Evaluate the model on a server or container

An alternative to launching the evaluation locally is to serve the model on a TGI-compatible server/container and then run the evaluation by sending requests to the server. The command is the same as before, except you specify a path to a yaml config file (detailed below):

lighteval endpoint {tgi,inference-endpoint} \
    "/path/to/config/file"\
    <task parameters>

There are two types of configuration files that can be provided for running on the server:

Hugging Face Inference Endpoints

To launch a model using HuggingFace’s Inference Endpoints, you need to provide the following file: endpoint_model.yaml. Lighteval will automatically deploy the endpoint, run the evaluation, and finally delete the endpoint (unless you specify an endpoint that was already launched, in which case the endpoint won’t be deleted afterwards).

configuration file example:

model:
  base_params:
    # Pass either model_name, or endpoint_name and true reuse_existing
    # endpoint_name: "llama-2-7B-lighteval" # needs to be lower case without special characters
    # reuse_existing: true # defaults to false; if true, ignore all params in instance, and don't delete the endpoint after evaluation
    model_name: "meta-llama/Llama-2-7b-hf"
    # revision: "main" # defaults to "main"
    dtype: "float16" # can be any of "awq", "eetq", "gptq", "4bit' or "8bit" (will use bitsandbytes), "bfloat16" or "float16"
  instance:
    accelerator: "gpu"
    region: "eu-west-1"
    vendor: "aws"
    instance_type: "nvidia-a10g"
    instance_size: "x1"
    framework: "pytorch"
    endpoint_type: "protected"
    namespace: null # The namespace under which to launch the endpoint. Defaults to the current user's namespace
    image_url: null # Optionally specify the docker image to use when launching the endpoint model. E.g., launching models with later releases of the TGI container with support for newer models.
    env_vars:
      null # Optional environment variables to include when launching the endpoint. e.g., `MAX_INPUT_LENGTH: 2048`

Text Generation Inference (TGI)

To use a model already deployed on a TGI server, for example on HuggingFace’s serverless inference.

configuration file example:

model:
  instance:
    inference_server_address: ""
    inference_server_auth: null
    model_id: null # Optional, only required if the TGI container was launched with model_id pointing to a local directory

OpenAI API

Lighteval also supports evaluating models on the OpenAI API. To do so you need to set your OpenAI API key in the environment variable.

export  OPENAI_API_KEY={your_key}

And then run the following command:

lighteval endpoint openai \
    {model-name} \
    <task parameters>

< > Update on GitHub