HuggingFace

Description: The HuggingFace class serves as a wrapper for interacting with the Hugging Face APIs for operations such as model inference, repository creation, file management, and validity checks for API tokens and Hugging Face hubs.

---

Method: tokenIsValid

Returns: A boolean indicating whether the provided API token is valid.

Description:

• This method checks the validity of the Hugging Face API token by attempting to list the available models.
• If the token is valid, the method returns true, otherwise it returns false.
• Catches and logs any errors during the validation process.

Example: Verify the Hugging Face API token is valid before proceeding with further API interactions.

const hf = new HuggingFace('your-token');
const isValid = await hf.tokenIsValid();
if (isValid) {
  console.log('Token is valid.');
} else {
  console.log('Invalid token, please check your credentials.');
}

Parameters: None.

---

Method: hubExists

Returns: A boolean indicating whether the target Hugging Face hub exists.

Description:

• This method verifies the existence of a Hugging Face hub by listing models.
• If the hub exists, it returns true, otherwise it returns false.
• Similar in functionality to tokenIsValid() and may be subject to change to better reflect its intended purpose.

Example: Check if the Hugging Face hub exists.

const hf = new HuggingFace('your-token');
const exists = await hf.hubExists();
if (exists) {
  console.log('Hub exists.');
} else {
  console.log('Hub does not exist.');
}

Parameters: None.

---

Method: questionAnswering

Returns: The method returns the result of a question answering model hosted on the Hugging Face platform.

Description:

• Calls upon a pre-trained question answering model, specified by the user, to process the provided inputs.

Example: Use a question answering model to obtain answers based on provided input.

const hf = new HuggingFace('your-token');
const model = "bert-base-uncased";
const inputs = {
  question: "What is Hugging Face?",
  context: "Hugging Face is a social AI company and platform for ML models."
};
const result = await hf.questionAnswering(model, inputs);
console.log(result);

Parameters:

Parameter Name	Description	Accepted Values/Data Types
model	The identifier for the Hugging Face question answering model.	String
inputs	The inputs containing the question and context for the model.	Object { question, context }

---

Method: createRepo

Returns: The response from the Hugging Face API upon the creation of a new repository.

Description:

• Creates a new repository or a folder in the Hugging Face hub to store models or datasets.
• Accepts a repository path and a type to specifically create a 'model' type repository if specified.

Example: Create a new model repository at the specified path.

const hf = new HuggingFace('your-token');
const repoPath = "your-username/your-new-model";
const repoType = "model";
const response = await hf.createRepo(repoPath, repoType);
console.log(response);

Parameters:

Parameter Name	Description	Accepted Values/Data Types
repoPath	The path where the new repository will reside on the hub.	String
type	The type of repository to create (optionally specify 'model').	String

---

Method: uploadFile

Returns: The response from the Hugging Face API upon successful file upload.

Description:

• Uploads a file (e.g., a model file) to a specific repository on the Hugging Face hub.

Example: Upload a model file to a specified repository.

const hf = new HuggingFace('your-token');
const repoPath = "your-username/your-model";
const filePath = "pytorch_model.bin";
const blob = new Blob([...]); // blob containing file data
const response = await hf.uploadFile(repoPath, filePath, blob);
console.log(response);

Parameters:

Parameter Name	Description	Accepted Values/Data Types
repoPath	The repository path to upload the file to.	String
filePath	The destination file path on the hub.	String
blob	The file content to be uploaded.	Blob

---

Method: deleteFiles

Returns: The response from the Hugging Face API after attempting to delete the specified files.

Description:

• Deletes one or more files within a repository or space on the Hugging Face hub.

Example: Delete specific files in a model repository.

const hf = new HuggingFace('your-token');
const type = "model";
const name = "your-username/your-model";
const paths = ["file1.bin", "file2.bin"];
const response = await hf.deleteFiles(type, name, paths);
console.log(response);

Parameters:

Parameter Name	Description	Accepted Values/Data Types
type	The type of the target to delete files from.	String
name	The path to the repo or space where the files are located.	String
paths	An array of strings representing the file paths to delete.	Array of String

---