Merged
Conversation
Signed-off-by: Alexandros Koumparoulis <akoumparouli@nvidia.com>
Signed-off-by: Alexandros Koumparoulis <akoumparouli@nvidia.com>
Signed-off-by: Alexandros Koumparoulis <akoumparouli@nvidia.com>
Signed-off-by: Alexandros Koumparoulis <akoumparouli@nvidia.com>
Signed-off-by: Alexandros Koumparoulis <akoumparouli@nvidia.com>
Signed-off-by: Alexandros Koumparoulis <akoumparouli@nvidia.com>
Signed-off-by: Alexandros Koumparoulis <akoumparouli@nvidia.com>
Signed-off-by: Alexandros Koumparoulis <akoumparouli@nvidia.com>
oot cause: nemo-run pins cryptography<43.0.0, but your project constrains cryptography>=46.0.5 (CVE fix). These are fundamentally incompatible, so uv lock can't resolve any extra that depends on nemo_run. Fix: Removed nemo_run from the cli extra -- it's now just ["pyyaml"] Removed the standalone nemo-run extra entirely (it was also unresolvable) Updated BREAKING_CHANGES.md, installation guide, and cluster guide to note that nemo-run should be installed separately (pip install nemo-run) if needed The SLURM and k8s launchers don't need nemo-run at all (they shell out to sbatch/kubectl), and the NemoRunLauncher already does a runtime import check, so users who need it just install nemo-run directly.
d5cd04a to
944138c
Compare
Contributor
Author
|
/ok to test 203a148 |
Signed-off-by: Alexandros Koumparoulis <akoumparouli@nvidia.com>
203a148 to
ffdadc3
Compare
Signed-off-by: akoumpa <akoumpa@users.noreply.github.com>
Signed-off-by: Alexandros Koumparoulis <akoumparouli@nvidia.com>
Contributor
Author
|
/ok to test 150e717 |
Contributor
Author
|
/ok to test fc5ea89 |
Contributor
Author
|
/ok to test 00ca581 |
Signed-off-by: Alexandros Koumparoulis <akoumparouli@nvidia.com>
Contributor
Author
|
/ok to test 550c136 |
Contributor
Author
|
/ok to test d208597 |
Collaborator
|
/claude review |
Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com>
Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com>
Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com>
Contributor
Author
|
/ok to test c507d8e |
Signed-off-by: Alexandros Koumparoulis <akoumparouli@nvidia.com>
Signed-off-by: Alexandros Koumparoulis <akoumparouli@nvidia.com>
Contributor
Author
|
/ok to test 5c6ccaa |
Contributor
Author
|
/ok to test 164e4f1 |
Contributor
Author
|
/claude review |
Signed-off-by: Alexandros Koumparoulis <akoumparouli@nvidia.com>
Contributor
Author
|
/ok to test 84ae1f8 |
Signed-off-by: Alexandros Koumparoulis <akoumparouli@nvidia.com>
Contributor
Author
|
/ok to test 85da4ee |
Contributor
Author
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What does this PR do ?
Previously, running a job could be accomplished with three different ways:
examples/<domain>_<task>/<task>.py -c config.yamlnemo_automodel/recipes/<domain>_<task>/<task>.py -c config.yamlautomodel llm finetune -c config.yamlThe first two ways to run a job, provide direct access to the script being run, however, it's still too long to type, and similarly the automodel CLI had to use the "llm finetune" part to specify the recipe.
This PR refactors the job launching, and now, config.yaml lists the recipe to use. As a result, we can now run
automodel config.yamlwhich is much easier to type than before. In addition, the python scripts under example/* will now be deprecated and instead users can use theautomodel/amCLI application. Removal will happen in 26.06, with 26.04 still containing them with a deprecation message.Example launch commands
python3 app.py config.yaml --nproc-per-node 8python3 app.py config.yaml(single GPU)torchrun --nproc-per-node 8 app.py config.yamlThe recommended command is to use app.py with python3, however, to avoid bad ux launching with torchrun is also supported.
Changelog
Before your PR is "Ready for review"
Pre checks:
If you haven't finished some of the above items you can still open "Draft" PR.
Additional Information