gpetters94
gpetters94
I have a 32-bit Wine prefix that used to work great with this, but for the past week or so when I try to launch Steam with this LD_PRELOADed, I...
As @silvasean requested in [this issue](https://github.com/llvm/torch-mlir/pull/917#discussion_r896154545) here is the shape code from Torch-MLIR for conv_transpose2d.input and convolution (updated for the transposed case).
It'd be nice to not have to specify the download location for every file.
Here's transposed support for convolution, both strided and unstrided.
I'm working on lowering OPT, and I'm running into the following: ``` error: failed to legalize operation 'torch.aten.view' that was explicitly marked illegal note: see current operation: %932 = "torch.aten.view"(%898,...
I'm unable to generate heavydep tests at TOM using the script in `build_tools` with the following output: ``` Traceback (most recent call last): File "/home/george/.pyenv/versions/3.8-dev-debug/lib/python3.8/runpy.py", line 194, in _run_module_as_main return...
I'm running into an issue in U-Net, which I was using to test #917 and which I originally attributed to an issue in my code. However, looking at it more...
Do we want to have `prim.is_cuda` always return true? Do we have a concept of on the GPU or off the GPU in torch-mlir?
Just wanted to get some feedback on implementing a linalg op in upstream that implements similar logic to #858, i.e. the same logic as the normal `conv_2d_nchw_fchw` op but dividing...
I'm still working on the nonzero padding case, and it's not very efficient, but here's the implementation for convolution_backward.