[WIP] support different tokenizers #1318
Draft
Conversation
This file contains hidden or 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
facebook-github-bot
added
the
CLA Signed
H-Huang
commented
Jun 18, 2025
| @@ -0,0 +1,225 @@ | |||
| #!/usr/bin/env python3 | |||
There was a problem hiding this comment.
Will remove this file, i was just using to test
wwwjn
reviewed
Jun 18, 2025
| model_name = repo_id.split("/")[-1] | ||
| model_dir = os.path.join(local_dir, model_name) | ||
|
|
||
| # Common tokenizer files to download |
There was a problem hiding this comment.
This might not be comprehensive and need user's double-check, eg, CLIP tokenizer used in FLUX needs special_tokens_map.json and vocab.json: https://huggingface.co/openai/clip-vit-large-patch14/tree/main
There was a problem hiding this comment.
Thanks, good point! Let me look into how image tokenizers work too
There was a problem hiding this comment.
CLIP is a text encoder instead of image encoder, and I feel like the "transformers" repo supports more tokenizers than the tokenizer repo
fegin
reviewed
Jun 18, 2025
| python scripts/download_tokenizer.py --repo_id meta-llama/Meta-Llama-3.1-8B --hf_token=... | ||
|
|
||
| # DeepSeek tokenizer (automatically downloads tokenizer.json and tokenizer_config.json) | ||
| python scripts/download_tokenizer.py --repo_id deepseek-ai/DeepSeek-V3 |
There was a problem hiding this comment.
Just curious, is hf_token not required?
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Outdated suggestions cannot be applied.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.
Requesting feedback since this will change the tokenizer directory for existing users. We need to add support more multiple tokenizers in titan (we currently only have the one used in llama3)
What happens currently:
scripts/download_tokenizer.pyassets/tokenizer/original/tokenizer.modeltokenizer.modelin the.tomlconfigs (https://github.com/pytorch/torchtitan/blob/main/torchtitan/models/llama3/train_configs/llama3_8b.toml#L21tokenizer.modeland wrap with our owntiktokentokenizer (https://github.com/pytorch/torchtitan/blob/main/torchtitan/datasets/tokenizer/tiktoken.py)New workflow:
scripts/download_tokenizer.pyassets/tokenizer/<model_name>/.tomlconfigsPros:
Cons:
original/tokenizer.model