[MPS] .backward() encounters old nn.Embedding() weights #101198
Labels
module: mps
Related to Apple Metal Performance Shaders framework
triaged
This issue has been looked at a team member, and triaged and prioritized into an appropriate module
馃悰 Describe the bug
Code to reproduce:
The error (some directory information omitted):
It works fine when
device = torch.device('cpu')
.Note the 2nd line of the error, where it seems the underlying MPS operation was working with a tensor of size <100x36> (i.e the dimensions of
emb_36_a
andemb_36_b
's underlying weights), but I'd expect it to be <100x48> (the dims ofemb_48
's weights).Versions
PyTorch version: 2.0.1
Is debug build: False
CUDA used to build PyTorch: None
ROCM used to build PyTorch: N/A
OS: macOS 12.6 (arm64)
GCC version: Could not collect
Clang version: 14.0.0 (clang-1400.0.29.202)
CMake version: version 3.22.2
Libc version: N/A
Python version: 3.8.11 (default, Feb 22 2022, 14:24:07) [Clang 13.0.0 (clang-1300.0.27.3)] (64-bit runtime)
Python platform: macOS-12.6-arm64-arm-64bit
Is CUDA available: False
CUDA runtime version: No CUDA
CUDA_MODULE_LOADING set to: N/A
GPU models and configuration: No CUDA
Nvidia driver version: No CUDA
cuDNN version: No CUDA
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: True
CPU:
Apple M1 Pro
Versions of relevant libraries:
[pip3] numpy==1.24.3
[pip3] torch==2.0.1
[conda] No relevant packages
cc @kulinseth @albanD @malfet @DenisVieriu97 @razarmehr @abhudev
The text was updated successfully, but these errors were encountered: