Commit Graph

  • 35383e2db3
    fix potential bug of i2v Zheng Guang Cong 2025-01-11 17:08:25 +08:00
  • 7dc8516bcb add comment as #653 zR 2025-01-11 12:53:32 +08:00
  • 2f275e82b5 Merge remote-tracking branch 'upstream/CogVideoX_dev' into dev OleehyO 2025-01-11 02:16:09 +00:00
  • caa24bdc36 feat: add SFT support with ZeRO optimization strategies OleehyO 2025-01-11 02:13:32 +00:00
  • e213b6c083 fix: pad latent frames to match patch_size_t requirements OleehyO 2025-01-11 02:08:07 +00:00
  • 70ca65300c
    upgrade to moviepy v2 Erfan Asgari 2025-01-11 00:18:24 +03:30
  • 6f43c5ef65
    Pin moviepy to v1.0.3 to avoid breaking changes in v2 Erfan Asgari 2025-01-10 20:25:01 +03:30
  • 6a78278c0b
    Update README.md Erfan Asgari 2025-01-10 20:22:25 +03:30
  • f6d722cec7 fix: remove copying first video frame as conditioning image OleehyO 2025-01-09 15:52:51 +00:00
  • 07766001f6 feat(dataset): pad short videos by repeating last frame OleehyO 2025-01-08 02:14:56 +00:00
  • 8f1829f1cd
    Merge pull request #642 from THUDM/CogVideoX_dev Yuxuan Zhang 2025-01-08 09:51:39 +08:00
  • 045e1b308b readme zR 2025-01-08 09:50:08 +08:00
  • 249fadfb76 docs: add hardware requirements for model training OleehyO 2025-01-08 01:39:37 +00:00
  • 10de04fc08 perf: cast VAE and text encoder to target dtype before precomputing cache OleehyO 2025-01-08 01:38:13 +00:00
  • 0e21d41b12 Merge remote-tracking branch 'upstream/CogVideoX_dev' into dev OleehyO 2025-01-07 09:51:48 +00:00
  • 392e37021a Add video path to error message for better debugging OleehyO 2025-01-07 09:50:21 +00:00
  • 11935892ae remove --image_column zR 2025-01-07 16:37:11 +08:00
  • ee1f666206 docs: update READMEs with auto first-frame extraction feature OleehyO 2025-01-07 06:45:10 +00:00
  • e084a4a270 feat: auto-extract first frames as conditioning images for i2v model OleehyO 2025-01-07 06:43:26 +00:00
  • 96e511b413 feat: add warning for fp16 mixed precision training OleehyO 2025-01-07 06:00:38 +00:00
  • 36427274d6 style: format import statements across finetune module OleehyO 2025-01-07 05:47:39 +00:00
  • 1789f07256 format and check fp16 for cogvideox2b zR 2025-01-07 13:16:18 +08:00
  • 1b886326b2 Merge remote-tracking branch 'upstream/CogVideoX_dev' into dev OleehyO 2025-01-06 10:47:56 +00:00
  • 9157e0cbc8 Adapt dataset for text embeddings and add noise padding OleehyO 2025-01-06 10:44:58 +00:00
  • 49dc370de6 fix: remove pipeline hooks after validation OleehyO 2025-01-04 06:21:17 +00:00
  • 93b906b3fb docs: clarify train_frames includes padding frame OleehyO 2025-01-04 06:20:25 +00:00
  • 7e1ac76847 feat(cogvideox): add prompt embedding caching support OleehyO 2025-01-04 06:17:56 +00:00
  • 66e4ba2592 fix(cogvideox): add prompt embedding caching and fix frame padding OleehyO 2025-01-03 09:27:33 +00:00
  • de5bef6611 feat(args): add train_resolution validation for video frames and resolution OleehyO 2025-01-03 08:48:42 +00:00
  • ffb6ee36b4 docs: update finetune documentation in all languages OleehyO 2025-01-03 08:30:41 +00:00
  • c817e7f062 chore: update default training parameters for t2v and i2v scripts OleehyO 2025-01-03 08:27:49 +00:00
  • e5b8f9a2ee feat: add caching for prompt embeddings OleehyO 2025-01-03 07:52:10 +00:00
  • f731c35f70 Add unload_model function OleehyO 2025-01-03 07:50:36 +00:00
  • ce2c299c1f Update diffusion_video.py zR 2025-01-03 08:45:42 +08:00
  • b080c6a010 put lora back(sat), unavailable running zR 2025-01-02 11:48:18 +08:00
  • a88c1ede69 feat(args): add validation for training resolution OleehyO 2025-01-02 03:12:09 +00:00
  • 362b7bf273 docs: update README in multiple languages OleehyO 2025-01-02 03:07:34 +00:00
  • aa240dc675
    Merge pull request #632 from THUDM/CogVideoX_dev Yuxuan Zhang 2025-01-02 08:31:25 +08:00
  • cf2fff7e55 Merge remote-tracking branch 'upstream/main' into dev OleehyO 2025-01-01 16:03:51 +00:00
  • 7fa1bb48be refactor: remove deprecated training scripts OleehyO 2025-01-01 15:56:14 +00:00
  • 48ad178818 Reorganize training script arguments OleehyO 2025-01-01 15:52:39 +00:00
  • 6ef15dd2a5 docs: update TOC and add friendly link in README files 三洋三洋 2024-12-27 19:37:08 +08:00
  • 6e79472417 feat: add training launch scripts for I2V and T2V models OleehyO 2025-01-01 14:53:45 +00:00
  • 26b87cd4ff feat(args): add validation and arg interface for training parameters OleehyO 2025-01-01 14:40:09 +00:00
  • 04a60e7435 Change logger name to trainer OleehyO 2025-01-01 14:22:59 +00:00
  • a001842834 feat: implement CogVideoX trainers for I2V and T2V tasks OleehyO 2024-12-31 17:27:47 +00:00
  • 91d79fd9a4 feat: add schemas module for configuration and state management OleehyO 2024-12-31 17:25:02 +00:00
  • 45d40450a1 refactor: simplify dataset implementation and add latent precomputation OleehyO 2024-12-30 16:14:46 +00:00
  • 6eae5c201e feat: add latent caching for video encodings OleehyO 2024-12-30 16:10:06 +00:00
  • 2a6cca0656 Add type conversion and validation checks OleehyO 2024-12-30 06:53:23 +00:00
  • fa4659fb2c feat(trainer): add validation functionality to Trainer class OleehyO 2024-12-30 06:51:03 +00:00
  • 6971364591 Export file_utils.py OleehyO 2024-12-30 06:49:45 +00:00
  • 60f6a3d7ee feat: add base trainer implementation and training script OleehyO 2024-12-29 15:27:43 +00:00
  • a505f2e312 Add constants.py OleehyO 2024-12-29 15:06:52 +00:00
  • 78f655a9a4 Add utils OleehyO 2024-12-29 15:06:33 +00:00
  • 85e00a1082 feat(models): add scaffolding OleehyO 2024-12-27 09:59:49 +00:00
  • 918ebb5a54 feat(datasets): implement video dataset modules OleehyO 2024-12-27 09:57:37 +00:00
  • e3f6def234 feat: add video frame extraction tool OleehyO 2024-12-27 09:50:47 +00:00
  • 7b282246dd chore: remove unused configuration files after refactoring OleehyO 2024-12-27 09:47:45 +00:00
  • 5cb9303286 chore: update .gitignore OleehyO 2024-12-27 09:42:42 +00:00
  • ba85627577 [docs] improve help messages in argument parser OleehyO 2024-12-18 12:30:13 +00:00
  • 2508c8353b [bugfix] fix specific resolution setting OleehyO 2024-12-18 07:38:10 +00:00
  • 48ac9c1066 [fix]fix typo in train_cogvideox_image_to_video_lora.py Gforky 2024-12-14 16:12:57 +08:00
  • 21693ca770 fix bugs of image-to-video without image-condition Zheng Guang Cong 2024-12-06 20:14:43 +08:00
  • a6e611e354 docs: update TOC and add friendly link in README files 三洋三洋 2024-12-27 19:37:08 +08:00
  • 7935bd58a1
    Merge pull request #615 from THUDM/CogVideoX_dev Yuxuan.Zhang 2024-12-19 12:57:56 +08:00
  • 1811c50e73 [docs] improve help messages in argument parser OleehyO 2024-12-18 12:30:13 +00:00
  • 92a589240f [bugfix] fix specific resolution setting OleehyO 2024-12-18 07:38:10 +00:00
  • 7add8f8437
    Merge pull request #607 from THUDM/CogVideoX_dev OleehyO 2024-12-17 09:58:10 +08:00
  • cfaca91cde Merge remote-tracking branch 'upstream/main' into dev OleehyO 2024-12-16 11:38:26 +00:00
  • d3a7d2dc91 Add resolution warning OleehyO 2024-12-16 11:34:51 +00:00
  • 46098f446b
    Merge pull request #603 from Gforky/fix-demo-issue Yuxuan.Zhang 2024-12-15 22:00:41 +08:00
  • 5a03e6fa79 [fix]fix typo in train_cogvideox_image_to_video_lora.py Gforky 2024-12-14 16:12:57 +08:00
  • 1605e95033
    Merge pull request #599 from THUDM/CogVideoX_dev Yuxuan.Zhang 2024-12-13 15:03:48 +08:00
  • 7b4c9db6d9 Fix for CogVideoX-{2B,5B} OleehyO 2024-12-12 14:18:55 +00:00
  • 36f1333788 Fix for deepspeed training OleehyO 2024-12-12 13:38:25 +00:00
  • 4d1b9fd166 Fix for Disney video dataset OleehyO 2024-12-12 13:35:30 +00:00
  • 3ff9d3049d docs: change "read this in English" to "中文阅读" OleehyO 2024-12-11 05:10:28 +00:00
  • 496e220463
    Merge pull request #585 from ZGCTroy/patch-1 Yuxuan.Zhang 2024-12-08 19:31:59 +08:00
  • a46d762cd9
    fix bugs of image-to-video without image-condition Zheng Guang Cong 2024-12-06 20:14:43 +08:00
  • 923f55e213 fix: invalid call to is_context_parallel_initialized Jade Zheng 2024-12-04 18:12:32 +08:00
  • 87ccd38cea
    Merge pull request #567 from THUDM/main Yuxuan.Zhang 2024-12-02 11:30:20 +08:00
  • 5aa6d3a9ee
    Merge pull request #515 from Gforky/fix_finetune_demo Yuxuan.Zhang 2024-12-02 11:29:42 +08:00
  • a094b34425
    Merge pull request #565 from THUDM/CogVideoX_dev Yuxuan.Zhang 2024-11-30 12:45:25 +08:00
  • 0fe46df21f new jobs of friendly link zR 2024-11-30 12:40:07 +08:00
  • f1a2b48974
    Merge pull request #556 from THUDM/main Yuxuan.Zhang 2024-11-27 12:11:12 +08:00
  • d82922cc79
    Merge pull request #538 from spacegoing/fix_rope_finetune_shape Yuxuan.Zhang 2024-11-23 21:24:39 +08:00
  • 2db58d97dc
    Update cli_demo.py devopseng99 2024-11-22 15:57:39 -06:00
  • 2fb763d25f [Fix] fix rope temporal patch size spacegoing 2024-11-21 16:21:30 +00:00
  • 04df6ec083 [Fix] fix rope temporal patch size spacegoing 2024-11-21 16:21:30 +00:00
  • ac2f2c78f7 [fix]fix deepspeed initialization issue in finetune examples luwen.miao 2024-11-18 09:49:31 +00:00
  • 2fdc59c3ce
    Merge pull request #507 from THUDM/CogVideoX_dev Yuxuan.Zhang 2024-11-17 21:54:47 +08:00
  • 17996f11f8 update zR 2024-11-16 10:06:22 +08:00
  • 5e3e3aabe0
    Merge pull request #500 from THUDM/main Yuxuan.Zhang 2024-11-13 21:15:49 +08:00
  • e7a35ea33b update friendly link zR 2024-11-13 17:06:16 +08:00
  • cd5ceca22b fix resolution docs zR 2024-11-12 00:41:23 +08:00
  • bb2cb130a0 add width and height zR 2024-11-12 00:17:19 +08:00
  • 2151a3bdfb update with diffusers zR 2024-11-11 22:41:28 +08:00
  • 68d93ce8fc fix zR 2024-11-09 22:51:39 +08:00
  • 155456befa update zR 2024-11-09 22:49:03 +08:00