Kirthi Shankar Sivamani
Kirthi Shankar Sivamani
# What does this PR do ? Adds a brief description of FP8 usage with NeMo and describes the relevant arguments. **PR Type**: - [ ] New Feature - [...
# What does this PR do ? Fixes the incorrectly rendered table in FP8 args documention. **Collection**: [Note which collection this PR will affect] # Changelog - Add specific line...
### Project URL https://pypi.org/project/transformer-engine-cu12/ ### Does this project already exist? - [X] Yes ### New Limit 500 ### Update issue title - [X] I have updated the title. ### Which...
# Description `setup_requires` is deprecated ([pep ref](https://peps.python.org/pep-0517/)) and will be removed in `setuptools` v75.0.0. This PR uses the newer `pyproject.toml` to specify build time dependencies. We can consider gradually moving...
# Description Various improvements to build system and wheels, detailed in the _**changes**_ section. ## Type of change - [ ] Documentation change (change only to the documentation, either a...
# Description Update the recommended docker container in README from 25.01 (v1.14) to 25.03 (v2.1) ## Type of change - [x] Documentation change (change only to the documentation, either a...
# Description Please include a brief summary of the changes, relevant motivation and context. Fixes # (issue) ## Type of change - [ ] Documentation change (change only to the...
# Description Based on single GPU profiling of the `GroupedLinear` module, implement some optimizations in order to reduce CPU overhead due to PyTorch. ## Type of change - [ ]...
# Description This is a small refactor of library loading logic during runtime to be more consistent and avoid duplication. The main point is to check python packages as a...