Search results

[model] video data SP handling inconsistence between Qwen3VL and Qwen2.5VL

Status: Open.

#369 In ByteDance-Seed/VeOmni;

· Yangr116 opened on Jan 5, 2026

[RFC] Introducing Ring-Based Context Parallelism

Status: Open.

#363 In ByteDance-Seed/VeOmni;

· TKONIY opened on Jan 4, 2026

Question: Tensor Parallel and Pipeline Parallel not supported yet?

Status: Open.

#344 In ByteDance-Seed/VeOmni;

· Hyubo opened on Dec 29, 2025

Trainer: add text trainer for veomni

Status: Open.

#326 In ByteDance-Seed/VeOmni;

· Fazziekey opened on Dec 24, 2025

[Bug] First-step Loss NaN when enabling deterministic mode via msprobe.seed_all in Wan2.1 training

Status: Open.

#306 In ByteDance-Seed/VeOmni;

· zbian99 opened on Dec 18, 2025

[Feature Request] Implement ViT rebalance

ci

enhancementNew feature or requestNew feature or request

fix

roadmap

Status: Open.

#280 In ByteDance-Seed/VeOmni;

· piyifan123 opened on Dec 11, 2025

[Ascend][RoadMap]: NPU support for VeOmni v0.2

ascendeverything about Ascend supporteverything about Ascend support

ci

ckptCheckpoint related.Checkpoint related.

docImprovements or additions to documentationImprovements or additions to documentation

enhancementNew feature or requestNew feature or request

hf_v5Related for transformers v5Related for transformers v5

roadmap

Status: Open.

#271 In ByteDance-Seed/VeOmni;

· onehaitao opened on Dec 9, 2025

[Roadmap] VeOmni v0.2 Release Plan

ci

docImprovements or additions to documentationImprovements or additions to documentation

enhancementNew feature or requestNew feature or request

example

hf_v5Related for transformers v5Related for transformers v5

roadmap

Status: Open.

#268 In ByteDance-Seed/VeOmni;

· FoolPlayer opened on Dec 9, 2025

[Future deprecation] SP: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9

bugSomething isn't workingSomething isn't working

Status: Open.

#267 In ByteDance-Seed/VeOmni;

· piyifan123 opened on Dec 8, 2025

RL post training for omni-modality models

enhancementNew feature or requestNew feature or request

roadmap

Status: Open.

#262 In ByteDance-Seed/VeOmni;

· hsliuustc0106 opened on Dec 7, 2025

[Feature Request] Add support for validation dataset during training

enhancementNew feature or requestNew feature or request

Status: Open.

#247 In ByteDance-Seed/VeOmni;

· piyifan123 opened on Dec 2, 2025

support non-intrusive pipeline parallelism

Status: Open.

#238 In ByteDance-Seed/VeOmni;

· shifangx opened on Dec 1, 2025