[RLlib] Speedup A3C up to 3x (new training_iteration function instead of execution_plan) and re-instate Pong learning test.
#22126
Conversation
This file contains 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
avnishn
approved these changes
Feb 4, 2022
|
|
||
| # Synch updated weights back to the particular worker. | ||
| with self._timers[SYNCH_WORKER_WEIGHTS_TIMER]: | ||
| weights = local_worker.get_weights(local_worker.get_policies_to_train()) |
rllib/agents/a3c/a3c.py
Outdated
| if global_vars: | ||
| local_worker.set_global_vars(global_vars) | ||
|
|
||
| # TODO: If we have processed more than one gradients |
There was a problem hiding this comment.
so to be clear we haven't written to result in this pr, right? But we want to for logging purposes.
There was a problem hiding this comment.
This is still WIP. I need to add proper compilation of the results dict. The only thing that's missing is to combine those learner stats from all workers that have returned something from the async_parallel_requests call further above. This shim implementation right now only returns the last one.
Let me finish this before merging, of course.
|
Merge pending results dict. |
6 tasks
simonsays1980
pushed a commit
to simonsays1980/ray
that referenced
this pull request
Feb 27, 2022
…ad of `execution_plan`) and re-instate Pong learning test. (ray-project#22126)
simonsays1980
pushed a commit
to simonsays1980/ray
that referenced
this pull request
Feb 27, 2022
…on instead of `execution_plan`) and re-instate Pong learning test." (ray-project#22250) Reverts ray-project#22126 Breaks rllib:tests/test_io
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Outdated suggestions cannot be applied.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.
This PR:
training_iterationfunction for A3C (alternative to existingexecution_plan)._disable_execution_plan_api=True).tuned_examples/a3c/pong-a3c.yaml(16 worker, LSTM+CNN Atari problem).tuned_examples/a3c/pong-a3c.yaml).Why are these changes needed?
Related issue number
Checks
scripts/format.shto lint the changes in this PR.