Add support for Azure, OpenAI, Palm, Anthropic, Cohere, Replicate Models - using litellm

[ishaan-jaff]

This PR adds support for models from all the above mentioned providers using litellm https://github.com/BerriAI/litellm

TLDR: gorilla gets:

• more models out of the box, everything litellm integrates with: https://litellm.readthedocs.io/en/latest/supported/
• instant integration of new models by modifying the model param (Easy to add models in the future, just change the model param)

Here's a sample of how it's used:

from litellm import completion

## set ENV variables
# ENV variables can be set in .env file, too. Example in .env.example
os.environ["OPENAI_API_KEY"] = "openai key"
os.environ["COHERE_API_KEY"] = "cohere key"

messages = [{ "content": "Hello, how are you?","role": "user"}]

# openai call
response = completion(model="gpt-3.5-turbo", messages=messages)

# cohere call
response = completion("command-nightly", messages)

# anthropic call
response = completion(model="claude-instant-1", messages=messages)

[ishaan-jaff]

@ShishirPatil can you please take a look at this ? Happy to add docs/tests if this initial commit looks good😊

[ishaan-jaff]

litellm manages model I/O for your model choice

[ShishirPatil]

Hello @ishaan-jaff, thank you for your PR! It's impressive to see a wrapper that can accommodate multiple models. However, I'm a bit concerned about introducing an additional dependency that might not be essential for standard use cases. While I see its potential value, I'm contemplating whether it's best to include it in a separate release or integrate it into the main releases. I'm flexible on this. Could you initiate a discussion thread titled "Request for Comments - LiteLLM Integration" so we can gather community feedback? Based on that I can review this appropriately. Though at a high level, it looks good, since you basically count on the user to create the message.
Thoughts?

[rajveer43]

Hey I would Like to contribute to this PR...

[ishaan-jaff]

Thanks for the feedback @ShishirPatil

• started a discussion here: Request for Comments - LiteLLM Integration #89
• as for adding additional dependencies, we can address that by only requiring dependencies when they are used

eg. only install litellm anthropic if a user is trying to use claude-2

[rajveer43]

Any suggestions where should I start from? and what is open to implement?

[krrishdholakia]

@ShishirPatil @rajveer43 happy to help as well :)