transformers-docs

What problem does it solve?

This Skill helps you quickly find the right guidance for using and optimizing the Hugging Face Transformers library, covering model loading, training, inference, tokenization, quantization, and serving integrations.

Core Features & Use Cases

• Transformers usage reference for core concepts like AutoModel/AutoTokenizer, from_pretrained, the pipeline() API, generate(), and the Trainer workflow.
• Inference optimization & serving guidance including attention backends, KV-cache strategies, and integration pointers for vLLM/SGLang/llama.cpp.
• Training & distributed setup support spanning FSDP, DeepSpeed, Accelerate, and Trainer-related configuration topics.
• Quantization roadmap for GPTQ, AWQ, bitsandbytes, and GGUF, with links to the most relevant subsections.

Quick Start

Ask it for guidance on how to load a pretrained causal language model and choose the correct decoding strategy for generate() while also understanding how quantization affects that flow.