Tokenizer overhaul (complete) #38
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
…frozendict dependency
…uple of `StepTokenizer`s.
…ns_backwards_compatible` and `test_to_legacy_tokenizer` passing
…to resolve circular imports. Tests not run yet. Will need to refactor imports in maze-transformer and other repos
… few key tests are passing, but I suspect that overriding `TokenizerElement.__hash__` wholesale is not a good idea.
…`Cardinal` and its references since VSCode pytest extension was crashing during test collection. When testing other part of the code, recommend nullifying `Cardinal` and `Relative`. This may be done by either commenting them and all references out or by adding `@abc.abstractmethod` to their method to remove them from consideration by `all_instances(StepTokenizer)`. Would still need to remove any uses that attempted to construct `Cardinal()` or `Relative()`, which right now only exist in `test_tokenizer.py`.
….Relative` since building and testing `PathTokenizers` is done for now. Need to reactivate them later.
…ad` failing, as expected.
…t be why tests have been taking so long.
…ed tokenizer regions. Lots of superfluous warnings
wandb fails in the backend when passing data with this key and a value of longer than 64 chars, see: wandb/wandb#8159 this was causing integration tests to hang idefinitely in understanding-search/maze-transformer#214
…tep2 Update to `mark_as_unsupported`
…nderstanding-search/maze-dataset into tokenizer-overhaul-integration
I think it's fine to have shorthands for these things, as well as to have backwards compatibility.
mivanit
commented
Aug 28, 2024
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Outdated suggestions cannot be applied.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.
TODOS:
(roughly in order of importance)
checks.yml:all-tok-checks.yml:depends on Tokenizer Overhaul Step 2 #37 being merged into
tokenizer-overhaul-integrationfirstdepends on tests passing in
maze-transformer#214make sure that tokenizer hash numpy file is shipped withmaze-datasetpackagemaybe make it optional, only iftokenizationextra is requested?go over existing
# TODOcomments, see if anything is criticaltests/unit/maze_dataset/generation/test_maze_dataset.pyclean up various superfluous warnings
consider optimizations to
all_instancesandsave_hashessave_hashesis being properly parallelized :/see if any more utils should be moved to muutils