A curated list of awesome libraries, models, datasets, tutorials, and tools from and for the Hugging Face ecosystem.
Hugging Face is a leading company in open-source AI, offering powerful NLP, computer vision, speech, and multimodal tools with a vibrant ecosystem of models, datasets, and transformers-based libraries.
- Official Resources
- Popular Libraries
- Models
- Datasets
- Tutorials & Courses
- Deployment & Hosting
- Integrations & Tools
- Community Projects
- Related Awesome Lists
- π€ Hugging Face Hub β The central repository for models, datasets, and spaces.
- π€ Transformers β State-of-the-art pre-trained models for NLP, vision, speech, and more.
- π€ Datasets β Easily share and load datasets for machine learning.
- π€ Tokenizers β Fast, customizable tokenization library.
- π€ Accelerate β Train models across multiple GPUs or TPUs with minimal code changes.
- π€ PEFT β Parameter-efficient fine-tuning methods like LoRA.
- π€ Diffusers β State-of-the-art diffusion models for generative tasks.
- π€ Safetensors β A safe and fast format for model weights.
- transformers β Core library for pretrained models.
- datasets β Community-driven dataset collection.
- evaluate β Standardized metrics and evaluation tools.
- gradio β Build web-based demos easily (integrates with HF Spaces).
- optimum β Performance optimization tools for Hugging Face models.
- inference endpoints β Managed model hosting service.
- text-generation-inference β Efficient inference for LLMs.
- BERT β Bidirectional encoder for language understanding.
- GPT-2 β Open-source autoregressive language model.
- T5 β Text-to-text transformer.
- DistilBERT β Smaller, faster BERT.
- Whisper β Speech recognition from OpenAI.
- CLIP β Vision-language model.
- Bloom β Open multilingual LLM.
- Falcon β Open-weight high-performance LLM.
- Hugging Face Course β Official in-depth free course.
- Fine-tuning Transformers
- Using Accelerate for Multi-GPU
- Hugging Face Transformers on Google Colab
- Hugging Face Spaces β Free app hosting platform for ML demos (uses Gradio or Streamlit).
- Inference Endpoints β Production-ready model APIs.
- Docker templates β Containers for deploying HF models.
- Weights & Biases β Integrated experiment tracking.
- Comet ML β Model training dashboards.
- Keras + Hugging Face β Use transformers in TensorFlow workflows.
- Ray Tune β Hyperparameter tuning.
- HuggingChat β Chat UI for open-source models.
- Text2Image Gradio Demo β DALL-E-style generator.
- CodeParrot β Code generation datasets/models.
- Open Assistant β Community-built assistant LLM.
- AutoTrain β Low-code fine-tuning for non-coders.
Contributions are welcome!