pytorch

pytorch copied to clipboard

Degreded preformance after first epoch

1

### 🐛 Describe the bug running imagenet main.py in pytorch examples [ github link ](https://github.com/pytorch/examples/tree/main/imagenet) ``` => creating model 'mobilenet_v3_small' Epoch: [0][ 1/1424] Time 60.000 (60.000) Data 15.133 (15.133) Loss...

YumingChang02

### 🐛 Describe the bug I installed the pytorch 2.1 with rocm 5.4.2. When I tested the one_hot function, it returned this: print(torch.nn.functional.one_hot(torch.tensor(0,device=‘cuda:0’), 2)) = [0,0] Do you know why...

akichinguyen

No CPU fallback?

2

## Issue description In our software, we install the ROCm enabled pytorch if we detect an AMD GPU on Linux. Otherwise, we install the CUDA-enabled pytorch. If the AMD GPU...

JeLuF

This PR uses hipify_torch as a submodule instead of torch.utils.hipify This PR is dependent on https://github.com/ROCmSoftwarePlatform/hipify_torch/pull/50

lcskrishna

### 🐛 Describe the bug rocm (rocrand or rocblas or something like that) defaultly use AVX instructions. When they are loaded by torch with rocm support, just import torch will...

echoidcf

[CI test do not merge] Skip rocm msg

4

Testing ci

jataylo

Add CUDAMallocManagedAllocator Backend With the new CUDAAllocator class, we have created a new CUDAMallocManagedAllocator, which will handle allocator requests from both cpu and cuda device types when the backend is...

dllehr-amd

Sow ms7 enable test reference numerics small all except complex

7

This PR is to enable tests which are skip on rocm from test package test_unary_ufuncs.py::TestUnaryUfuncsCUDA test_file | test_name | test_class -- | -- | -- test_unary_ufuncs | test_reference_numerics_large_polygamma_polygamma_n_2_cuda_float16 | (__main__.TestUnaryUfuncsCUDA)...

bmedishe

Enable test_det_logdet_slogdet on ROCm

5

BLOrange-AMD

Hey @jithunnair-amd This PR is the change to fix the ldl_factor tests regarding that "hermitian" flag. I know we wanted to wait until hipsolver was enabled by default (hence the...

alugorey