HuggingFace Requests

A free browser-based approach to experiment with text-to-text models hosted on the HuggingFace Inference API

MichaelCurrin - flan-t5-tests

Inputs

Result

Instructions

To use this form, you need a HuggingFace account and API token. You can use the free tier (within limits) and do not need to enter billing details.

Follow the instructions below and see the HuggingFace tokens documentation for more help if needed.

  1. Go to the HuggingFace website: huggingface.co.
  2. Sign in
  3. Click the Profile icon.
  4. Select Settings.
  5. Click on Access Tokens.
  6. Create a token - "readonly" access is sufficient
  7. Copy and paste the token into the form.
  8. Enter a model and prompt and submit

🔐 Your form inputs are persisted in localstorage for convenience. But your secret token is still secure as it is not stored anywhere except in your own browser.

Notes

ℹ️ The FLAN-5 model as used in the default setup is one of the popular recent "transformer" models that supports text to text tasks. It can handle tasks like summarization and language translation but may not be suited to a certain tasks and is not a chatbots or LLM.

⚠️ The Inference API takes a few seconds to warm up if it is not already active. So if you get an error around that, wait and try again.

ℹ️ As an alternative model to the default, try the "-large" model or other text to text generation models on the Models section of HuggingFace. If you try another kind of model, the code on this page will fail to run it.