Yu Yao
Yu Yao
### Question This line seems to do nothing? https://github.com/haotian-liu/LLaVA/blob/b080bb98b7445a5a11aecec44c9bf9ab1c5feefc/llava/model/llava.py#L142 Is it a placeholder or anything?
# What does this PR do ? Add a one line overview of what this PR aims to accomplish. **Collection**: [Note which collection this PR will affect] # Changelog -...
# What does this PR do ? Unify HF converter scripts and arguments. Moved from https://github.com/NVIDIA/NeMo/pull/8307. Change target to main. # Jenkins CI To run Jenkins, a NeMo User with...
# What does this PR do ? Add llama3 and distributed checkpoint support in NeVA **Collection**: [multimodal] # Changelog - Add specific line by line info of high level changes...
# What does this PR do ? In this pr, we updated docs for TP/PP/VP/CP and EP (moved from https://github.com/NVIDIA/NeMo/pull/9148). We added a section for GQA/MQA. We also added Seq...
# What does this PR do ? Add guards to SD imports. So that pip install'ed NeMo won't fail the imports. **Collection**: [multimodal] # Changelog - Add specific line by...
# What does this PR do ? Update NeVA, GPT and optimizer part of the code to enable per token loss in NeVA. This can be turned on by `calculate_per_token_loss`....
# What does this PR do ? add Per Token Loss Support in NeVA **Collection**: [Note which collection this PR will affect] # Changelog - Add specific line by line...
# What does this PR do ? Fix trainer builder when exp_manager is not in config **Collection**: [Note which collection this PR will affect] # Changelog - Add specific line...
# What does this PR do ? Update neva conversion script from and to HF **Collection**: [Note which collection this PR will affect] # Changelog - Add specific line by...