Support converting Qwen2.5-based MiniVLA checkpoint to HF format#2
Open
360ZMEM wants to merge 2 commits intoStanford-ILIAD:mainfrom
Open
Support converting Qwen2.5-based MiniVLA checkpoint to HF format#2360ZMEM wants to merge 2 commits intoStanford-ILIAD:mainfrom
360ZMEM wants to merge 2 commits intoStanford-ILIAD:mainfrom
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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:
This PR modifies file
vla-scripts/extern/convert_openvla_weights_to_hf.pyto enable the conversion of MiniVLA checkpoints (Qwen2.5 backbone) to standard HF format.Changes:
num_key_value_heads(GQA) which is present in Qwen2.5 configs but not in the default Llama mapping.<|extra_i|>tokens for models trained with theextratokenizer configuration.vocab_sizein the config matches the padded vocabulary size used during Prismatic training.Scope:
prism-qwen25-extra-dinosiglip-224px+0_5b(Standard Action Binning).VQActionTokenizermodels.