HuggingFace

promptfoo includes support for the HuggingFace Inference API, for text generation, classification, and embeddings related tasks, as well as HuggingFace Datasets.

To run a model, specify the task type and model name. Supported models include:

• huggingface:text-generation:<model name>
• huggingface:text-classification:<model name>
• huggingface:token-classification:<model name>
• huggingface:feature-extraction:<model name>
• huggingface:sentence-similarity:<model name>

Examples

For example, autocomplete with GPT-2:

huggingface:text-generation:gpt2

Generate text with Mistral:

huggingface:text-generation:mistralai/Mistral-7B-v0.1

Embeddings similarity with sentence-transformers:

# Model supports the sentence similarity API
huggingface:sentence-similarity:sentence-transformers/all-MiniLM-L6-v2

# Model supports the feature extraction API
huggingface:feature-extraction:sentence-transformers/paraphrase-xlm-r-multilingual-v1

Configuration

These common HuggingFace config parameters are supported:

Parameter	Type	Description
top_k	number	Controls diversity via the top-k sampling strategy.
top_p	number	Controls diversity via nucleus sampling.
temperature	number	Controls randomness in generation.
repetition_penalty	number	Penalty for repetition.
max_new_tokens	number	The maximum number of new tokens to generate.
max_time	number	The maximum time in seconds model has to respond.
return_full_text	boolean	Whether to return the full text or just new text.
num_return_sequences	number	The number of sequences to return.
do_sample	boolean	Whether to sample the output.
use_cache	boolean	Whether to use caching.
wait_for_model	boolean	Whether to wait for the model to be ready. This is useful to work around the "model is currently loading" error

Additionally, any other keys on the config object are passed through directly to HuggingFace. Be sure to check the specific parameters supported by the model you're using.

The provider also supports these built-in promptfoo parameters:

Parameter	Type	Description
apiKey	string	Your HuggingFace API key.
apiEndpoint	string	Custom API endpoint for the model.

Supported environment variables:

• HF_API_TOKEN - your HuggingFace API key

The provider can pass through configuration parameters to the API. See text generation parameters and feature extraction parameters.

Here's an example of how this provider might appear in your promptfoo config:

providers:
  - id: huggingface:text-generation:mistralai/Mistral-7B-v0.1
    config:
      temperature: 0.1
      max_length: 1024

Inference endpoints

HuggingFace provides the ability to pay for private hosted inference endpoints. First, go the Create a new Endpoint and select a model and hosting setup.

Once the endpoint is created, take the Endpoint URL shown on the page:

Then set up your promptfoo config like this:

description: 'HF private inference endpoint'

prompts:
  - 'Write a tweet about {{topic}}:'

providers:
  - id: huggingface:text-generation:gemma-7b-it
    config:
      apiEndpoint: https://v9igsezez4ei3cq4.us-east-1.aws.endpoints.huggingface.cloud
      # apiKey: abc123   # Or set HF_API_TOKEN environment variable

tests:
  - vars:
      topic: bananas
  - vars:
      topic: potatoes

Local inference

If you're running the Huggingface Text Generation Inference server locally, override the apiEndpoint:

providers:
  - id: huggingface:text-generation:my-local-model
    config:
      apiEndpoint: http://127.0.0.1:8080/generate

Datasets

promptfoo can load test cases directly from HuggingFace datasets. To use a dataset, specify it in your config using the huggingface://datasets/ prefix:

tests: huggingface://datasets/fka/awesome-chatgpt-prompts

You can customize the dataset split and other parameters using query parameters:

# Load from training split
tests: huggingface://datasets/fka/awesome-chatgpt-prompts?split=train

# Load from validation split with custom config
tests: huggingface://datasets/fka/awesome-chatgpt-prompts?split=validation&config=custom

The dataset rows will be automatically converted to test cases. Each field in the dataset becomes a variable that can be used in your prompts. For example, if your dataset has fields question and answer, you can reference them in your prompts like this:

prompts:
  - "Question: {{question}}\nAnswer:"

tests: huggingface://datasets/rajpurkar/squad

Query Parameters

The dataset loader supports all query parameters from the HuggingFace Datasets API. Common parameters include:

Parameter	Description	Default
split	Dataset split to load (train/test/validation)	test
config	Dataset configuration name	default

Any additional query parameters will be passed directly to the HuggingFace Datasets API.

Example Configuration

Here's a complete example that loads test cases from a HuggingFace dataset:

description: Testing with HuggingFace dataset

prompts:
  - 'Act as {{act}}. {{prompt}}'

providers:
  - id: openai:gpt-4o-mini

tests: huggingface://datasets/fka/awesome-chatgpt-prompts?split=train

This will load all rows from the dataset and use the act and prompt fields as variables in the prompt template.