Skip to content

Pull requests: NVIDIA/Megatron-LM

New pull request New
194 Open 294 Closed
194 Open 294 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
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Add DistTrain, Allow Encoder to Have Different DP Size
#1605 opened May 30, 2025 by zidanehuang001 Loading…
add node_rank argument for example scripts
#1604 opened May 30, 2025 by xylllllllll Loading…
CLIPViTModel support SP and CP
#1600 opened May 28, 2025 by Thaurun Loading…
Support Multiple Input Formats for checkpoint
#1599 opened May 28, 2025 by Thaurun Loading…
bugfix: cross_entropy inplace operations may cause backward error
#1594 opened May 24, 2025 by ChangWeiming Loading…
support for qwen2.5vl window attention
#1591 opened May 22, 2025 by Agoniii Loading…
fix bug: the loss of aux_loss and mtp will be tracked twice
#1585 opened May 18, 2025 by hyleepp Loading…
[fix] Fix get_transformer_layer_offset incorrect when virtual pipeline is enabled and num_layers_in_last_pipeline_stage is set but num_layers_in_first_pipeline_stage is not
#1583 opened May 17, 2025 by yqy3214 Loading…
[Draft ]FP8 param support for MXFP8
#1581 opened May 14, 2025 by WanZzzzzz • Draft
use multiple yaml files to avoid passing annoying model configs from cmd lines
#1579 opened May 14, 2025 by nrailg Loading…
1
The phrase "need to want to" is grammatically incorrect
#1574 opened May 13, 2025 by A-transformer Loading…
1
fix: count_zeros protection in chained optimizer
#1569 opened May 8, 2025 by clumsy Loading…
2
Fix incorrect softmax_factor calculation in MLA
#1562 opened May 2, 2025 by HowardZorn Loading…
1
[bugfix] fix the bug that loss: 0 will not be printed
#1555 opened Apr 28, 2025 by leisuzz Loading…
Fix: training arguments print format
#1552 opened Apr 24, 2025 by vicoooo26 Loading…
Fp8 LM-head
#1551 opened Apr 22, 2025 by dhia680 Loading…
param_copy_back_gpu_hook should sync to h2d stream
#1543 opened Apr 16, 2025 by ariverhorse Loading…
3
Fix parameter error in text_generation_server.py file
#1542 opened Apr 16, 2025 by xichengpro Loading…
2
lora offload
#1540 opened Apr 15, 2025 by sanandaraj5597 Loading…
Lora offload
#1539 opened Apr 15, 2025 by sanandaraj5597 Loading…
Swiglu fusion
#1538 opened Apr 15, 2025 by rachitgarg91 Loading…
Add fused swiglu for MLP
#1536 opened Apr 15, 2025 by michal2409 Loading…
[BUGFIX] Save dist_checkpointing metadata on all nodes for multi-node training
#1531 opened Apr 13, 2025 by Pranaykarvi Loading…
1
1
Fix AttributeError in MultiTokenPredictionLayer
#1529 opened Apr 12, 2025 by shenyunhang Loading…
added fix to avoid overflow with new numpy casting behaviour (Issue: #1519)
#1520 opened Apr 4, 2025 by Apsod Loading…
ProTip! Filter pull requests by the default branch with base:main.