-
Notifications
You must be signed in to change notification settings - Fork 1.1k
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
Add ClearMLHandler to track all MONAI Experiments #6013
Conversation
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.
Thanks for the contribution.
Please add some unit tests to cover the feature.
Thanks.
Working on adding some unittests. |
for more information, see https://pre-commit.ci
for more information, see https://pre-commit.ci
for more information, see https://pre-commit.ci
@wyli , @Nic-Ma , @binliunls, is the memory ran out of error on these Tests normal? The failing tests are blocking the PR. |
yes, that issue is probably introduced by the code changes in this PR. any idea what the root cause is? we can skip the windows test using Line 247 in b85e2c6
|
@wyli , Could you please clarify what the Also, I would request that can you please check monitoring on the runner machine, so we can understand if it truly ran out of memory. |
that's testing the bundle command: python -m monai.bundle run --runner_id evaluating ... you can run the test locally with:
@Nic-Ma @binliunls could you please help the debugging here? |
/build |
Signed-off-by: skinan <[email protected]>
Signed-off-by: skinan <[email protected]>
looking at the error messages, it seems the error is from the clearml fire binding on windows.
https://github.com/Project-MONAI/MONAI/actions/runs/4371805968/jobs/7648041295 |
Signed-off-by: skinan <[email protected]>
Signed-off-by: skinan <[email protected]>
I, Victor Sonck <[email protected]>, hereby add my Signed-off-by to this commit: ce0ac23 Signed-off-by: Victor Sonck <[email protected] Signed-off-by: Victor Sonck <[email protected]>
Signed-off-by: skinan <[email protected]>
/build |
the docker test is blocked by the unit test https://github.com/Project-MONAI/MONAI/actions/runs/4397243843/jobs/7700566567 could you please help debug it? @skinan the test is running on a github runner https://github.com/Project-MONAI/MONAI/actions/runs/4397243843/workflow#L83-L99 |
@wyli , This is weird. We are trying to reproduce it locally, to find out what goes wrong. |
thanks, I try to reproduce it on github by manually triggering the test dev...wyli:MONAI:trigger-tests but it works fine now https://github.com/wyli/MONAI/actions/runs/4415391874 I'll monitor this test for the dev branch |
**Issue:** fixes #6148 **Previous Pull-request:** #6013 --------- Signed-off-by: skinan <[email protected]>
**Issue:** fixes Project-MONAI#6148 **Previous Pull-request:** Project-MONAI#6013 --------- Signed-off-by: skinan <[email protected]>
**Issue:** fixes Project-MONAI#6148 **Previous Pull-request:** Project-MONAI#6013 --------- Signed-off-by: skinan <[email protected]>
### Description Add option to install `clearml` as optional dependency with `pip install monai[clearml]`. All of the docstrings were updated and unit tests added in [PR #6013](#6013). ### Types of changes <!--- Put an `x` in all the boxes that apply, and remove the not applicable items --> - [x] Non-breaking change (fix or new feature that would not break existing functionality). - [ ] Breaking change (fix or new feature that would cause existing functionality to change). - [ ] Integration tests passed locally by running `./runtests.sh -f -u --net --coverage`. - [ ] Quick tests passed locally by running `./runtests.sh --quick --unittests --disttests`. --------- Signed-off-by: revital <[email protected]>
ClearML is a leading MLOps stack that can supercharge dialogues research with its state-of-the-art experiment tracking capability. ClearML: https://clear.ml/
I have added a
clearml_handler.py
which contains ClearMLHandler, ClearMLStatsHandler, and ClearMLImageHandler class. Basically, clearml can track everything which is tracked by tensorboard including scalars, and debug samples, and can also store models, artifacts, and the console in the ClearML server which can be easily accessed from ClearML WebUI shown as below:ClearMLStatsHandler and ClearMLImageHandler can be used with Pytorch Trainer just like TensorboardStatsHandler and TensorboardImageHandler.
Use
ClearMLStatsHandler()
&ClearMLImageHandler(log_dir="./runs/", batch_transform=from_engine(["image", "label"]),output_transform=from_engine(["pred"]),)
with any MONAI example to test its functionality.Also, please let us where should we put further documentation and tutorials regarding this MLOps tool in the MONAI.