Skip to content

Pull requests: ModelTC/LightX2V

New pull request New
27 Open 1,151 Closed
27 Open 1,151 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

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(metax): add ops for compile
#1415 opened Aug 21, 2026 by Watebear Collaborator Loading…
HunyuanImage 3.0 path for the classic four-GPU topology
#1414 opened Aug 21, 2026 by Chernobyllight Contributor Loading…
feat(hunyuan-image3): optimize four-GPU hybrid inference
#1412 opened Aug 21, 2026 by Chernobyllight Contributor Loading…
feat(openpi): add native PyTorch LIBERO integration
#1401 opened Aug 19, 2026 by Chernobyllight Contributor Loading…
Add aggressive MiniMax-H3 VAE decode tile shapes
#1398 opened Aug 19, 2026 by STwangyingrui 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
Add sol_attn for 5090(sm120)
#1385 opened Aug 15, 2026 by wangshankun Collaborator Loading…
fix: transcode non-AAC audio to AAC for MP4 container
#1357 opened Aug 10, 2026 by Tingting-Y Collaborator Loading…
4 tasks done
feat(hunyuan-image3): add multi-micro hybrid parallelism
#1350 opened Aug 10, 2026 by Chernobyllight Contributor Loading…
feat(thor): add FA4 and stride-based split-N NVFP4 support for Wan2.2
#1344 opened Aug 10, 2026 by Michael20070814 Contributor Loading…
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…
lightx2v_train big refactor
#1325 opened Aug 3, 2026 by helloyongyang Contributor Loading…
1
Add FastWAM RoboDojo evaluation adapter
#1311 opened Jul 30, 2026 by q6y6y6 Contributor Loading…
2
Optimize Ulysses FP8 pre-quant row tiling
#1304 opened Jul 28, 2026 by STwangyingrui Contributor Loading…
1
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
update docker
#1225 opened Jul 5, 2026 by helloyongyang Contributor Loading…
12
Optimize offload weight pinning with mmap host registration
#1206 opened Jul 1, 2026 by Fatemanx Contributor Loading…
2
perf: improve HunyuanVideo1.5 I2V runtime and VAE decode controls
#1201 opened Jun 30, 2026 by starrkk Loading…
6
perf: optimize Hunyuan DiT Ulysses and non-attention paths
#1200 opened Jun 30, 2026 by starrkk Loading…
12
perf: optimize ffmpeg video saving
#1198 opened Jun 30, 2026 by starrkk Loading…
6
ProTip! What’s not been updated in a month: updated:<2026-07-23.