TensorRT
TensorRT copied to clipboard
pytorch
PyTorch/TorchScript/FX compiler for NVIDIA GPUs using TensorRT
## Bug Description When using torch_tensorrt.Input and torch_tensorrt.compile, I get an error: ` File "/mnt/c/Coding/Testing/PyTorch/MultiClassImageClassification/src/compressmodel.py", line 48, in trt_gm = torch_tensorrt.compile(model, ir="dynamo", inputs=[input], enabled_precisions = {torch.half, torch.float}, output_format="torchscript") ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File...
https://github.com/pytorch/TensorRT/blob/b21fb5fe5cce4145006efbea29b37f4319850ade/py/torch_tensorrt/dynamo/conversion/_TRTInterpreter.py#L492-L505
The `torch.ops.aten._local_scalar_dense` operation is a low-level operation in PyTorch used internally to extract a scalar value from a tensor containing a single element.
## Bug Description Implement dynamic batch and dynamic shapes support for layer norm converter. Add the following testcase once it is implemented ```py def test_layernorm_with_dynamic_shape(self): class LayerNorm(torch.nn.Module): def forward(self, x):...
## Bug Description The code below produces the following error: ``` RuntimeError: The following operation failed in the TorchScript interpreter. Traceback of TorchScript (most recent call last): RuntimeError: [Error thrown...
- See [here](https://github.com/pytorch/pytorch/blob/4e3022dbe94af5db41a151fc7c2cbf717c583d0f/aten/src/ATen/native/native_functions.yaml#L9164-L9170) for operator schema - See [here](https://pytorch.org/docs/stable/generated/torch.masked_select.html) for `torch.masked_select` documentation
# Description Please include a summary of the change and which issue is fixed. Please also include relevant motivation and context. List any dependencies that are required for this change....
The issue in this PR is for cases where there is index collision. Example of such cases- `scatter_add(input_tensor = torch.zeros(3,5), dim=1, index_tensor = torch.tensor([[0, 1, 2, 0]]).cuda(), src_tensor = torch.tensor([[1,...