Added ability to configure concurrent_requests in litellm_model.py

[dameikle]

This pull request updates the way concurrent API requests are configured for the LiteLLM model endpoint. Instead of using a hardcoded value, the number of concurrent requests is now configurable through the model's configuration object, making the code more flexible and easier to tune for different environments.

Configuration improvements:

• Added a new field concurrent_requests to the LiteLLMModelConfig class, allowing the number of concurrent API requests to be set via configuration instead of being hardcoded.
• Updated the LiteLLMClient class to use the new concurrent_requests configuration value, removing the old hardcoded CONCURRENT_CALLS attribute.

Concurrency handling:

• Modified the __call_api_parallel method to use the configurable concurrent_requests value when creating the ThreadPoolExecutor, improving flexibility and maintainability.

[NathanHB]

hey ! thanks for the fix, if the tests pass we are good to merge

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[dameikle]

hey ! thanks for the fix, if the tests pass we are good to merge

Thanks for the review @NathanHB 🙇 I've updated with the suggested change.

[NathanHB]

Hey, this is good to merge, i just missed that you need to add the new parametr to the docstring, thanks !

[dameikle]

Hey, this is good to merge, i just missed that you need to add the new parametr to the docstring, thanks !

Sorry totally should have thought of that. Update made now