Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Build manylinux wheels #3

Merged
merged 11 commits into from
Jul 26, 2023
Merged

Build manylinux wheels #3

merged 11 commits into from
Jul 26, 2023

Conversation

daskol
Copy link
Member

@daskol daskol commented Feb 23, 2023

Sandbox Docker image and scripts for building wheels are provided.

daskol added 10 commits June 17, 2022 02:02
- Do not use init-stmt in if-clause for NVCC 10.
- Do not use obsolete GELU (approximated) implementation.
- Fix NVCC flags usage in Torch 1.10.
- Add automatic inference for CMake prefix.
- Change default value for CUDA architectures for code generation.
- Compose fewbit version from pattern below.

      <torch-public>.<fewbit-public>[+<torch-local>]

- Make hard dependency on PyTorch version.
- Also, add stub for package version.
- Add fallback for activations functions
- Use bits argument for activations layers
- Fix test on forward of randomized linear layers.
- Add argument validattion to forward of randomized layers.
- Fix integration of approximation error of quantized gradients of
  activation functions.
@daskol daskol mentioned this pull request Feb 23, 2023
@daskol daskol force-pushed the wheels branch 2 times, most recently from eda89bc to 4055007 Compare July 26, 2023 04:35
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant