OpenAI and Huggingface Providers #16
Conversation
| @@ -0,0 +1,9 @@ | |||
|
|
|||
| from redisvl.providers.openai import OpenAIProvider | |||
There was a problem hiding this comment.
should we consider organizing it one layer deeper, under embeddings? that way its clear whether its for Embedding creation or something else (future proofing)
There was a problem hiding this comment.
That's a long import but I see what you mean. A couple options
from redisvl.api.embedding.openai
from redisvl.embedding.openai import OpenAIEmbedding
from redisvl.model.embedding.openai import ...others?
| @@ -0,0 +1,29 @@ | |||
| from typing import Dict, List, Optional | |||
|
|
|||
| from sentence_transformers import SentenceTransformer | |||
There was a problem hiding this comment.
we could consider importing this scoped within the class -- and then we can prevent folks from having to install SentenceTransformers (torch, etc) if they don't plan to need it?
There was a problem hiding this comment.
It's the default for LLMCache... should we not assume that? I do like the scoped gaurd approach.
| @@ -0,0 +1,48 @@ | |||
| from typing import Dict, List, Optional | |||
|
|
|||
| import openai | |||
There was a problem hiding this comment.
same point as above?
| openai.api_key = api_config.get("api_key", None) | ||
| self._model_client = openai.Embedding | ||
|
|
||
| def embed_many( |
There was a problem hiding this comment.
should we look to make this immune to rate limit issues? One pattern I have seen is to use tenacity and their exponential backoff decorator.
There was a problem hiding this comment.
@tylerhutcherson can you point me to or write a code sample for this?
|
addressing changes in Next PR |
Providers for creating embeddings with OpenAI and Hugginface.
Includes