Learner test build #75
Merged
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
…p, this prevents from passing data through ray object store
…ray object store instead of having each actor loading the data.
… as a slice. i.e. if __getitem__(4) was called, would return everything from 0 to 4. This was crashing torch DataLoader. Issue was fixed by re-creating a dataframe when being asked single idx.
- stack during forward was not well executed - target shape was incorrect
suzannejin
reviewed
Jan 31, 2025
| validation=TorchDataset( | ||
| config_path=config_path, | ||
| csv_path=data_path, | ||
| # Configure trainable with resources and dataset parameters |
There was a problem hiding this comment.
I don't remember exactly why, but I think initially I put them here to avoid the need for reencoding each time it is tuned.
suzannejin
reviewed
Jan 31, 2025
| @@ -138,13 +139,31 @@ def tuner_initialization( | |||
|
|
|||
| logging.info(f"PER_TRIAL resources -> GPU: {self.gpu_per_trial} CPU: {self.cpu_per_trial}") | |||
|
|
|||
| # Pre-load and encode datasets once, then put them in Ray's object store | |||
| @ray.remote | |||
There was a problem hiding this comment.
ahhh so ray.remote is the correct way to do it :)
suzannejin
reviewed
Jan 31, 2025
|
|
||
| # Put datasets in Ray's object store | ||
| datasets_ref = create_datasets.remote(data_config_path, data_path, encoder_loader) | ||
| self.config["_training_ref"] = training_ref |
There was a problem hiding this comment.
i like how your thought process evolved and went back XD
Btw, why you chose to use config again, instead of remote function?
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Outdated suggestions cannot be applied.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.
added test for learner and made sur it runs end to end,
TODO: add proper debugging tool to assess size missmatch