Skip to content
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

[RLlib] Use observations (input_dict) for exploration #27979

Closed
wants to merge 761 commits into from

Conversation

Stefan-1313
Copy link

@Stefan-1313 Stefan-1313 commented Aug 18, 2022

@kouroshHakha

Why are these changes needed?

Use observations (input_dict) for exploration.
For details, please see issue: #26437 .

Related issue number

If this pull request is correct(ed), it closes enhancement issue #26437.

Checks

  • [ I did, some afterward, I don't know if this is how you want it... ] I've signed off every commit(by using the -s flag, i.e., git commit -s) in this PR.
  • [ I'm on Windows, does not work ] I've run scripts/format.sh to lint the changes in this PR.
  • [ I think this will be auto generated, I updated the doc string in the code for the developer API. ] I've included any doc changes needed for https://docs.ray.io/en/master/.
  • [ Yes, except for tests that already failed before my changes. Among them tests when using framework "tf", perhaps this is due to my environment, when using frameworks "torch" and "tf2" tests pass ] I've made sure the tests are passing. Note that there might be a few flaky tests, see the recent failures at https://flakey-tests.ray.io/
  • Testing Strategy
    • [ X ] Unit tests
    • Release tests
    • This PR is not tested :(
    • [ X ] A test program I normally use, runs fine for both "torch" and "tf2" frameworks.

I did my best, this is my first Pull Request ever so please go easy on me :).
I'm not an very experienced Git user, so I had trouble fixing all the sign off stuff.

EDIT:

I tried to fix the sign off issue with the instructions in the DCO check (git rebase HEAD~7 --signoff and git push --force-with-lease origin master), and now I have seen to signed off everyone's commit :O. That cannot be how it should be... But the DCO check is not complaining any more! :)

kouroshHakha and others added 22 commits August 18, 2022 11:35
…s` (ray-project#27284)

Add optional last_activity_at field to /api/component_activities to record end time of most recently finished activity

Signed-off-by: Nikita Vemuri <[email protected]>
Signed-off-by: Stefan van der Kleij <[email protected]>
Signed-off-by: Eric Liang <[email protected]>
Signed-off-by: Stefan van der Kleij <[email protected]>
…-project#27370)

We decided to allow escaping the parent pg via `PlacementGroupSchedulingStrategy(placement_group=None)` instead of using "DEFAULT". Our doc is updated with that but in the code it's still not allowed.

Signed-off-by: Stefan van der Kleij <[email protected]>
…c_code dir (ray-project#27350)

Edit pass. Move code into doc_code dir. Code in doc_code is verified by CI

Signed-off-by: Stefan van der Kleij <[email protected]>
Co-authored-by: Kai Fricke <[email protected]>
Signed-off-by: Stefan van der Kleij <[email protected]>
…ay-project#26969)

Moves FastAPI into its own section instead of appearing in a duplicated note.

Co-authored-by: simon-mo <[email protected]>
Signed-off-by: Stefan van der Kleij <[email protected]>
…opy-editing

Signed-off-by: Eric Liang <[email protected]>
Signed-off-by: Stefan van der Kleij <[email protected]>
…t#27377)

Why are these changes needed?
python/tests/test_state_api.py runs for 5min in normal run

Signed-off-by: Stefan van der Kleij <[email protected]>
…an a NumPy conversion (ray-project#26891)

Converting a Pandas DataFrame column to an ndarray (e.g. via df[col].values) can often result in a full copy of the column in order to construct the ndarray due to Pandas' 2D block management. This PR ports tensor extension type checking to checking the dtype, which is always an O(1) check.

Signed-off-by: Clark Zinzow <[email protected]>
Signed-off-by: Stefan van der Kleij <[email protected]>
Support a GPU column for the new dashboard

Have first node be default expanded

Signed-off-by: Alan Guo [email protected]

fixes ray-project#13889

Addresses comment from ray-project#26996

Signed-off-by: Stefan van der Kleij <[email protected]>
    Signed-off-by: Dmitri Gekhtman <[email protected]>

This PR

adds a page of guidance on GPU deployment with Ray/K8s. This page is a modified and slightly expanded version of the existing page https://docs.ray.io/en/latest/cluster/kubernetes-gpu.html
moves managed K8s service intro links to their own page

Signed-off-by: Stefan van der Kleij <[email protected]>
Update API references to beta. Needed as we are going to beta in 2.0.

I left out RL/Scikit-Learn/HuggingFace.

Signed-off-by: Stefan van der Kleij <[email protected]>
- Currently not all code under ray-core/doc_code is covered by CI.
- tf_example.py and torch_example.py are not used anywhere.

Signed-off-by: Jiajun Yao <[email protected]>
Signed-off-by: Stefan van der Kleij <[email protected]>
christy and others added 7 commits August 18, 2022 12:56
Signed-off-by: Cheng Su <[email protected]>
This is to run ray 2.0.0rc0 on https://docs.ray.io/en/master/data/examples/nyc_taxi_basic_processing.html and fix the notebook end-to-end, make sure the output and wording is matched.

The page after this PR - https://ray--27927.org.readthedocs.build/en/27927/data/examples/nyc_taxi_basic_processing.html .

Signed-off-by: Stefan van der Kleij <[email protected]>
Signed-off-by: Stefan van der Kleij <[email protected]>
…y-project#27901)

Fix 2.0.0 release blocker bug where Ray State API and Jobs not accessible if the override URL doesn't support adding additional subpaths. This PR keeps the localhost dashboard URL in the internal KV store and only overrides in values printed or returned to the user.
images.githubusercontent.com/6900234/184809934-8d150874-90fe-4b45-a13d-bce1807047de.png">

Signed-off-by: Stefan van der Kleij <[email protected]>
Signed-off-by: Stefan van der Kleij <[email protected]>
Signed-off-by: Stefan van der Kleij <[email protected]>
Signed-off-by: Stefan van der Kleij <[email protected]>
…both PyTorch and TensorFlow.

Signed-off-by: Stefan van der Kleij
Signed-off-by: Stefan van der Kleij <[email protected]>
Signed-off-by: Stefan van der Kleij <[email protected]>
…ring.

Signed-off-by: Stefan van der Kleij
Signed-off-by: Stefan van der Kleij <[email protected]>
Signed-off-by: Stefan van der Kleij <[email protected]>
Signed-off-by: Stefan van der Kleij
Signed-off-by: Stefan van der Kleij <[email protected]>
Signed-off-by: Stefan van der Kleij <[email protected]>
@stale
Copy link

stale bot commented Sep 20, 2022

This pull request has been automatically marked as stale because it has not had recent activity. It will be closed in 14 days if no further activity occurs. Thank you for your contributions.

  • If you'd like to keep this open, just leave any comment, and the stale label will be removed.

@stale stale bot added the stale The issue is stale. It will be closed within 7 days unless there are further conversation label Sep 20, 2022
@Stefan-1313
Copy link
Author

Commenting because of stale[bot].

This pull request is still usefull, and should still be able to merge without (much) problems.
Only a few files have changed and the sign off command I did somehow made it such that all commits that where done since I created my fork are seen as commits for this pull request.

Only look at the latest commits, only these are mine and are the relevant ones.

@xwjiang2010
Copy link
Contributor

@Stefan-1313
Thanks for the PR!
Could you clean up the commit history? With its current state, it is hard to get it reviewed and merged.

To fix it, i would just start a new brach, cherry-pick your PR onto that branch. Please also do git commit --amend --signoff to make sure that DCO check will pass. Then the same procedure with git push stuff (that will give you a new PR but I think it's ok in this case).

@kouroshHakha assigning this to you as you have the most context about the problem the PR is addressing.

@stale stale bot removed the stale The issue is stale. It will be closed within 7 days unless there are further conversation label Sep 21, 2022
@Stefan-1313
Copy link
Author

@xwjiang2010 I will try to clean up my commit history.
Currently I'm a bit low on time, but I will do it when I find some time.
Should I then create a new pull request? If yes, then I close this one after I created the new pull request.

@xwjiang2010
Copy link
Contributor

Yes, that would work. Thank you!

@kouroshHakha
Copy link
Contributor

@Stefan-1313 Please assign the clean PR to me. Thanks.

@stale
Copy link

stale bot commented Oct 28, 2022

This pull request has been automatically marked as stale because it has not had recent activity. It will be closed in 14 days if no further activity occurs. Thank you for your contributions.

  • If you'd like to keep this open, just leave any comment, and the stale label will be removed.

@stale stale bot added the stale The issue is stale. It will be closed within 7 days unless there are further conversation label Oct 28, 2022
@stale
Copy link

stale bot commented Nov 12, 2022

Hi again! The issue will be closed because there has been no more activity in the 14 days since the last message.

Please feel free to reopen or open a new issue if you'd still like it to be addressed.

Again, you can always ask for help on our discussion forum or Ray's public slack channel.

Thanks again for opening the issue!

@stale stale bot closed this Nov 12, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
stale The issue is stale. It will be closed within 7 days unless there are further conversation
Projects
None yet
Development

Successfully merging this pull request may close these issues.