This repository contains a simple example of how to run inference on HTDemucs model using ONNX runtime. The model is first converted to ONNX format using the new dynamo torch.export (as described here) and then loaded into ONNX runtime for inference. htdemucs_optimized.onnx comes from this script

Run npm run dev to start the vite development server. Run the model, check errors in the console.