Make VBoost activation conditional by bdubauski · Pull Request #14458

Make VBoost activation conditional by bdubauski · Pull Request #14458 · NVIDIA-NeMo/NeMo

Signed-off-by: Guyue Huang <guyueh@nvidia.com>
Signed-off-by: Barys Dubauski <bdubauski@nvdia.com>

* Add notebooks for LoRA on GPT-OSS-20B

Signed-off-by: Shashank Verma <shashank3959@gmail.com>

* Update link to nb in README

Signed-off-by: Shashank Verma <shashank3959@gmail.com>

* Minor tweaks to configs to improve recipe

Signed-off-by: Shashank Verma <shashank3959@gmail.com>

* Move folders

Signed-off-by: Shashank Verma <shashank3959@gmail.com>

* Fix broken link

Signed-off-by: Shashank Verma <shashank3959@gmail.com>

* Minor adjustments to documentation in the notebook

Signed-off-by: Shashank Verma <shashank3959@gmail.com>

* Add link to NeMo docs in README

Signed-off-by: Shashank Verma <shashank3959@gmail.com>

* Remedy review feedback around docs

Signed-off-by: Shashank Verma <shashank3959@gmail.com>

* change gpt-oss to official branding

Signed-off-by: Shashank Verma <shashank3959@gmail.com>

---------

Signed-off-by: Shashank Verma <shashank3959@gmail.com>
Signed-off-by: Barys Dubauski <bdubauski@nvdia.com>

* Refactor performance scripts to use build_perf_env_plugin function

* Replaced direct instantiation of PerfEnvPlugin with build_perf_env_plugin in multiple LLM finetuning and pretraining scripts for consistency and maintainability.
* Added build_perf_env_plugin function to helpers.py to streamline performance environment setup based on GPU and pipeline parallelism settings.

This change enhances code readability and reduces redundancy across scripts.

* control vboost enablement via cli

* Update finetune_llama4_e128.py to import build_perf_env_plugin function

* Added the build_perf_env_plugin import to enhance performance environment setup consistency across scripts.

This change aligns with recent refactoring efforts to streamline performance script management.

---------

Co-authored-by: Barys Dubauski <bdubauski@nvdia.com>
Signed-off-by: Barys Dubauski <bdubauski@nvdia.com>

Signed-off-by: Barys Dubauski <bdubauski@nvdia.com>

Signed-off-by: Barys Dubauski <bdubauski@nvdia.com>

Signed-off-by: Barys Dubauski <bdubauski@nvdia.com>

Updated multiple LLM finetuning and pretraining scripts to replace direct instantiation of PerfEnvPlugin with the new build_perf_env_plugin function. This change enhances code consistency and maintainability across the scripts by streamlining the performance environment setup based on GPU and pipeline parallelism settings.

Signed-off-by: Barys Dubauski <bdubauski@nvdia.com>

Signed-off-by: bdubauski <bdubauski@users.noreply.github.com>
Signed-off-by: Barys Dubauski <bdubauski@nvdia.com>

Signed-off-by: bdubauski <bdubauski@users.noreply.github.com>
Signed-off-by: Barys Dubauski <bdubauski@nvdia.com>

Updated multiple LLM and VLM finetuning and pretraining scripts to eliminate unnecessary imports of PerfEnvPlugin, enhancing code clarity and maintainability. This change streamlines the import statements across the scripts.

Signed-off-by: Barys Dubauski <bdubauski@nvdia.com>

* fix_subsampling_chunking_test

Signed-off-by: monica-sekoyan <msekoyan@nvidia.com>

* Apply isort and black reformatting

Signed-off-by: monica-sekoyan <monica-sekoyan@users.noreply.github.com>

* increasing threshold

Signed-off-by: monica-sekoyan <msekoyan@nvidia.com>

---------

Signed-off-by: monica-sekoyan <msekoyan@nvidia.com>
Signed-off-by: monica-sekoyan <monica-sekoyan@users.noreply.github.com>
Co-authored-by: monica-sekoyan <monica-sekoyan@users.noreply.github.com>
Signed-off-by: Barys Dubauski <bdubauski@nvdia.com>

guyueh1 pushed a commit to guyueh1/NeMo that referenced this pull request

Aug 25, 2025

Signed-off-by: Guyue Huang <guyueh@nvidia.com>