A torch.fx Based Compression Toolkit Empowered by torch_musa - Fan Mo, Moore Threads
A torch.fx Based Compression Toolkit Empowered by torch_musa - Fan Mo, Moore Threads
An introduction of torch_musa, which is built by Moore Threads to provide PyTorch backend support for MUSA architecture, and with its customized support for quantization, we built a light-weight model compression toolkit named NeuroTrim, which can quantize a variety of deep learning models and run it on native PyTorch.
We expanded the graph tracing capability of PyTorch by integrating fx graph, torchInductor, and ExportedIR, and added many lowered operators in torch_musa, which can be easily used for debugging, and then export to ONNX or other formats for custom accelerator hardwares.
An introduction of torch_musa, which is built by Moore Threads to provide PyTorch backend support for MUSA architecture, and with its customized support for quantization, we built a light-weight model compression toolkit named NeuroTrim, which can quantize a variety of deep learning models and run it on native PyTorch.
We expanded the graph tracing capability of PyTorch by integrating fx graph, torchInductor, and ExportedIR, and added many lowered operators in torch_musa, which can be easily used for debugging, and then export to ONNX or other formats for custom accelerator hardwares.
PyTorch
Welcome to the official PyTorch YouTube Channel. Learn about the latest PyTorch tutorials, new, and more.
PyTorch is an open source machine learning framework that is used by both researchers and developers to build, train, and deploy ML systems that so...
torch.compile and Diffusers: A Hands-On Guide to Peak Performance - Sayak Paul, Hugging Face
PyTorch
torch compile and Diffusers - A Hands On Guide to Peak Performance - PyTorch Compiler Series
PyTorch
verl: An Open Source Large Scale LLM RL Framework for Agentic Tasks - Yuxuan Tong, Bytedance
PyTorch
Lightning Talk: Sparsifying Vision Transformers with Minimal Accuracy Loss - Jesse Cai, Meta
PyTorch