I am unable to connect gptscript to Azure's implementation of OpenAI.
I have a gpt 3.5 turbo deployment on Azure OpenAI named, coincidentally enough, gpt-35-turbo.
It is reachable with:
curl https://<endpoint>.openai.azure.com/openai/deployments/gpt-35-turbo/completions?api-version=2023-05-15 -H "Content-Type: application/json" -H "api-key: <key>" -d "{
\"prompt\": \"Once upon a time\",
\"max_tokens\": 5
}"
and returns, for example:
{"id":"cmpl-8uobBPlV6d0UDqVFPwQdKEkZJaGgl","object":"text_completion","created":1708551561,"model":"gpt-35-turbo","choices":[{"text":" on a small island,","index":0,"finish_reason":"length","logprobs":null}],"usage":{"prompt_tokens":4,"completion_tokens":5,"total_tokens":9}}
16:45:22 started [main] [input=--input Hello, World! --openai-api-key <key> --openai-api-type AZURE --openai-base-url https://<endpoint>.openai.azure.com/openai/deployments/gpt-35-turbo/completions --openai-api-version 2023-05-15]
16:45:22 sent [main]
content [1] content | Waiting for model response...2024/02/21 16:45:22 error, status code: 404, message: Resource not found
16:51:33 started [main] [input=--input Hello, World! --debug]
16:51:33 sent [main]
content [1] content | Waiting for model response...2024/02/21 16:51:33 error, status code: 404, message: Resource not found
I am unable to connect gptscript to Azure's implementation of OpenAI.
I have a gpt 3.5 turbo deployment on Azure OpenAI named, coincidentally enough, gpt-35-turbo.
It is reachable with:
and returns, for example:
With gptscript, I am running:
and getting:
I can reproduce the same on Linux:
Gets: