Support Huggingface tokenizer #1269
Merged
+82
−12
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
khatwanimohit
requested review from
gobbleturk,
bvandermoon,
vipannalla and
RissyRan
as code owners
khatwanimohit
force-pushed
the
mohit/hf_tokenizer
branch
6 times, most recently
from
6e95d6b to
265f09b
Compare
anfals
reviewed
Feb 13, 2025
khatwanimohit
force-pushed
the
mohit/hf_tokenizer
branch
4 times, most recently
from
a618206 to
52c8ade
Compare
RissyRan
approved these changes
Feb 18, 2025
MaxText/tokenizer.py
Outdated
| return self.tokenizer.decode(t) | ||
|
|
||
|
|
||
| def build_tokenizer(tokenizer_path, tokenizer_type, add_bos, add_eos, hf_access_token): | ||
| """Loads the tokenizer at `tokenizer_path`""" | ||
| max_logging.log(f"Tokenizer path: {tokenizer_path}") | ||
| if "tiktoken" in tokenizer_path: |
There was a problem hiding this comment.
Wondering why the tiktoken type is slightly different here? i.e. why if tokenizer_type == "tiktoken" does not work?
| @@ -99,5 +107,23 @@ def test_detokenize(self): | |||
| self.assertEqual(np.asarray(self.source_tokenizer.decode(tokens)), np.asarray(text)) | |||
|
|
|||
|
|
|||
| class HFTokenizerTest(unittest.TestCase): | |||
khatwanimohit
force-pushed
the
mohit/hf_tokenizer
branch
3 times, most recently
from
412b0eb to
aa15d33
Compare
khatwanimohit
force-pushed
the
mohit/hf_tokenizer
branch
from
aa15d33 to
0aa574f
Compare
khatwanimohit
force-pushed
the
mohit/hf_tokenizer
branch
from
0aa574f to
3fdd18f
Compare
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Outdated suggestions cannot be applied.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.
Description
Support Huggingface tokenizer and decouple it out of the input pipeline.
If the change fixes a bug or a Github issue, please include a link, e.g.,:
b/394635939
Tests
Tested using
python MaxText/train.py MaxText/configs/base.yml tokenizer_path='google/gemma-2-2b-it' tokenizer_type=huggingface run_name=${USER}-$RANDOM model_name=gemma2-2b base_output_directory=gs://runner-maxtext-logs dataset_path=gs://maxtext-dataset per_device_batch_size=1.0 enable_checkpointing=false steps=2and added a test for comparing Gemma2-2b HF tokenizer with Sentencepiece.Checklist
Before submitting this PR, please make sure (put X in square brackets):