logs/zeroshot_vis.log

WAN_GPU set explicitly: using GPU 2 (no headroom check)
task:     vace-1.3B
size:     832*480
ckpt_dir: /home/quang/.cache/huggingface/hub/models--Wan-AI--Wan2.1-VACE-1.3B/snapshots/574e6a744642ce3bee319afc31496b88bde8aac4
save to:  /srv/data/foundation-physics-graph-model/outputs/zeroshot/zs_vis.mp4
GPU:      2
[2026-08-03 18:26:49,780] INFO: offload_model is not specified, set to True.
[2026-08-03 18:26:49,780] INFO: Generation job args: Namespace(task='vace-1.3B', size='832*480', frame_num=81, ckpt_dir='/home/quang/.cache/huggingface/hub/models--Wan-AI--Wan2.1-VACE-1.3B/snapshots/574e6a744642ce3bee319afc31496b88bde8aac4', offload_model=True, ulysses_size=1, ring_size=1, t5_fsdp=False, t5_cpu=False, dit_fsdp=False, save_file='/srv/data/foundation-physics-graph-model/outputs/zeroshot/zs_vis.mp4', src_video='/srv/data/foundation-physics-graph-model/outputs/zeroshot/control_vis.mkv', src_mask=None, src_ref_images='/srv/data/foundation-physics-graph-model/outputs/datagen/AUTOLab+0d4edc83+2023-10-21-19h-06m-10s/22008760/vace/window_00000_00081/ref_plate.png', prompt='A Franka Panda robot arm with a Robotiq gripper on a wooden workbench: Put brick in drawer shelf and close drawer', use_prompt_extend=False, prompt_extend_method='local_qwen', prompt_extend_model=None, prompt_extend_target_lang='zh', base_seed=42, image=None, first_frame=None, last_frame=None, sample_solver='unipc', sample_steps=50, sample_shift=16, sample_guide_scale=5.0)
[2026-08-03 18:26:49,780] INFO: Generation model config: {'__name__': 'Config: Wan T2V 1.3B', 't5_model': 'umt5_xxl', 't5_dtype': torch.bfloat16, 'text_len': 512, 'param_dtype': torch.bfloat16, 'num_train_timesteps': 1000, 'sample_fps': 16, 'sample_neg_prompt': '色调艳丽，过曝，静态，细节模糊不清，字幕，风格，作品，画作，画面，静止，整体发灰，最差质量，低质量，JPEG压缩残留，丑陋的，残缺的，多余的手指，画得不好的手部，画得不好的脸部，畸形的，毁容的，形态畸形的肢体，手指融合，静止不动的画面，杂乱的背景，三条腿，背景人很多，倒着走', 't5_checkpoint': 'models_t5_umt5-xxl-enc-bf16.pth', 't5_tokenizer': 'google/umt5-xxl', 'vae_checkpoint': 'Wan2.1_VAE.pth', 'vae_stride': (4, 8, 8), 'patch_size': (1, 2, 2), 'dim': 1536, 'ffn_dim': 8960, 'freq_dim': 256, 'num_heads': 12, 'num_layers': 30, 'window_size': (-1, -1), 'qk_norm': True, 'cross_attn_norm': True, 'eps': 1e-06}
[2026-08-03 18:26:49,780] INFO: Input prompt: A Franka Panda robot arm with a Robotiq gripper on a wooden workbench: Put brick in drawer shelf and close drawer
[2026-08-03 18:26:49,781] INFO: Creating VACE pipeline.
[2026-08-03 18:27:35,364] INFO: loading /home/quang/.cache/huggingface/hub/models--Wan-AI--Wan2.1-VACE-1.3B/snapshots/574e6a744642ce3bee319afc31496b88bde8aac4/models_t5_umt5-xxl-enc-bf16.pth
[2026-08-03 18:27:42,672] INFO: loading /home/quang/.cache/huggingface/hub/models--Wan-AI--Wan2.1-VACE-1.3B/snapshots/574e6a744642ce3bee319afc31496b88bde8aac4/Wan2.1_VAE.pth
[2026-08-03 18:27:42,827] INFO: Creating VaceWanModel from /home/quang/.cache/huggingface/hub/models--Wan-AI--Wan2.1-VACE-1.3B/snapshots/574e6a744642ce3bee319afc31496b88bde8aac4
[2026-08-03 18:27:46,677] INFO: Generating video...

  0%|          | 0/50 [00:00<?, ?it/s]
  2%|▏         | 1/50 [00:03<02:37,  3.21s/it]
  4%|▍         | 2/50 [00:06<02:33,  3.21s/it]
  6%|▌         | 3/50 [00:09<02:35,  3.31s/it]
  8%|▊         | 4/50 [00:13<02:30,  3.27s/it]
 10%|█         | 5/50 [00:16<02:26,  3.25s/it]
 12%|█▏        | 6/50 [00:19<02:22,  3.24s/it]
 14%|█▍        | 7/50 [00:22<02:18,  3.23s/it]
 16%|█▌        | 8/50 [00:25<02:15,  3.22s/it]
 18%|█▊        | 9/50 [00:29<02:12,  3.23s/it]
 20%|██        | 10/50 [00:32<02:09,  3.23s/it]
 22%|██▏       | 11/50 [00:35<02:05,  3.22s/it]
 24%|██▍       | 12/50 [00:38<02:02,  3.22s/it]
 26%|██▌       | 13/50 [00:42<01:59,  3.22s/it]
 28%|██▊       | 14/50 [00:45<02:02,  3.39s/it]
 30%|███       | 15/50 [00:49<01:56,  3.34s/it]
 32%|███▏      | 16/50 [00:52<01:52,  3.30s/it]
 34%|███▍      | 17/50 [00:55<01:48,  3.28s/it]
 36%|███▌      | 18/50 [00:58<01:44,  3.26s/it]
 38%|███▊      | 19/50 [01:01<01:40,  3.25s/it]
 40%|████      | 20/50 [01:05<01:37,  3.24s/it]
 42%|████▏     | 21/50 [01:08<01:33,  3.24s/it]
 44%|████▍     | 22/50 [01:11<01:30,  3.23s/it]
 46%|████▌     | 23/50 [01:14<01:27,  3.23s/it]
 48%|████▊     | 24/50 [01:18<01:23,  3.23s/it]
 50%|█████     | 25/50 [01:21<01:20,  3.22s/it]
 52%|█████▏    | 26/50 [01:24<01:17,  3.22s/it]
 54%|█████▍    | 27/50 [01:27<01:14,  3.22s/it]
 56%|█████▌    | 28/50 [01:30<01:10,  3.22s/it]
 58%|█████▊    | 29/50 [01:34<01:07,  3.22s/it]
 60%|██████    | 30/50 [01:37<01:04,  3.23s/it]
 62%|██████▏   | 31/50 [01:40<01:01,  3.23s/it]
 64%|██████▍   | 32/50 [01:43<00:58,  3.23s/it]
 66%|██████▌   | 33/50 [01:47<00:54,  3.22s/it]
 68%|██████▊   | 34/50 [01:50<00:51,  3.22s/it]
 70%|███████   | 35/50 [01:53<00:48,  3.22s/it]
 72%|███████▏  | 36/50 [01:56<00:45,  3.22s/it]
 74%|███████▍  | 37/50 [01:59<00:41,  3.22s/it]
 76%|███████▌  | 38/50 [02:03<00:38,  3.22s/it]
 78%|███████▊  | 39/50 [02:06<00:35,  3.22s/it]
 80%|████████  | 40/50 [02:09<00:32,  3.22s/it]
 82%|████████▏ | 41/50 [02:12<00:28,  3.22s/it]
 84%|████████▍ | 42/50 [02:15<00:25,  3.22s/it]
 86%|████████▌ | 43/50 [02:19<00:22,  3.22s/it]
 88%|████████▊ | 44/50 [02:22<00:19,  3.22s/it]
 90%|█████████ | 45/50 [02:25<00:16,  3.22s/it]
 92%|█████████▏| 46/50 [02:28<00:12,  3.22s/it]
 94%|█████████▍| 47/50 [02:32<00:09,  3.22s/it]
 96%|█████████▌| 48/50 [02:35<00:06,  3.22s/it]
 98%|█████████▊| 49/50 [02:38<00:03,  3.22s/it]
100%|██████████| 50/50 [02:41<00:00,  3.22s/it]
100%|██████████| 50/50 [02:41<00:00,  3.23s/it]
[2026-08-03 18:30:42,298] INFO: Saving generated video to /srv/data/foundation-physics-graph-model/outputs/zeroshot/zs_vis.mp4
[2026-08-03 18:30:43,425] INFO: Finished.