latest openai embedding support is not working and still assuming the model name is from hugging face

[Felixchong882003]

Describe the bug
I saw the recent update 956bb63 on v1.25.0-rc1 says it is supporting latest OpenAI embedding model text-embedding-3-large and text-embedding-3-small, But when i use EmbeddingRetriever, it still try to find the huggingface model of text-embedding-3-large, instead of using OpenAI embedding.

Error message

 text-embedding-3-large is not a local folder and is not a valid model identifier listed on 'https://huggingface.co/models'
If this is a private repository, make sure to pass a token having permission to this repo either by logging in with `huggingface-cli login` or by passing `token=<your_token>`

Expected behavior
OpenAI instead of searching Huggingface for the OpenAI embedding model

Additional context
Seems like the new model name is not included in _infer_model_format of EmbeddingRetriever.

To Reproduce
initialize a retriever like the follow:

retriever = EmbeddingRetriever(
                    document_store=document_store,
                    embedding_model='text-embedding-3-large',
                    model_format=None,
                    use_gpu=False,
                    batch_size=8,
                    api_key=api_key,
                    max_seq_len=8192,
)

FAQ Check

• Have you had a look at our new FAQ page?

System:

• OS: Debian GNU/Linux 10 (buster)
• GPU/CPU: Tesla T4 /Intel(R) Xeon(R) CPU @ 2.20GHz (GCP n1-standard-4)
• Haystack version (commit or version number): v1.25.0-rc1
• DocumentStore: elastic search
• Reader: N/A
• Retriever: text-embedding-3-large

[anakin87]

Thanks for reporting the issue!

A fix is coming...

[anakin87]

fixed in #7297

[mrm1001]

Hi @Felixchong882003, did the fix help solve your problem? I would love to hear more about how you're using Haystack.