Skip to content

Actions: NVIDIA/Megatron-LM

Actions

Claude Copy PR to Main

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
claude-copy-to-main.yml
247 workflow runs
247 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

[Question] Enable GDN Packed Sequence Support for Context Parallelism in Qwen 3.5 series Claude Copy PR to Main #247: Issue comment #4043 (comment) created by yuzhongw-nvidia
1s
1s
Add Dual Chunk Attention (DCA) for long-context training Claude Copy PR to Main #246: Issue comment #4048 (comment) created by Ternura143
1s
1s
Feature Request: Dual Chunk Attention for Long Context Claude Copy PR to Main #245: Issue comment #2797 (comment) created by Ternura143
1s
1s
[Bug] HybridEP dispatcher passes incorrect max_num_of_tokens_per_rank to DeepEP, causing RDMA QP assertion failure Claude Copy PR to Main #244: Issue comment #3999 (comment) created by Shreyas-S-809
1s
1s
Add Dual Chunk Attention (DCA) for long-context training Claude Copy PR to Main #243: Issue comment #4048 (comment) created by copy-pr-bot bot
1s
1s
[draft]DistTrain demo for mutil module training Claude Copy PR to Main #242: Issue comment #1995 (comment) created by chenhongyu2048
2s
2s
[QUESTION] Support for Xiaomi MiMo-V2-Flash Architecture (Hybrid Attention + Fine-Grained MoE) Claude Copy PR to Main #241: Issue comment #2976 (comment) created by beccohov
1s
1s
megatron.core.datasets SFTDataset and get_batch* functions refactor Claude Copy PR to Main #240: Issue comment #3542 (comment) created by asolergi-nv
1s
1s
Miscellaneous inference fixes Claude Copy PR to Main #239: Issue comment #4030 (comment) created by santhnm2
1s
1s
Miscellaneous inference fixes Claude Copy PR to Main #238: Issue comment #4030 (comment) created by santhnm2
1s
1s
Miscellaneous inference fixes Claude Copy PR to Main #237: Issue comment #4030 (comment) created by santhnm2
2s
2s
Energy Efficiency: 10 Mathematical Techniques for 60-70% AI Energy Reduction (Phi6Simple, FFT-Mix, Phi MoE) Claude Copy PR to Main #236: Issue comment #4037 (comment) created by sbhavani
1s
1s
[ENHANCEMENT] RFC: upstreaming Qwen3.5 multimodal support into Megatron-LM Claude Copy PR to Main #235: Issue comment #3985 (comment) created by sbhavani
2s
2s
fix: wait for async P2P send before deallocating output tensor Claude Copy PR to Main #234: Issue comment #4047 (comment) created by copy-pr-bot bot
1s
1s
Implement scatter_gather_tensors_in_pipeline logic Claude Copy PR to Main #233: Issue comment #4032 (comment) created by ko3n1g
1s
1s
Implement scatter_gather_tensors_in_pipeline logic Claude Copy PR to Main #232: Issue comment #4032 (comment) created by QingZhou-YangHY
1s
1s
Fix HybridDeviceOptimizer KeyError after mixed-precision param replacement Claude Copy PR to Main #231: Issue comment #4046 (comment) created by copy-pr-bot bot
7s
7s
Custom step batch size rampup schedules Claude Copy PR to Main #230: Issue comment #3779 (comment) created by mkhona-nvidia
1s
1s
[QUESTION] Does Megatron-LM support torch 2.10.0 Claude Copy PR to Main #229: Issue comment #4034 (comment) created by longnguyen1802
2s
2s
Create a Protocol for the MLP layer of TransformerLayer Claude Copy PR to Main #228: Issue comment #3435 (comment) created by nschank
1s
1s
build: Bump TE to 2.14 Claude Copy PR to Main #227: Issue comment #4025 (comment) created by svcnvidia-nemo-ci
10s
10s
Refactor setup.py to use get_pybind_include Claude Copy PR to Main #226: Issue comment #3658 (comment) created by svcnvidia-nemo-ci
1s
1s
Refactor setup.py to use get_pybind_include Claude Copy PR to Main #225: Issue comment #3658 (comment) created by svcnvidia-nemo-ci
2s
2s
[QUESTION] Support for Heterogeneous Parallelism in Multimodal Training Claude Copy PR to Main #224: Issue comment #1375 (comment) created by chenhongyu2048
2s
2s
Refactor setup.py to use get_pybind_include Claude Copy PR to Main #223: Issue comment #3658 (comment) created by ko3n1g
1s
1s