chat_completion": {
    "frequency_penalty": 0,
    "log_bias": {},
    "max_tokens": 256,
    "messages": [
        {
            "role": "user",
            "content": "select_prompt + input_field"

        }
    ],
    "model": null,
    "n": 1,
    "presence_penalty": 0,
    "stop": "",
    "stream": false,
    "temperature": 0.2,
    "temperature_decay_rate": 0,
    "top_p": 1,
    "user": ""
},
"api_completion": {
    "best_of": 1,
    "echo": false,
    "frequency_penalty": 0,
    "logit_bias": {},
    "logprobs": null,
    "max_tokens": 256,
    "model": null,
    "n": 1,
    "presence_penalty": 0,
    "prompt": "",
    "stop": "",
    "stream": false,
    "suffix": "",
    "temperature": 0.2,
    "temperature_decay_rate": 0,
    "top_p": 1,
    "user": ""
},