ValueError: Unknown model 'gpt-4o'.

[aiwalter]

I am getting

ValueError: Unknown model 'gpt-4o'. Please provide a valid OpenAI model name in: gpt-4, gpt-4-32k, gpt-4-1106-preview, gpt-4-vision-preview, gpt-4-0613, gpt-4-32k-0613, gpt-4-0314, gpt-4-32k-0314, gpt-3.5-turbo, gpt-3.5-turbo-16k, gpt-3.5-turbo-1106, gpt-3.5-turbo-0613, gpt-3.5-turbo-16k-0613, gpt-3.5-turbo-0301, text-davinci-003, text-davinci-002, gpt-3.5-turbo-instruct, text-ada-001, text-babbage-001, text-curie-001, ada, babbage, curie, davinci, gpt-35-turbo-16k, gpt-35-turbo

when running this

service_context = ServiceContext.from_defaults(
    embed_model=embed_model,
    llm=llm,
)

But when I rung the llm and embed_model separately they work well like doing this is working

embed_model.get_text_embedding("hi")
llm.complete("hi")

What is the issue here? I wanted to update my model to gpt-4o from Azure. I am using an older Llamaindex version llama-index==0.9.8.post1. Can I not use gpt-4o with this older version?

[dosubot[bot]]

You cannot use the 'gpt-4o' model with the older Llamaindex version llama-index==0.9.8.post1. The error ValueError: Unknown model 'gpt-4o' occurs because the model 'gpt-4o' is not recognized in that version. The list of known models is defined in the code, and 'gpt-4o' is included in a later version ^[1].

To resolve this issue, you need to upgrade to a version of LlamaIndex that includes support for the 'gpt-4o' model. The latest version that includes this support is 0.9.45 ^[2][3].

Keep in mind that there are backward compatibility issues when upgrading from llama-index==0.9.8.post1 to the latest version. Specifically, querying an index created in an older version using the latest version can result in errors such as:

ValueError: doc_type simple_dict not found in type_to_struct. Make sure that it was registered in the index registry.

Recreating indexes in the latest version is suggested to fix these problems ^[4][5].

To continue talking to Dosu, mention @dosu.

[aiwalter]

its seems like there is hardcoded some model versions, this is probably not the best idea:

GPT4_MODELS: Dict[str, int] = {
    # stable model names:
    #   resolves to gpt-4-0314 before 2023-06-27,
    #   resolves to gpt-4-0613 after
    "gpt-4": 8192,
    "gpt-4-32k": 32768,
    # 1106 model (Turbo, JSON mode)
    "gpt-4-1106-preview": 128000,
    # multimodal model
    "gpt-4-vision-preview": 128000,
    # 0613 models (function calling):
    #   https://openai.com/blog/function-calling-and-other-api-updates
    "gpt-4-0613": 8192,
    "gpt-4-32k-0613": 32768,
    # 0314 models
    "gpt-4-0314": 8192,
    "gpt-4-32k-0314": 32768,
}

@dosubot what do the integer values mean here?