-
Notifications
You must be signed in to change notification settings - Fork 24
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
ci: Add build-test-publish-wheel
workflow
#114
base: main
Are you sure you want to change the base?
Changes from all commits
File filter
Filter by extension
Conversations
Jump to
Diff view
Diff view
There are no files selected for viewing
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,46 @@ | ||
# Copyright (c) 2020-2021, NVIDIA CORPORATION. | ||
# | ||
# Licensed under the Apache License, Version 2.0 (the "License"); | ||
# you may not use this file except in compliance with the License. | ||
# You may obtain a copy of the License at | ||
# | ||
# http://www.apache.org/licenses/LICENSE-2.0 | ||
# | ||
# Unless required by applicable law or agreed to in writing, software | ||
# distributed under the License is distributed on an "AS IS" BASIS, | ||
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | ||
# See the License for the specific language governing permissions and | ||
# limitations under the License. | ||
|
||
name: Build, test, and publish a PyPi wheel (to testpypi) | ||
|
||
on: | ||
push: | ||
branches: | ||
- main | ||
- 'r**' | ||
|
||
defaults: | ||
run: | ||
shell: bash -x -e -u -o pipefail {0} | ||
|
||
jobs: | ||
build-test-publish-wheel: | ||
uses: NVIDIA/NeMo-FW-CI-templates/.github/workflows/[email protected] | ||
with: | ||
image-name: nemo_run_container | ||
dockerfile: Dockerfile | ||
image-label: nemo-run | ||
build-args: | | ||
IMAGE_LABEL=nemo-run | ||
NEMO_RUN_COMMIT=${{ github.event.pull_request.head.sha || github.sha }} | ||
prune-filter-timerange: 24h | ||
dry-run: true | ||
python-package: nemo_run/src | ||
container-workdir: /opt/NeMo-Run | ||
environment: public | ||
secrets: | ||
TWINE_USERNAME: ${{ secrets.TWINE_USERNAME }} | ||
TWINE_PASSWORD: ${{ secrets.TWINE_PASSWORD }} | ||
SLACK_WEBHOOK_ADMIN: ${{ secrets.SLACK_WEBHOOK_ADMIN }} | ||
SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }} |
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,31 @@ | ||
FROM python:3.12 as nemo-run-update | ||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. We don't really need a docker file for NeMo-Run since it can be pip installed everywhere and is relatively limited in scope. There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. I'd need to refactor the workflows a bit for this, but I think it would be worth doing so. Let me compare across projects if there some barriers I'm not seeing yet |
||
WORKDIR /opt | ||
ARG NEMO_RUN_COMMIT | ||
RUN <<"EOF" bash -exu | ||
if [[ ! -d NeMo-Run ]]; then | ||
git clone https://github.com/NVIDIA/NeMo-Run.git | ||
fi | ||
cd NeMo-Run | ||
git init | ||
git remote add origin https://github.com/NVIDIA/NeMo-Aligner.git | ||
git fetch --all | ||
git fetch origin '+refs/pull/*/merge:refs/remotes/pull/*/merge' | ||
git checkout $NEMO_RUN_COMMIT | ||
EOF | ||
|
||
FROM python:3.12 | ||
|
||
RUN \ | ||
--mount=type=bind,source=/opt/NeMo-Run/src/nemo_run/__init__.py,target=/opt/NeMo-Run/src/nemo_run/__init__.py,from=nemo-run-update \ | ||
--mount=type=bind,source=/opt/NeMo-Run/src/nemo_run/package_info.py,target=/opt/NeMo-Run/src/nemo_run/package_info.py,from=nemo-run-update \ | ||
--mount=type=bind,source=/opt/NeMo-Run/pyproject.toml,target=/opt/NeMo-Curator/pyproject.toml,from=nemo-run-update \ | ||
cd /opt/NeMo-Run && \ | ||
pip install . | ||
|
||
COPY --from=nemo-run-update /opt/NeMo-Run/ /opt/NeMo-Run/ | ||
|
||
# Clone the user's repository, find the relevant commit, and install everything we need | ||
RUN bash -exu <<EOF | ||
cd /opt/NeMo-Run/ | ||
pip install . | ||
EOF |
Original file line number | Diff line number | Diff line change |
---|---|---|
|
@@ -16,7 +16,8 @@ | |
from nemo_run import cli | ||
from nemo_run.api import autoconvert, dryrun_fn | ||
from nemo_run.config import Config, ConfigurableMixin, Partial, Script | ||
from nemo_run.core.execution.base import Executor, ExecutorMacros, FaultTolerance, Torchrun | ||
from nemo_run.core.execution.base import (Executor, ExecutorMacros, | ||
FaultTolerance, Torchrun) | ||
from nemo_run.core.execution.docker import DockerExecutor | ||
from nemo_run.core.execution.local import LocalExecutor | ||
from nemo_run.core.execution.skypilot import SkypilotExecutor | ||
|
@@ -32,6 +33,11 @@ | |
from nemo_run.run.experiment import Experiment | ||
from nemo_run.run.plugin import ExperimentPlugin as Plugin | ||
|
||
from .package_info import (__contact_emails__, __contact_names__, | ||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Can we use absolute path here? |
||
__description__, __download_url__, __keywords__, | ||
__license__, __package_name__, __repository_url__, | ||
__shortversion__, __version__) | ||
|
||
__all__ = [ | ||
"autoconvert", | ||
"cli", | ||
|
@@ -61,9 +67,14 @@ | |
"SlurmExecutor", | ||
"SSHTunnel", | ||
"Torchrun", | ||
"__contact_emails__", | ||
"__contact_names__", | ||
"__description__", | ||
"__download_url__", | ||
"__keywords__", | ||
"__license__", | ||
"__package_name__", | ||
"__repository_url__", | ||
"__shortversion__", | ||
"__version__" | ||
] | ||
|
||
try: | ||
from nemo_run._version import __version__ | ||
except Exception: | ||
__version__ = "0.0.1" |
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,37 @@ | ||
# Copyright (c) 2023, NVIDIA CORPORATION. All rights reserved. | ||
# | ||
# Licensed under the Apache License, Version 2.0 (the "License"); | ||
# you may not use this file except in compliance with the License. | ||
# You may obtain a copy of the License at | ||
# | ||
# http://www.apache.org/licenses/LICENSE-2.0 | ||
# | ||
# Unless required by applicable law or agreed to in writing, software | ||
# distributed under the License is distributed on an "AS IS" BASIS, | ||
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | ||
# See the License for the specific language governing permissions and | ||
# limitations under the License. | ||
|
||
|
||
MAJOR = 0 | ||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. We have hatch's vcs for versioning currently. See here - https://github.com/NVIDIA/NeMo-Run/blob/main/pyproject.toml#L118-L122. Can you remove it if the versions are going to be manually maintained? There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Oh that's sad, hatch seems like a cool tool. But I can't see how it fits into Nemo's release cycle. Because both pre-release and releases are rather github-workflow than some-other-cli-tool driven. I'd remove hatch for now |
||
MINOR = 0 | ||
PATCH = 1 | ||
PRE_RELEASE = "rc0" | ||
DEV = "dev1" | ||
|
||
# Use the following formatting: (major, minor, patch, pre-release) | ||
VERSION = (MAJOR, MINOR, PATCH, PRE_RELEASE, DEV) | ||
|
||
__shortversion__ = ".".join(map(str, VERSION[:3])) | ||
__version__ = __shortversion__ + VERSION[3] + "." + ".".join(VERSION[4:]) | ||
|
||
__package_name__ = "nemo_run" | ||
__contact_names__ = "NVIDIA" | ||
__contact_emails__ = "[email protected]" | ||
__repository_url__ = "https://github.com/NVIDIA/NeMo-Run" | ||
__download_url__ = "https://github.com/NVIDIA/NeMo-Run/releases" | ||
__description__ = ( | ||
"NeMo-Run - A powerful tool designed to streamline the configuration, execution and management of Machine Learning experiments across various computing environments." | ||
) | ||
__license__ = "Apache2" | ||
__keywords__ = "deep learning, machine learning, gpu, NLP, NeMo, nvidia, pytorch, torch, language, preprocessing, LLM, large language model" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We use rye/uv for pacakge management. Can we use it to build and publish the package as well?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Out of curiosity, what would be the benefit? My workflows use standard
build
andtwine
for building and publishing. It should be complaint to rye and uv since it also follows PEP standards. Not all projects use rye/uv so I need to find a common denominator across projects. Of course we could allow injection your own custom tool, but I wonder which benefit it would bring