Sana/diffusion · sana/Sana

GGitHubfeat: add SANA-Video 2.0 four-step preview (#471)

文件	最后提交记录	最后更新时间
data	SANA Streaming Training Code (#421) * bidirectional training release * release sana streaming training code * Fix pre-commit formatting * test: cover streaming paired data and chunks * fix typo, remove unused code, add re-try boundary * pre-commit * test: cover bidirectional streaming training * fix: bound V2V batch retries locally * fix reverse generation overlap * add online validation * test: cover reverse overlap and V2V validation --------- Co-authored-by: junsong <cjs1020440147@icloud.com>	2 个月前
distributed	Support stage-1 CP training, SANA-WM Benchmark & fp8/fp4 quantized streaming inference (#404) * feat(sana-wm): support stage-1 CP training * docs(sana-wm): expose chunk-causal training entry * style: run pre-commit on sana-wm training changes * docs: add SANA-WM benchmark guide * SANA-WM streaming: per-component fp8/fp4 quantized inference Add --stage1_precision/--refiner_precision (bf16/fp8/fp4) with per-block fp8_autocast scoping that preserves action-following, fp8 recipe selector for stage-1 + refiner, default Transformer Engine install, and benchmark docs. * Polish Sana-WM PR docs and formatting * Use end-to-end Sana-WM inference CI * Add SANA-WM stage1 training CI * Run SANA-WM stage1 smoke in video CI * Keep SANA-WM training in dedicated CI job --------- Co-authored-by: junsong <cjs1020440147@icloud.com>	3 个月前
guiders	LongSANA: support SANA-Video + [LongLive](https://github.com/NVlabs/LongLive) (#341) * longsana training and inference code release * [CI-Lint] Fix code style issues with pre-commit 51181934710b656bf0792c4133cd44c32f4040c0 * update convert script; * update pre-commit version; * ci update; * revoke keys and update related code; * 1. add train val txt samples files for longsana; 2. clean yaml file for better vis; * pre-commit; * 1. add some TODOs where path need to be updated; 2. change all Chinese into English; 3. remove unused commands and comments; (stop before diffusion/longsana/trainer/longsana_trainer.py) * update data path and longsana docs * change `asset/longsana` to `data/longsana`, since the ./asset is not in gitignore, but ./data is. * 1. add longsana CI training test; 2. remove some unsed code; 3. add max_iters to the config in training scripts; * strict loss print to 4 digits; * 1. keep cleaning unsed code; 2. add longsana CI inference test; 3. add diffusers pipeline in md; * update readme; Co-authored-by: Shuai <a18310626323@163.com> * update model zoo * remove all unsed functions; * update readme; Co-authored-by: LoveSy <shana@zju.edu.cn> Co-authored-by: Enze Xie <johnny_ez@163.com> * fix some TODOs * pre-commit; * pre-commit; * ci update; * fix bug; * fix bugs; * fix bug; --------- Co-authored-by: GitHub Action <action@github.com> Co-authored-by: junsong <cjs1020440147@icloud.com> Co-authored-by: Shuai <a18310626323@163.com> Co-authored-by: LoveSy <shana@zju.edu.cn> Co-authored-by: Enze Xie <johnny_ez@163.com>	9 个月前
longsana	SANA Streaming Training Code (#421) * bidirectional training release * release sana streaming training code * Fix pre-commit formatting * test: cover streaming paired data and chunks * fix typo, remove unused code, add re-try boundary * pre-commit * test: cover bidirectional streaming training * fix: bound V2V batch retries locally * fix reverse generation overlap * add online validation * test: cover reverse overlap and V2V validation --------- Co-authored-by: junsong <cjs1020440147@icloud.com>	2 个月前
model	feat: open-source Sana-Video 2.0 (#439) * feat: add Sana-Video 2.0 release code * refactor: reuse shared Sana-Video workflows * docs: release SANA-Video 2.0 5B checkpoint * fix: validate SANA-Video 2.0 release inference * docs: update SANA-Video 2.0 demo seed * fix: align SANA-Video 2.0 linear attention * docs: link SANA-Video 2.0 live demo	1 个月前
post_training	[codex review] fix all chinese comment to english	3 个月前
refiner	Support stage-1 CP training, SANA-WM Benchmark & fp8/fp4 quantized streaming inference (#404) * feat(sana-wm): support stage-1 CP training * docs(sana-wm): expose chunk-causal training entry * style: run pre-commit on sana-wm training changes * docs: add SANA-WM benchmark guide * SANA-WM streaming: per-component fp8/fp4 quantized inference Add --stage1_precision/--refiner_precision (bf16/fp8/fp4) with per-block fp8_autocast scoping that preserves action-following, fp8 recipe selector for stage-1 + refiner, default Transformer Engine install, and benchmark docs. * Polish Sana-WM PR docs and formatting * Use end-to-end Sana-WM inference CI * Add SANA-WM stage1 training CI * Run SANA-WM stage1 smoke in video CI * Keep SANA-WM training in dedicated CI job --------- Co-authored-by: junsong <cjs1020440147@icloud.com>	3 个月前
scheduler	feat: add SANA-Video 2.0 four-step preview (#471) * feat: add SANA-Video 2.0 four-step preview * docs: add verified four-step preview * docs: align four-step preview controls Co-authored-by: Shuchen Xue <74777491+scxue@users.noreply.github.com> Co-authored-by: Yuyang Zhao <43061147+HeliosZhao@users.noreply.github.com> Co-authored-by: Haozhe Liu <53887227+HaozheLiu-ST@users.noreply.github.com> Co-authored-by: Yitong Li <135998170+lyttttt3333@users.noreply.github.com> Co-authored-by: Enze Xie <20984575+xieenze@users.noreply.github.com> Co-authored-by: Jincheng Yu <19924374+loveSy@users.noreply.github.com> * docs: announce four-step preview * docs: simplify four-step preview news --------- Co-authored-by: Shuchen Xue <74777491+scxue@users.noreply.github.com> Co-authored-by: Yuyang Zhao <43061147+HeliosZhao@users.noreply.github.com> Co-authored-by: Haozhe Liu <53887227+HaozheLiu-ST@users.noreply.github.com> Co-authored-by: Yitong Li <135998170+lyttttt3333@users.noreply.github.com> Co-authored-by: Enze Xie <20984575+xieenze@users.noreply.github.com> Co-authored-by: Jincheng Yu <19924374+loveSy@users.noreply.github.com>	18 天前
utils	feat: open-source Sana-Video 2.0 (#439) * feat: add Sana-Video 2.0 release code * refactor: reuse shared Sana-Video workflows * docs: release SANA-Video 2.0 5B checkpoint * fix: validate SANA-Video 2.0 release inference * docs: update SANA-Video 2.0 demo seed * fix: align SANA-Video 2.0 linear attention * docs: link SANA-Video 2.0 live demo	1 个月前
__init__.py	feat: add SANA-Video 2.0 four-step preview (#471) * feat: add SANA-Video 2.0 four-step preview * docs: add verified four-step preview * docs: align four-step preview controls Co-authored-by: Shuchen Xue <74777491+scxue@users.noreply.github.com> Co-authored-by: Yuyang Zhao <43061147+HeliosZhao@users.noreply.github.com> Co-authored-by: Haozhe Liu <53887227+HaozheLiu-ST@users.noreply.github.com> Co-authored-by: Yitong Li <135998170+lyttttt3333@users.noreply.github.com> Co-authored-by: Enze Xie <20984575+xieenze@users.noreply.github.com> Co-authored-by: Jincheng Yu <19924374+loveSy@users.noreply.github.com> * docs: announce four-step preview * docs: simplify four-step preview news --------- Co-authored-by: Shuchen Xue <74777491+scxue@users.noreply.github.com> Co-authored-by: Yuyang Zhao <43061147+HeliosZhao@users.noreply.github.com> Co-authored-by: Haozhe Liu <53887227+HaozheLiu-ST@users.noreply.github.com> Co-authored-by: Yitong Li <135998170+lyttttt3333@users.noreply.github.com> Co-authored-by: Enze Xie <20984575+xieenze@users.noreply.github.com> Co-authored-by: Jincheng Yu <19924374+loveSy@users.noreply.github.com>	18 天前