Stage 2: Optimizing Engine Conversion - [ ] Research different engine export methods and benchmark results/create performance report - [ ] ex. PyTorch -> TensorRT (Torch-TRT or torchscript) - [ ] ex. PyTorch -> ONNX -> TensorRT - [ ] Research different TensorRT runtimes - [ ] Research different TensorRT Engine Plugins Helpful Documentation Sites: - [ ] [ONNX to TRT Project](https://github.com/onnx/onnx-tensorrt/tree/main) - [ ] [Nvidia Library Converter](https://github.com/NVIDIA-AI-IOT/torch2trt?tab=readme-ov-file#models) - [ ] [Basic Explanation Article](https://pytorch.org/TensorRT/user_guide/torch_tensorrt_explained.html) - [ ] [Deploying Programs](https://pytorch.org/TensorRT/user_guide/runtime.html) - [ ] [Quick Start](https://pytorch.org/TensorRT/getting_started/quick_start.html) - [ ] [Dynamo Export](https://pytorch.org/tutorials/advanced/super_resolution_with_onnxruntime.html) - [ ] [Engine Caching](https://pytorch.org/TensorRT/tutorials/_rendered_examples/dynamo/engine_caching_example.html) - [ ] [Compiling Exported Programs](https://pytorch.org/TensorRT/dynamo/dynamo_export.html) - [ ] [Exporting CUDA Graphs](https://pytorch.org/TensorRT/tutorials/_rendered_examples/dynamo/torch_export_cudagraphs.html) - [ ] [Polygraphy](https://github.com/NVIDIA/TensorRT/tree/main/tools/Polygraphy/how-to)
Stage 2: Optimizing Engine Conversion
Helpful Documentation Sites: