Skip to content

Pull requests: ISEEKYAN/mbridge

New pull request New
10 Open 80 Closed
10 Open 80 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

[feat] Optimize export_weights with bucketed all-gather
#116 opened Apr 3, 2026 by Begunner Loading…
[fix] handle FSDP DTensor in broadcast_from_megatron_pp
#113 opened Apr 2, 2026 by yxs Loading…
1
[fix] Fix qwen3.5 MoE vp_stage propagation under VPP
#112 opened Apr 1, 2026 by none0663 Loading…
try fit in new transformers moe
#110 opened Mar 30, 2026 by ETOgaosion Collaborator Loading…
1
fix(qwen3_vl): fix IndexPutBackward0 invalid gradient shape in vision…
#100 opened Mar 23, 2026 by alexqdh Loading…
support mtp layer support for qwen3.5 series models
#98 opened Mar 20, 2026 by zpltys Loading…
3
feat: Support quantization before weight resharding
#92 opened Mar 8, 2026 by hy2826 Loading…
feat: support LongCat-Flash-Chat Model
#82 opened Feb 14, 2026 by VegetaPn Loading…
feat: online load/save huggingface model weights for Megatron-FSDP
#80 opened Feb 12, 2026 by conver334 Loading…
fix: get_tensor_model_parallel_group_if_none is missing for mcore v0.12.1
#53 opened Nov 13, 2025 by quancs Contributor Loading…
ProTip! Follow long discussions with comments:>50.