Llama2 - Huggingface Tutorial

Huggingface is an open source platform to deploy machine-learnings models.

Call Llama2 with Huggingface Inference Endpoints

LiteLLM makes it easy to call your public, private or the default huggingface endpoints.

In this case, let's try and call 3 models:

Model	Type of Endpoint
deepset/deberta-v3-large-squad2	Default Huggingface Endpoint
meta-llama/Llama-2-7b-hf	Public Endpoint
meta-llama/Llama-2-7b-chat-hf	Private Endpoint

Case 1: Call default huggingface endpoint

Here's the complete example:

from litellm import completion 

model = "deepset/deberta-v3-large-squad2"
messages = [{"role": "user", "content": "Hey, how's it going?"}] # LiteLLM follows the OpenAI format 

### CALLING ENDPOINT
completion(model=model, messages=messages, custom_llm_provider="huggingface")

What's happening?

• model: This is the name of the deployed model on huggingface
• messages: This is the input. We accept the OpenAI chat format. For huggingface, by default we iterate through the list and add the message["content"] to the prompt. Relevant Code
• custom_llm_provider: Optional param. This is an optional flag, needed only for Azure, Replicate, Huggingface and Together-ai (platforms where you deploy your own models). This enables litellm to route to the right provider, for your model.

Case 2: Call Llama2 public Huggingface endpoint

We've deployed meta-llama/Llama-2-7b-hf behind a public endpoint - https://ag3dkq4zui5nu8g3.us-east-1.aws.endpoints.huggingface.cloud.

Let's try it out:

from litellm import completion 

model = "meta-llama/Llama-2-7b-hf"
messages = [{"role": "user", "content": "Hey, how's it going?"}] # LiteLLM follows the OpenAI format 
api_base = "https://ag3dkq4zui5nu8g3.us-east-1.aws.endpoints.huggingface.cloud"

### CALLING ENDPOINT
completion(model=model, messages=messages, custom_llm_provider="huggingface", api_base=api_base)

What's happening?

• api_base: Optional param. Since this uses a deployed endpoint (not the default huggingface inference endpoint), we pass that to LiteLLM.

Case 3: Call Llama2 private Huggingface endpoint

The only difference between this and the public endpoint, is that you need an api_key for this.

On LiteLLM there's 3 ways you can pass in an api_key.

Either via environment variables, by setting it as a package variable or when calling completion().

Setting via environment variables
Here's the 1 line of code you need to add

os.environ["HF_TOKEN"] = "..."

Here's the full code:

from litellm import completion 

os.environ["HF_TOKEN"] = "..."

model = "meta-llama/Llama-2-7b-hf"
messages = [{"role": "user", "content": "Hey, how's it going?"}] # LiteLLM follows the OpenAI format 
api_base = "https://ag3dkq4zui5nu8g3.us-east-1.aws.endpoints.huggingface.cloud"

### CALLING ENDPOINT
completion(model=model, messages=messages, custom_llm_provider="huggingface", api_base=api_base)

Setting it as package variable
Here's the 1 line of code you need to add

litellm.huggingface_key = "..."

Here's the full code:

import litellm
from litellm import completion 

litellm.huggingface_key = "..."

model = "meta-llama/Llama-2-7b-hf"
messages = [{"role": "user", "content": "Hey, how's it going?"}] # LiteLLM follows the OpenAI format 
api_base = "https://ag3dkq4zui5nu8g3.us-east-1.aws.endpoints.huggingface.cloud"

### CALLING ENDPOINT
completion(model=model, messages=messages, custom_llm_provider="huggingface", api_base=api_base)

Passed in during completion call

completion(..., api_key="...")

Here's the full code:

from litellm import completion 

model = "meta-llama/Llama-2-7b-hf"
messages = [{"role": "user", "content": "Hey, how's it going?"}] # LiteLLM follows the OpenAI format 
api_base = "https://ag3dkq4zui5nu8g3.us-east-1.aws.endpoints.huggingface.cloud"

### CALLING ENDPOINT
completion(model=model, messages=messages, custom_llm_provider="huggingface", api_base=api_base, api_key="...")