Commit Graph
257 Commits
Author SHA1 Message Date
SolitaryThinker ea8c154176 ckpt 2025-06-28 14:56:21 -07:00
Kevin Lin 580d6dfe1f [CI] Add tests to Modal (#562) 2025-06-28 14:02:16 -05:00
Wenxuan Tan 344e43006a [CI] Fix SSIM and transformers CI (#564) 2025-06-28 00:26:20 -05:00
Wenxuan Tan c5155b256e [Feature] Load weights from distributed (#470) 2025-06-27 22:52:40 -05:00
William Lin e005c7f3ac [Docs] [Training] add readme for example training (#563) 2025-06-27 14:50:42 -05:00
Yongqi Chen ff5a79ef60 [Feature][Inference] Add VSA inference script (#561) 2025-06-27 02:19:23 -05:00
William Lin ab01dc4ba5 [Feature] [Training] Add i2v training (#559) 2025-06-27 01:56:50 -05:00
William Lin 285a950c1b [CI] fix vae and ssim tests (#557) 2025-06-26 23:53:01 -05:00
William Lin 46a0a85d85 [Training] Fixes SP for training; Improve Datasets and schema (#555) 2025-06-26 21:13:28 -05:00
Yongqi Chen 4aeabbc629 [Feature][Training] Add cfg rate for dataset loader (#556) 2025-06-26 18:22:37 -04:00
Wenxuan Tan 949bb5c835 [CI] Fix CI checks (#553) 2025-06-25 14:07:51 -05:00
Wenxuan Tan aab74c1271 [Kernel] Remove all syncs from STA & VSA kernels (#517) 2025-06-23 13:13:09 -07:00
Yongqi Chen f89d86944f [Feature][Training]Add diffusers format checkpoint saving for inference (#542) 2025-06-22 01:23:41 -04:00
William Lin 8741d204a5 [Training] Refactor and improve validation datasets (#539) 2025-06-21 17:58:35 -07:00
Wenxuan Tan cdc85f58a8 [chore] Bump torch to 2.7.1 to support Blackwell (#483) 2025-06-20 22:10:56 -07:00
William Lin 0262d2f089 [misc] [training] Reorganize training pipeline (#533) 2025-06-20 20:42:25 -07:00
William Lin 62c0343465 [bugfix] [VSA] Fix layernorm type for VSA Wan2.1 TransformerBlock (#534) 2025-06-20 00:24:51 -07:00
William Lin 1e1a023fb0 [bugfix] Fix stage validator for multi text encoder models (#535) 2025-06-19 22:49:16 -07:00
William Lin 1d2517ad8e [misc] Remove gradient checking code (#532) 2025-06-18 23:29:25 -07:00
William Lin d41186cb4a [Feat] Add Stage input and output verification (#523) 2025-06-18 23:29:11 -07:00
78e0c7eec9 Specify cu128 Pytorch installation (#530)
Co-authored-by: Edenzzzz <wtan45@wisc.edu>
Co-authored-by: Wenxuan Tan <wenxuan.tan@wisc.edu>
2025-06-18 20:02:50 -05:00
Wenxuan Tan 1c41a94b62 [Refactor] Move dict_to_3d_list under utils (#507) 2025-06-18 13:34:37 -07:00
Yongqi Chen 2e66aafe20 [Bugfix][Readme]Fix readme website bugs and add VSA finetune docs (#531) 2025-06-17 22:48:29 -07:00
Yongqi ChenandWill Lin 55074bda76 [CI] Add STA-inference/VSA-training test (#527)
Co-authored-by: Will Lin <wlsaidhi@gmail.com>
2025-06-17 21:13:06 -07:00
William Lin de65bec2b7 [Ci] add sta and vsa install to docker image (#528) 2025-06-17 18:09:48 -07:00
Yongqi Chen 7664dd0de3 [Bugfix][Inference]Fix envs.attn_backend (#525) 2025-06-17 18:38:06 -05:00
William Linandkevin314 019a88ced4 [CI][bugfix] Use new 3.12 docker image (#526)
Co-authored-by: kevin314 <kevin.lin.cs1@gmail.com>
2025-06-17 15:37:08 -07:00
Kevin Lin 72de11abcc [CI] Add current PR test workflow to Buildkite/Modal (#512) 2025-06-17 13:29:22 -07:00
Kevin Lin d71a4ebffc [CI] Update Docker image to flash-attn 2.8.0 / CUDA 12.8 (#524) 2025-06-16 17:48:23 -07:00
William Lin 1089ab43bf [bugfix] [Training] use diffusers fp32layernorm for wan2.1 (#490) 2025-06-15 22:45:48 -07:00
William Lin 97d4b984c9 [misc] [ci] fix e2e preprocess+training data path (#521) 2025-06-14 22:37:51 -07:00
Wenxuan Tan 2a8953d74d [Refactor] Fix attn backend selection not correctly setting env variable (#516) 2025-06-15 00:04:54 -05:00
Yongqi Chen 8801b10da7 [Bugfix][Preprocess]fix mini dataset name (#520) 2025-06-14 22:03:22 -07:00
William Lin 6b413f2ec4 [CI] [Training] drop negative prompt in validation dataset and CI test for preprocess + training overfit (#519) 2025-06-14 18:50:17 -07:00
Yongqi Chen 28b72694aa [Feature][Preprocess]Add Readme doc for preprocess (#518) 2025-06-14 20:41:13 -04:00
Yongqi Chen 4afb0cfe4f [Feature][Training]vsa for t2v training ready (#513) 2025-06-14 01:08:00 -04:00
Zhang Peiyuan 3eec1281cf [misc] Fix preprocessing and dataloader extra padding (#514) 2025-06-13 15:15:33 -07:00
Wenxuan Tan 0660489e38 [CI] Restrict training CI to v1 (#508) 2025-06-12 15:26:05 -07:00
Zhang Peiyuan dd871a17bf fix logging (#509) 2025-06-12 15:24:12 -07:00
dc11529862 [Refactor][Configurations] clean config orgnization (#505)
Co-authored-by: Peiyuan Zhang <a1286225768@gmail.com>
Co-authored-by: Will Lin <wlsaidhi@gmail.com>
2025-06-12 13:27:08 -07:00
Zhang Peiyuan ffabf85e31 [feat] Add parquet iterable dataset. (#506) 2025-06-12 04:30:56 -04:00
William Lin c0026ca5ba [CI] [Training] Initial e2e small training test (#504) 2025-06-11 13:53:36 -07:00
Zhang Peiyuan 0f2bbe71ac [misc] rename dp_size to hdsp_replicate_dim (#491) 2025-06-10 16:36:56 -07:00
Yongqi ChenandJerryZhou54 2a46902ecb [Feature][VSA]Update STA publish workflow (#498)
Co-authored-by: JerryZhou54 <zhouw.jerry2017@outlook.com>
2025-06-10 19:33:34 -04:00
Zhang Peiyuan 66012d3a4c [Feat][Dataloader] 1/n Refactor parquet map-style dataloader (#492) 2025-06-10 16:00:13 -07:00
William Lin f666b9de41 [misc] Add missing license headers (#499) 2025-06-10 14:25:32 -07:00
Yongqi Chen 7e3c073b55 [Feature] Adding VSA inference (#478) 2025-06-10 16:03:53 -04:00
Wei Zhou a6aa21bd07 [bugfix][Cli Inference] Resolve runtime errors when running fastvideo generate (#495) 2025-06-10 02:50:21 -04:00
Wenxuan Tan 6519b57aab [chore] Fix main pre-commit CI failure (#494) 2025-06-10 00:54:37 -05:00
Wei Zhou 675aea6ece [bugfix][Cli Inference] Resolve runtime errors when running fastvideo generate (#493) 2025-06-09 19:49:34 -07:00