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