Commit Graph
53 Commits
Author SHA1 Message Date
rlsu9 bd0ef67632 update readme 2024-12-14 20:33:12 +00:00
rlsu9 61ab8221ef add data propress readme 2024-12-14 20:25:10 +00:00
Your Name d9d7714195 update 2024-12-14 12:24:40 -08:00
Your Name 7e7e9cbbcf update 2024-12-13 21:00:27 -08:00
rlsu9 566a6cda53 update 2024-12-13 20:22:06 +00:00
Peiyuan Zhang 1bf20f30a7 scripts 2024-12-13 20:06:17 +00:00
Zhang Peiyuan 85639d1101 [feat] add hunyuan adv (#79) 2024-12-13 11:52:57 -08:00
Zhang Peiyuanandforeverpiano 6ab2263f3a [Feat] Add HunyuanVideo (#78)
Co-authored-by: foreverpiano <pianoqwz@qq.com>
2024-12-12 14:14:09 -08:00
Zhang Peiyuanandrlsu9 b421c2e183 Cleanup (#77)
Co-authored-by: rlsu9 <r3su@ucsd.edu>
2024-12-12 14:04:02 -08:00
Zhang Peiyuan de1e8d868e Cleanup (#75) 2024-12-06 20:56:06 -08:00
Brian ChenandBrianChenn1129 98b92be25e add web demo (#73)
Co-authored-by: BrianChenn1129 <yonqgich@umich>
2024-12-06 09:47:22 -08:00
Zhang Peiyuanandforeverpiano 8d41d505fe [cleanup] (#72)
Co-authored-by: foreverpiano <pianoqwz@gmail.com>
2024-12-05 21:02:10 -08:00
Brian Chen 8cfdf58a17 [Feat] HF Lora
yongqich@umich.edu
2024-12-05 18:50:53 -08:00
Zhang Peiyuan cf15594055 [Feat & Debug] fix uncond; multi guidance validaiton; multiphase schedule; linear range (#64)
typo

update

update

typo

[Debug] Typo (#65)

debug gradient accumulation loss

update gitignore

typo

typo

update scripts

update experiment 10

new script

typo

update

update

update

update

wandb offline and dir

runlong

update

update

update

update

update

update

update

fix finetune code bug

update

update

update

add l2

update
2024-11-30 16:43:24 -08:00
Zhang Peiyuan ce95c2df29 [Feat] EMA Distill; Distributed validation (#63)
update
2024-11-29 21:46:40 -08:00
Zhang Peiyuan d417e4c7c4 [Feat] Refactor GAN; State saving & Resume; Experiments script (#59)
typo

add upload command

add

add ema_transform

add ema transformer

remove harcode

ok

ema

remove hardcode

update env

update script; no sp

revert to sp=4, sp bs=2, full shard

add aws efo env

distributed validation

readme

distributed validation

add gupload

linear range; fix uncond; multi guidance validation

add script
2024-11-28 21:09:31 -08:00
Zhang Peiyuan 2a70d05b4f [Feat] Training precision (#57) 2024-11-27 16:25:07 -08:00
Zhang Peiyuan 03187fd83a [Feat][Debug] linear quadratic distill; HF precision bug (#56) 2024-11-27 15:13:20 -08:00
Zhang Peiyuan 8a128ad815 [Fix] Squeeze bug (#55) 2024-11-26 13:29:03 -08:00
Zhang Peiyuan 13f665e455 [Feat] PCM Distill; Refactor FM logit to be compatible with all SD3/Flux scheduler. (#54) 2024-11-26 12:26:24 -08:00
rlsu9andRunlong 94ba0ab6ea [feat]: Add batchy data preprocess (#53)
Co-authored-by:Runlong <rlsu9@ucsd.edu>
2024-11-25 21:52:36 -08:00
rlsu9andrunlong 5a5d0ef1a0 [feat]: Add Image-Video Mixture training to main repo (#50)
Co-authored-by: runlong <r3su@ucsd.edu>
2024-11-24 14:36:45 -08:00
Zhang Peiyuan 45e4adca4d [Fix]: Resolve config bug and seed (#51)
Co-authored-by: Peiyuan Zhang <a1286225768@gmail.com>
2024-11-15 20:29:48 -08:00
Zhang Peiyuanandforeverpiano 7106eadffc [feat]: Add LADD (#45)
Co-authored-by: Peiyuan Zhang <a1286225768@gmail.com>
Co-authored-by: foreverpiano <pianoqwz@gmail.com>
2024-11-15 19:28:41 -08:00
Yongqi ChenandPeiyuan Zhang 2f3a8661bf [feat]: add lr scheduler; precision bug fix; add naive dataloader resume (#49)
Co-authored-by: Peiyuan Zhang <a1286225768@gmail.com>
2024-11-14 22:57:38 -05:00
Yongqi ChenandPeiyuan Zhang 6d0082c1a9 [Feat] Lora resume (#48)
Co-authored-by: Peiyuan Zhang <a1286225768@gmail.com>
2024-11-13 01:30:58 -05:00
Zhang PeiyuanandYongqi Chen 3d9189571a [feat]: Add lora (#47)
Co-authored-by: Yongqi Chen <144848849+BrianChen1129@users.noreply.github.com>
2024-11-11 13:12:14 -08:00
Zhang Peiyuan b042e321a1 [Refactor] Switch to FSDP (#42) 2024-11-09 15:46:25 -08:00
rlsu9 44bda9f8a3 [feat]: Add adaptive fps dataloader and remove redundant code (#41)
No checkout mochi
2024-11-06 19:52:48 -08:00
rlsu9 035ba5f5cf [feat]: Add vae encoder embedded generator to main (#30) 2024-11-06 11:58:15 -08:00
Zhang Peiyuan 52ba538e8e [feat]: Add validation logging with SP (#36) 2024-11-05 14:46:20 -08:00
Zhang Peiyuanandforeverpiano f0bc297260 [Feat] Sequence Parallel (#31)
Co-authored-by: foreverpiano <pianoqwz@gmail.com>
2024-11-05 08:13:01 -08:00
rlsu9 7413b1dd5f [feat]: update data preprocess 2024-10-31 00:13:05 +00:00
Peiyuan Zhang 8ec82cbc37 Generate synthetic dataset
Delete Open Sora Plan Modeling

amend log validation

Change name to fast video

Remove OSP modeling

clean up name changing

remove files

 Deleted unnecessary files

commit first

commit first

training

ok

Update generate_synthetic.sh and deepspeed_zero2_config.yaml

14

debug

overfitting ....

debugging ..

Debug successful!

Add zero3

OK

update optimizer

load

small bug

random seed args

sp enable & still has bug

rename

update inference sp code / can run / still has bug / don't output normal mp4

switch to deepspeed dummyoptim

fix some bugs; still output green

SP inference done!

fix typos in readme and latent dataset debug file
2024-10-28 00:44:02 +00:00
Peiyuan Zhang 54e74aec3f Mochi Inferenfce & Diffusers
typo

Original pipeline
2024-10-27 22:57:14 +00:00
Peiyuan Zhang c1c276b616 Refactor OpenSora sample_t2v.py and update download_hf 2024-10-26 21:59:22 +00:00
Peiyuan Zhang 8e5fa4d383 remove vae loss 2024-10-26 19:20:04 +00:00
Peiyuan Zhang 06ff1c912f Remove files in causalvae 2024-10-26 19:06:46 +00:00
Peiyuan Zhang 857f5df51b normalize 255; vae reconstruct 2024-10-26 18:57:07 +00:00
Peiyuan Zhang 18c5ca131d Add mochi download & Change output dir 2024-10-26 18:08:50 +00:00
Peiyuan Zhang c16625242e Delete merge_data.txt 2024-10-26 18:00:16 +00:00
Peiyuan Zhang fcc45701c9 Remove unused adaptor files 2024-10-26 17:59:04 +00:00
Peiyuan Zhang 6c87e003aa Remove unused arguments in train_t2v_diffusers.py 2024-10-26 17:54:43 +00:00
Peiyuan Zhang 23181aac96 Update T5Base 2024-10-26 17:53:51 +00:00
Peiyuan Zhang 26b65a8baf Update PyTorch installation command 2024-10-26 17:36:23 +00:00
Peiyuan Zhang 25498a9d85 Update model path and cache directory 2024-10-25 10:21:31 +00:00
Peiyuan Zhang afb3ac9fc1 Update t2v_debug_multi.sh with video_length_tolerance_range and dataloader_num_workers 2024-10-25 09:58:41 +00:00
Peiyuan Zhang 3d385600e9 update version 2024-10-25 09:22:16 +00:00
Peiyuan Zhang 5680039dbe Add environment setup and training instructions to README.md
Update dependencies; Setup code for debugging

Delete unused files and code

Remove all npu code

Update torchvision imports

Update EMA model and t2v_debug.sh script

Delete npu related stuff and remove inpaint module

Remove compress kv

Fix warning with dataset handling and model loading

Update PyTorch index URLs and video length tolerance range

Remove UDIT and inpaint

fix typo for dataset download

 include pretrained open-sora

Add pretrained model for OpenSoraT2V-ROPE-L

Update max height and width for video processing
2024-10-24 17:47:34 +00:00
jzhang38 2c1eb3b0e2 Remove NPU related code and update training process 2024-10-24 02:36:18 +00:00