Pytorch release 2.0 - rc1 is available

PyTorch core and Domain Libraries are available for download from pytorch-test channel.

Reminder of key dates:

  • Domain libraries to cut RC Branch 2/17 (completed)
  • RC1 Binaries for PyTorch core and Domain Libraries are available 2/20 (completed)
  • Last Cherry Pick 2/27
  • Release date 3/15

Following are instructions on how to download different versions of RC for testing.

CONDA CPU:

Windows/LInux:
conda install pytorch torchvision torchaudio cpuonly -c pytorch-test

MacOS:
conda install pytorch torchvision torchaudio -c pytorch-test

CUDA 11.7, 11.8

Windows/Linux:
conda install pytorch torchvision torchaudio pytorch-cuda=11.7 -c pytorch-test -c nvidia
conda install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch-test -c nvidia

PIP CPU:
Windows/Linux/MacOS:
pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/test/cpu

PIP CUDA 11.7, 11.8
Windows/Linux:
pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/test/cu117
pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/test/cu118

PIP ROCM 5.4
pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/test/rocm5.4.2/

Libtorch:

CPU Linux:

(Pre-cxx11 ABI):

https://download.pytorch.org/libtorch/test/cpu/libtorch-shared-with-deps-latest.zip

(cxx11 ABI):

https://download.pytorch.org/libtorch/test/cpu/libtorch-cxx11-abi-shared-with-deps-latest.zip

CPU Windows:

Download here (Release version):

https://download.pytorch.org/libtorch/test/cpu/libtorch-win-shared-with-deps-latest.zip

Download here (Debug version):

https://download.pytorch.org/libtorch/test/cpu/libtorch-win-shared-with-deps-debug-latest.zip

CPU MacOS:

https://download.pytorch.org/libtorch/test/cpu/libtorch-macos-latest.zip

CUDA 11.7, 11.8 Linux

(Pre-cxx11 ABI):

https://download.pytorch.org/libtorch/test/cu117/libtorch-shared-with-deps-latest.zip

https://download.pytorch.org/libtorch/test/cu118/libtorch-shared-with-deps-latest.zip

(cxx11 ABI):

https://download.pytorch.org/libtorch/test/cu117/libtorch-cxx11-abi-shared-with-deps-latest.zip

https://download.pytorch.org/libtorch/test/cu118/libtorch-cxx11-abi-shared-with-deps-latest.zip

Windows CUDA 11.7, 11.8

(Release version):

https://download.pytorch.org/libtorch/test/cu117/libtorch-win-shared-with-deps-latest.zip

https://download.pytorch.org/libtorch/test/cu118/libtorch-win-shared-with-deps-latest.zip

(Debug version):

https://download.pytorch.org/libtorch/test/cu117/libtorch-win-shared-with-deps-debug-latest.zip

https://download.pytorch.org/libtorch/test/cu118/libtorch-win-shared-with-deps-debug-latest.zip

If you have any questions please comment below or reach out to us on this thread.

1 Like

I believe you may have a mistake in the index url for pip -

This one fails:
https://download.pytorch.org/whl/test/cu117/torch_test.html

And it seems this is the correct one:
https://download.pytorch.org/whl/test/cu117

Indeed you are correct PIP install instructions should be:

PIP CPU:
Windows/Linux/MacOS:
pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/test/cpu

PIP CUDA 11.7, 11.8

Windows/Linux:
pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/test/cu117
pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/test/cu118

Will a pytorch 2.0 wheel with cuda 12 available at the time of the release?

I’m hitting some issues when trying to use the pytorch-nightly conda dev binaries for TorchVision and TorchAudio (see the pytorch/builder GitHub issue here). Is there any documentation, or does anyone have advice, about how to get a custom conda build of PyTorch 2.0.0 to work with the pytorch-nightly conda channel binaries for TorchAudio and TorchVision? I’m hitting dependency issues when trying to make a custom conda channel with these binaries as described in the linked issue. Any help would be greatly appreciated!

Please refer to Release Compatibility Matrix: pytorch/RELEASE.md at master · pytorch/pytorch · GitHub

We will be releasing PyTorch 2.0 with CUDA 11.7 and CUDA 11.8.

@dkey-amazon : Replied you on the issue you mentioned