ExperimentHparams class; Set state.train_dataloader
#966
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
…ps_per_epoch`. 1. Made the `state.dataloader` optional, since it will not be provided on `__init__` as part of mosaicml#40. 2. Binding the active dataloader to the state on `Event.FIT_START`, and switching the dataloader to each evaluation dataloader before `Event.EVAL_START`. Restoring the previous (training) dataloader after `Event.EVAL_END`. 3. Moved `Event.EVAL_START` and `Event.EVAL_END` to run for each evaluator, instead of once for all evaluators. With mosaicml#40, the eval() will take in a dataloader, which would then require `Event.EVAL_START` and `Event.EVAL_END`. This change also permits for algorithms that wish to modify (each) evalution dataloader. 4. Moved scaling of the LR schedulers to `Trainer.fit()` before `Event.FIT_START` fires. Schedulers will be passed in on `Trainer.fit()` as part of mosaicml#40. 5. Removed `steps_per_epoch` as part of the state. Instead, algorithms and callbacks can read len(state.dataloader) directly. While this change will make schedulers no longer accurate when using `train_subset_num_batches`, that flag should only be used for performance measurements. As such, it is not necessarry that SSR behaves correctly for performance runs. Added a warning for the `train_subset_num_batches` field. Implements the first part of mosaicml#40. Closes mosaicml#363.
It can be useful for algorithms and callbacks to know which dataloader is active, so added the `dataloader_label` to the state. Removed `evaluators` from state, as nothing is using that anymore.
…er into ravi/optional_dataloader
…()`. Preferable to keep variables on the state object rather than as trainer members, where appropriate. Before, the state.schedulers was empty after `__init__()` but before `fit()`. Now, state.schedulers contains the compiled composer schedulers or original pytorch schedulers. Restored optimizers on Event.INIT; it will be a bigger issue to rewrite algs to not depend on optimizers on init.
* Removed `precision_context` from state * Switched `train_subset_num_batches` and `eval_subset_num_batches` to use `-1` as the default value instead of `None`.
…er into ravi/optional_dataloader
This is a non-breaking change; all existing trainer hparams will work as-is. |
…b.com:ravi-mosaicml/ravi-composer into experiment_hparams
f8bd71b to
d3689af
Compare
A-Jacobson
approved these changes
May 11, 2022
ExperimentHparams classExperimentHparams class; Set state.train_dataloader
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.
ExperimentHparamsclass. This class describes how to run a training job that may have multiple calls toTrainer.fitand/orTrainer.eval. Specifically,ExperimentHparams.initialize_object()returns a(Trainer, List[FitKwargs], List[EvalKwargs])tuple, that then the user's entrypoint can consome.This class does not automatically train the model, nor does it include an entrypoint.
FitKwargsandEvalKwargs, along with test cases to ensure they stay in sync with the Trainer signature.Trainer.fit()#948, which removed the setting ofState.train_dataloader. Added back the lines to correctly set the train dataloader.