HuggingFace Models

Compile and accelerate HuggingFace models with Torch-TensorRT: large language models and visual language models via the tools/llm toolkit, Stable Diffusion via torch.compile, Flux via torch.export, and LoRA weight-swapping via the Mutable Torch-TensorRT Module.

• Compiling LLM models from Huggingface
• Example: Compiling Stable Diffusion with torch.compile
• Example: Compiling FLUX.1-dev with the dynamo backend
• Example: FLUX.1-dev INT4 WOQ
• Example: FLUX.1-dev NVFP4 WOQ
• Example: FLUX.1-dev MXFP4
• Example: Qwen3-8B INT4 WOQ
• Example: Mutable Torch TensorRT Module