Skip to content

Pull requests: ModelTC/LightX2V

New pull request New
31 Open 1,227 Closed
31 Open 1,227 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

Refactor offload
#1501 opened Sep 9, 2026 by Watebear Collaborator Loading…
feat(openpi): add aligned LIBERO training support
#1500 opened Sep 9, 2026 by Chernobyllight Contributor Loading…
feat(minimax_h3): add MPS support, Turbo config and VAE fixes
#1497 opened Sep 8, 2026 by q6y6y6 Contributor Loading…
feat: integrate OpenPI π0.5-LIBERO inference and ROS2 evaluation
#1491 opened Sep 7, 2026 by Chernobyllight Contributor Loading…
fix(minimax-h3): align Ref2AV output
#1490 opened Sep 6, 2026 by Watebear Collaborator Loading…
9
Accelerate The Inference Speed Of Wan2.2 On NVIDIA Thor T5000
#1489 opened Sep 6, 2026 by Michael20070814 Contributor Loading…
feat(musa): add sparse attention settings
#1488 opened Sep 5, 2026 by Watebear Collaborator Loading…
fix(audio_io): remove unreachable duplicate except in torchcodec probe
#1487 opened Sep 5, 2026 by Anai-Guo Loading…
perf(kernel): optimize SM120 FP8 GEMM with FP16 accumulation
#1479 opened Sep 4, 2026 by STwangyingrui Contributor Loading…
feat: integrate OpenPI π0.5-LIBERO inference and ROS2 evaluation
#1476 opened Sep 4, 2026 by Chernobyllight Contributor Loading…
feat: integrate OpenPI LIBERO inference
#1473 opened Sep 3, 2026 by Chernobyllight Contributor Loading…
Fix pre-weights ignoring dit_quant_scheme (fp8 dtype mismatch)
#1470 opened Sep 1, 2026 by li-lizhe Loading…
1
fix(minimax_h3): enable torch.compile for INT8 ConvRot
#1464 opened Aug 31, 2026 by STwangyingrui Contributor Loading…
fix(wan-audio): restore the phase argument at the three multi-line _apply_rope_with_cache_range call sites
#1460 opened Aug 31, 2026 by Anai-Guo Loading…
Add OpenPI pi0.5 support and LIBERO evaluation
#1440 opened Aug 26, 2026 by Chernobyllight Contributor Loading…
Replace deprecated v2.ToTensor() with v2.ToImage() + v2.ToDtype(scale=True)
#1419 opened Aug 23, 2026 by xyf5432 Loading…
1
fix(qwen_image): forward multimodal token type ids
#1417 opened Aug 22, 2026 by llc1123 Loading…
1
feat(openpi): add native PyTorch LIBERO integration
#1401 opened Aug 19, 2026 by Chernobyllight Contributor Loading…
feat(minimax-h3): support AdaLN curve checkpoints
#1391 opened Aug 17, 2026 by Leonccaa Loading…
3
fix(minimax-h3): materialize TP column shards
#1390 opened Aug 17, 2026 by Leonccaa Loading…
2
fix: transcode non-AAC audio to AAC for MP4 container
#1357 opened Aug 10, 2026 by Tingting-Y Collaborator Loading…
4 tasks done
feat(flux2/metax): optimize distributed encoder and decoder，tp8 pipeline 53.99s-> 20.23s on metax platform
#1342 opened Aug 8, 2026 by Watebear Collaborator Loading…
feat(qwen-image/metax): optimize distributed offload，metax cfg2_sp4 pipeline cost 52.76s-> 21.68s
#1341 opened Aug 8, 2026 by Watebear Collaborator Loading…
Inference Sparse Optimization of Wan2.2 on Ascend Platform
#1301 opened Jul 27, 2026 by Watebear Collaborator Loading…
1
feat: support pipefusion for flux2
#1268 opened Jul 19, 2026 by Qin-sx Loading…
11
ProTip! Add no:assignee to see everything that’s not assigned.