Change8

v4.51.3-MLCD-preview

📦 transformersView on GitHub →
3 features🔧 3 symbols

Summary

This release introduces a preview of the MLCD vision model, a foundational visual model optimized for multimodal LLMs like LLaVA, developed by DeepGlint-AI.

Migration Steps

  1. Install the preview version using: pip install git+https://github.com/huggingface/transformers@v4.51.3-MLCD-preview

✨ New Features

  • Added support for the MLCD (Multi-modal Large-scale Contrastive Distillation) model architecture.
  • Integration of MLCDVisionModel for visual feature extraction in multimodal pipelines.
  • Support for MLCD-specific processing via AutoProcessor.

Affected Symbols