Transformers
Provides state-of-the-art machine learning models for text, computer vision, audio, video, and multimodal tasks, supporting inference and training across PyTorch, TensorFlow, and JAX.
关于
Transformers is a comprehensive library offering pre-trained models for various modalities, including text, computer vision, audio, video, and multimodal applications. It facilitates fine-tuning models on user data, constructing inference applications, and generative AI across multiple modalities. With access to over 500K+ model checkpoints on the Hugging Face Hub, it allows developers to get started immediately with a wide array of pre-trained models and frameworks.
主要功能
- Unified API for text, audio, vision, and multimodal tasks
- Access to over 500K+ pre-trained models on the Hugging Face Hub
- 144,497 GitHub stars
- Pipeline API for quick start inference
- Supports PyTorch, TensorFlow, and JAX frameworks
- Tools for fine-tuning models
使用案例
- Natural Language Understanding and Generation
- Image Classification and Object Detection
- Speech Recognition and Audio Processing
Sponsored