Cheng Hang

Support shape analysis for dynamic fallback

1

Signed-off-by: Cheng Hang # Description Previously, when encountered ops/modules that have to fall back and meanwhile dynamic shape is used, shape analysis would fail when doing `generateRandomInputs()`, because tensors with...

↔ [Converter] Add support for aten::atan2 in Torch-TensorRT

2

## Unsupported operator: aten::atan2(Tensor self, Tensor other) -> (Tensor) - **Function Schema**: aten::atan2(Tensor self, Tensor other) -> (Tensor) - **Original PyTorch API**: torch.atan2(input, other, *, out=None) → [Tensor] - **Relevant...

🐛 [Bug] Shape analysis failure when encountered dynamic fallback

5

## Bug Description In dynamic fallback scenarios, the shape analysis will fail in `generateRandomInputs` phase. ```bash Traceback (most recent call last): File "main.py", line 56, in main() File "main.py", line...

Add trt_executed_modules & default_torch_execution interface

2

Signed-off-by: Cheng Hang # Description According to the current design, every module is by default compiled to TensorRT execpt those modules included in torch_executed_modules. However, in some cases, for example,...

A FasterRMSNorm implementation (similar to FasterLayerNorm).