Huggingface

The huggingface module is a API module that features a core module can talk to the inference API of Huggingface. It will hopefully in the future have a suite of submodules or modules dependent on it.

Features

• Use one of the following rules on any model, including private, to generate/classify content with the AI Interpolator module: text-to-image, fill-mask, summarization, question-answering, table-question-answering, sentence-similarity, text-classification, text-generation, token-classification, translation, zero-shot-classification, conversational, automatic-speech-recognition, audio-classification, image-classification, object-detection, image-segmentation.
• Use the Image classification models together with the Augmentor AI Tags Widget.
• Finetune Stable Diffusion modules completely inside Drupal using Dreambooth on Runpod servers and host them in your Huggingface account.

Runpod integration

I'm working on integrating Runpod on multiple models finetuning, so you can automise finetuning, but also do finetuning without needing to learn Python, server setups and more.

If you end up using the model, I would urge you to sign up with the following link: https://runpod.io?ref=i7zrhoj7. This will give me credits to do more cases of finetuning in the future.

Post-Installation

1. Get a Huggingface token.
2. Visit /admin/config/huggingface/settings and paste it in there and save.
3. Create one of the field type of the rules, enable AI Interpolator and make sure you give the model namespace or a dedicated host.

Maintainers

Anyone looking to generate a Huggingface integration and want to co-maintain, please let me know.