add huggingface

docs/my-website/docs/providers/huggingface.md

@@ -0,0 +1,56 @@
+import Image from '@theme/IdealImage';
+
+# Huggingface
+
+LiteLLM supports Huggingface Inference Endpoints. It uses the [text-generation-inference](https://github.com/huggingface/text-generation-inference) format. You can use any chat/text model from Hugging Face with the following steps:
+
+* Copy your model id/url from Huggingface Inference Endpoints
+    - [ ] Go to https://ui.endpoints.huggingface.co/
+    - [ ] Copy the url of the specific model you'd like to use 
+    <Image img={require('../../img/hf_inference_endpoint.png')} alt="HF_Dashboard" style={{ maxWidth: '50%', height: 'auto' }}/>
+* Set it as your model name
+* Set your HUGGINGFACE_API_KEY as an environment variable
+
+Need help deploying a model on huggingface? [Check out this guide.](https://huggingface.co/docs/inference-endpoints/guides/create_endpoint)
+
+## usage 
+
+In this case our model id is the same as the model url - `https://ji16r2iys9a8rjk2.us-east-1.aws.endpoints.huggingface.cloud`
+
+
+
+You need to tell LiteLLM when you're calling Huggingface. You can do that in 2 ways:
+
+* By passing in the custom llm provider as part of the model name -  
+completion(model="<custom_llm_provider>/<model_id>",...). 
+
+```
+import os 
+from litellm import completion 
+
+# Set env variables
+os.environ["HUGGINGFACE_API_KEY"] = "huggingface_api_key"
+
+messages = [{ "content": "There's a llama in my garden 😱 What should I do?","role": "user"}]
+
+# model = <custom_llm_provider>/<model_id>
+response = completion(model="huggingface/https://ji16r2iys9a8rjk2.us-east-1.aws.endpoints.huggingface.cloud", messages=messages)
+
+print(response)
+```
+
+* By passing in a `custom_llm_provider` argument in the completion call
+
+```
+import os 
+from litellm import completion 
+
+# Set env variables
+os.environ["HUGGINGFACE_API_KEY"] = "huggingface_api_key"
+
+messages = [{ "content": "There's a llama in my garden 😱 What should I do?","role": "user"}]
+
+response = completion(model="https://ji16r2iys9a8rjk2.us-east-1.aws.endpoints.huggingface.cloud", messages=messages, custom_llm_provider="huggingface")
+# Add any assertions here to check the response
+print(response)
+```

docs/my-website/sidebars.js

@@ -30,6 +30,17 @@ const sidebars = {
       items: ["embedding/supported_embedding"],
     },
     'completion/supported',
+    {
+      type: "category",
+      label: "Providers",
+      link: {
+        type: 'generated-index',
+        title: 'Providers',
+        description: 'Learn how to deploy + call models from different providers on LiteLLM',
+        slug: '/providers',
+      },
+      items: ["providers/huggingface"],
+    },
     "token_usage",
     "exception_mapping",
     'debugging/hosted_debugging',

litellm/llms/huggingface_restapi.py

@@ -46,8 +46,8 @@ class HuggingfaceRestAPILLM:
         logger_fn=None,
     ):  # logic for parsing in - calling - parsing out model completion calls
         completion_url: str = ""
-        if custom_api_base:
+        if "https" in model:
-            completion_url = custom_api_base
+            completion_url = model
         elif "HF_API_BASE" in os.environ:
             completion_url = os.getenv("HF_API_BASE", "")
         else:

litellm/tests/test_completion.py

@@ -87,7 +87,7 @@ def test_completion_hf_deployed_api():
     try:
         user_message = "There's a llama in my garden 😱 What should I do?"
         messages = [{ "content": user_message,"role": "user"}]
-        response = completion(model="huggingface/WizardLM/WizardCoder-Python-34B-V1.0", messages=messages, custom_api_base="https://ji16r2iys9a8rjk2.us-east-1.aws.endpoints.huggingface.cloud", logger_fn=logger_fn)
+        response = completion(model="huggingface/https://ji16r2iys9a8rjk2.us-east-1.aws.endpoints.huggingface.cloud", messages=messages, logger_fn=logger_fn)
         # Add any assertions here to check the response
         print(response)
     except Exception as e: