Skip to content

Pull requests: ROCm/Megatron-LM

New pull request New
10 Open 48 Closed
10 Open 48 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Fix data preprocessing bug Something isn't working
#58 opened Feb 11, 2025 by nsakkine • Draft
@nsakkine
1
fix: explicitly set weights_only to False during checkpoint loading to support PyTorch 2.6
#56 opened Feb 9, 2025 by mpashkovskii Loading…
feat: add Grok-1 transformer layer and training scripts
#55 opened Feb 7, 2025 by mpashkovskii • Draft
feat: add LoRA adapter layer and Mixtral LoRA training
#53 opened Jan 31, 2025 by mpashkovskii Loading…
59
Add FSDP arguments and example script to train model with FSDP-v2
#52 opened Jan 28, 2025 by ryang-amd Loading…
55
[Perf] Skip creating attention mask in llama dataloader
#40 opened Dec 13, 2024 by billishyahao Loading…
3
Enable pretraining qwen2 model
#33 opened Dec 10, 2024 by billishyahao Loading…
4
Make llama dataset preparation more generic.
#32 opened Dec 10, 2024 by lcskrishna Loading…
7
Renwuli/disable te qwen1.5
#27 opened Nov 21, 2024 by amdrenwuli Loading…
1
Renwuli/disable te qwen
#26 opened Nov 21, 2024 by amdrenwuli Loading…
2
ProTip! What’s not been updated in a month: updated:<2025-01-12.