ExecuTorch Brings PyTorch Models to Browser GPUs with WebGPU
ExecuTorch’s new WebGPU backend lets PyTorch models run on browser GPUs without converting them to ONNX, GGUF, or TFLite. The same .pte artifact can support local inference across WebGPU-enabled browsers and a broad set of text, vision, and speech workloads.
Read more