forked from NVIDIA/Megatron-LM
-
Notifications
You must be signed in to change notification settings - Fork 0
Pull requests: yashaswikarnati/Megatron-LM
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
NMFW-464: set production Energon workers to 2
#38
opened May 23, 2026 by
yashaswikarnati
Owner
•
Draft
NMFW-478: Add --load-vision-from for hetero MIMO encoder DCP loading
#28
opened May 16, 2026 by
yashaswikarnati
Owner
•
Draft
7 tasks
Add detailed hetero MIMO timeline profiling
#25
opened May 14, 2026 by
yashaswikarnati
Owner
•
Draft
Thread custom process groups through MoE grad finalization
#19
opened May 13, 2026 by
yashaswikarnati
Owner
Loading…
Pass explicit process groups to hybrid logging
#18
opened May 13, 2026 by
yashaswikarnati
Owner
Loading…
NMFW-464: HyperCommGrid alt-factorization + Nemotron VLM E2E (MIMO hetero parallel)
#17
opened May 10, 2026 by
yashaswikarnati
Owner
•
Draft
11 tasks done
Add PP>1 support for colocated MIMO training (NMFW-19)
#15
opened Apr 28, 2026 by
yashaswikarnati
Owner
•
Draft
docs: add Megatron-LM skills for unit testing and SLURM execution
#14
opened Apr 28, 2026 by
yashaswikarnati
Owner
•
Draft
1 of 3 tasks
docs: add agentic heterogeneous parallelism discussion article
#13
opened Apr 28, 2026 by
yashaswikarnati
Owner
•
Draft
1 of 3 tasks
[Fork review] NMFW-17: dest CP>1 support on top of pr-a
#11
opened Apr 18, 2026 by
yashaswikarnati
Owner
•
Draft
Add PP>1 support for LLM in colocated MIMO training (NMFW-19)
#9
opened Apr 17, 2026 by
yashaswikarnati
Owner
•
Draft
2 tasks done
ProTip!
Follow long discussions with comments:>50.