-
Notifications
You must be signed in to change notification settings - Fork 185
Pull requests: ByteDance-Seed/VeOmni
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
[ops, model] refactor: register moe_experts in OpSpec registry
#718
opened May 1, 2026 by
TimYangst
Collaborator
Loading…
4 tasks done
[BREAKING][ops, model] feat: GPU-optimal ops defaults + strict NPU validation
ascend
everything about Ascend support
[model] fix: qwen3_moe router double-softmax and idempotent _init_weights wrap
fix
#715
opened Apr 30, 2026 by
TimYangst
Collaborator
Loading…
1 of 4 tasks
[ops, model] feat: dispatch Qwen3.5 linear-attention kernels via OpSlot
#714
opened Apr 30, 2026 by
TimYangst
Collaborator
Loading…
4 tasks done
[ops, model] feat: log post-build kernel selection summary
#712
opened Apr 29, 2026 by
TimYangst
Collaborator
Loading…
3 of 4 tasks
add NPU support for Qwen3/VL/MoE, Qwen3.5/MoE
ascend
everything about Ascend support
#710
opened Apr 29, 2026 by
yanghw116
Loading…
[model] fix: read tie_word_embeddings from inner text/decoder config
fix
#693
opened Apr 26, 2026 by
TimYangst
Collaborator
Loading…
4 of 5 tasks
[docker] fix: datasets version fix
docker
fix
#691
opened Apr 25, 2026 by
phdddd
Contributor
Loading…
[ci] refactor: consolidate workflows and reorganize tests
ci
#687
opened Apr 23, 2026 by
FoolPlayer
Collaborator
Loading…
5 tasks done
[parallel, perf] fix: sp gather && optimize: input embeds fusing
fix
#681
opened Apr 21, 2026 by
JorgenWan
Collaborator
Loading…
[ci] fix: update npu ci workflow yml
ascend
everything about Ascend support
ci
fix
#657
opened Apr 15, 2026 by
FoolPlayer
Collaborator
Loading…
[parallel] fix: NPU Hang/Deadlock during DTensor parameter loading in FSDP2
ascend
everything about Ascend support
fix
#642
opened Apr 10, 2026 by
First-Frost-code
Loading…
[docker] feat: update to torch2.10 + cu130
docker
#629
opened Apr 2, 2026 by
FoolPlayer
Collaborator
Loading…
[model]feat: Qwen3.5 is compatible with NPU
ascend
everything about Ascend support
#600
opened Mar 23, 2026 by
wang-hua-2019
Contributor
Loading…
[model] feat: [transformers-v5] Introduce new registration based kernel replacement.
#569
opened Mar 16, 2026 by
piyifan123
Collaborator
Loading…
[parallel] feat: Vision Data Parallel — O(1) communication alternative to patch-level SP
#505
opened Feb 24, 2026 by
aoshen524
Loading…
2 of 3 tasks
[models] chore: Change transformers v5 support for qwen3_moe to use HF v5 style expert weight layout and add a converter impl.
hf_v5
Related for transformers v5
misc
Every misc
#500
opened Feb 24, 2026 by
piyifan123
Collaborator
•
Draft
[task] feat: support sequence classification tasks
#470
opened Feb 11, 2026 by
yiwzhao
Collaborator
Loading…
[model] fix: Incorrect usage of the 'check_model_inputs' decorator
fix
#457
opened Feb 5, 2026 by
HSYZhang
Contributor
Loading…
6 tasks done
Previous Next
ProTip!
Add no:assignee to see everything that’s not assigned.