Files
unifolm-world-model-action/run_all_cases_20260211_194859.log
2026-02-12 10:47:21 +08:00

2469 lines
137 KiB
Plaintext

2026-02-11 19:49:03.885238: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2026-02-11 19:49:03.934263: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2026-02-11 19:49:03.934309: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2026-02-11 19:49:03.935622: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2026-02-11 19:49:03.943041: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2026-02-11 19:49:04.852993: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Global seed set to 123
INFO:mainlogger:LatentVisualDiffusion: Running in v-prediction mode
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
AE working on z of shape (1, 4, 32, 32) = 4096 dimensions.
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
>>> model checkpoint loaded.
>>> Load pre-trained model ...
INFO:root:***** Configing Data *****
>>> unitree_z1_stackbox: 1 data samples loaded.
>>> unitree_z1_stackbox: data stats loaded.
>>> unitree_z1_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox: data stats loaded.
>>> unitree_z1_dual_arm_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox_v2: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox_v2: data stats loaded.
>>> unitree_z1_dual_arm_stackbox_v2: normalizer initiated.
>>> unitree_z1_dual_arm_cleanup_pencils: 1 data samples loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: data stats loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: normalizer initiated.
>>> unitree_g1_pack_camera: 1 data samples loaded.
>>> unitree_g1_pack_camera: data stats loaded.
>>> unitree_g1_pack_camera: normalizer initiated.
>>> Dataset is successfully loaded ...
>>> Generate 16 frames under each generation ...
DEBUG:h5py._conv:Creating converter from 3 to 5
DEBUG:PIL.PngImagePlugin:STREAM b'IHDR' 16 13
DEBUG:PIL.PngImagePlugin:STREAM b'pHYs' 41 9
DEBUG:PIL.PngImagePlugin:STREAM b'IDAT' 62 4096
0%| | 0/11 [00:00<?, ?it/s]>>> Step 0: generating actions ...
>>> Step 0: interacting with world model ...
DEBUG:PIL.Image:Importing BlpImagePlugin
DEBUG:PIL.Image:Importing BmpImagePlugin
DEBUG:PIL.Image:Importing BufrStubImagePlugin
DEBUG:PIL.Image:Importing CurImagePlugin
DEBUG:PIL.Image:Importing DcxImagePlugin
DEBUG:PIL.Image:Importing DdsImagePlugin
DEBUG:PIL.Image:Importing EpsImagePlugin
DEBUG:PIL.Image:Importing FitsImagePlugin
DEBUG:PIL.Image:Importing FitsStubImagePlugin
DEBUG:PIL.Image:Importing FliImagePlugin
DEBUG:PIL.Image:Importing FpxImagePlugin
DEBUG:PIL.Image:Image: failed to import FpxImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing FtexImagePlugin
DEBUG:PIL.Image:Importing GbrImagePlugin
DEBUG:PIL.Image:Importing GifImagePlugin
DEBUG:PIL.Image:Importing GribStubImagePlugin
DEBUG:PIL.Image:Importing Hdf5StubImagePlugin
DEBUG:PIL.Image:Importing IcnsImagePlugin
DEBUG:PIL.Image:Importing IcoImagePlugin
DEBUG:PIL.Image:Importing ImImagePlugin
DEBUG:PIL.Image:Importing ImtImagePlugin
DEBUG:PIL.Image:Importing IptcImagePlugin
DEBUG:PIL.Image:Importing JpegImagePlugin
DEBUG:PIL.Image:Importing Jpeg2KImagePlugin
DEBUG:PIL.Image:Importing McIdasImagePlugin
DEBUG:PIL.Image:Importing MicImagePlugin
DEBUG:PIL.Image:Image: failed to import MicImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing MpegImagePlugin
DEBUG:PIL.Image:Importing MpoImagePlugin
DEBUG:PIL.Image:Importing MspImagePlugin
DEBUG:PIL.Image:Importing PalmImagePlugin
DEBUG:PIL.Image:Importing PcdImagePlugin
DEBUG:PIL.Image:Importing PcxImagePlugin
DEBUG:PIL.Image:Importing PdfImagePlugin
DEBUG:PIL.Image:Importing PixarImagePlugin
DEBUG:PIL.Image:Importing PngImagePlugin
DEBUG:PIL.Image:Importing PpmImagePlugin
DEBUG:PIL.Image:Importing PsdImagePlugin
DEBUG:PIL.Image:Importing QoiImagePlugin
DEBUG:PIL.Image:Importing SgiImagePlugin
DEBUG:PIL.Image:Importing SpiderImagePlugin
DEBUG:PIL.Image:Importing SunImagePlugin
DEBUG:PIL.Image:Importing TgaImagePlugin
DEBUG:PIL.Image:Importing TiffImagePlugin
DEBUG:PIL.Image:Importing WebPImagePlugin
DEBUG:PIL.Image:Importing WmfImagePlugin
DEBUG:PIL.Image:Importing XbmImagePlugin
DEBUG:PIL.Image:Importing XpmImagePlugin
DEBUG:PIL.Image:Importing XVThumbImagePlugin
9%|▉ | 1/11 [01:12<12:09, 72.95s/it]
18%|█▊ | 2/11 [02:26<10:58, 73.19s/it]
27%|██▋ | 3/11 [03:39<09:45, 73.21s/it]
36%|███▋ | 4/11 [04:52<08:32, 73.21s/it]
45%|████▌ | 5/11 [06:05<07:19, 73.22s/it]
55%|█████▍ | 6/11 [07:19<06:06, 73.20s/it]
64%|██████▎ | 7/11 [08:32<04:52, 73.16s/it]
73%|███████▎ | 8/11 [09:45<03:39, 73.14s/it]
82%|████████▏ | 9/11 [10:58<02:26, 73.13s/it]
91%|█████████ | 10/11 [12:11<01:13, 73.15s/it]
100%|██████████| 11/11 [13:24<00:00, 73.18s/it]
100%|██████████| 11/11 [13:24<00:00, 73.17s/it]
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 1: generating actions ...
>>> Step 1: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 2: generating actions ...
>>> Step 2: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 3: generating actions ...
>>> Step 3: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 4: generating actions ...
>>> Step 4: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 5: generating actions ...
>>> Step 5: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 6: generating actions ...
>>> Step 6: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 7: generating actions ...
>>> Step 7: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 8: generating actions ...
>>> Step 8: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 9: generating actions ...
>>> Step 9: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 10: generating actions ...
>>> Step 10: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
real 15m0.258s
user 65m49.394s
sys 0m45.300s
2026-02-11 20:04:06.049535: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2026-02-11 20:04:06.099186: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2026-02-11 20:04:06.099232: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2026-02-11 20:04:06.100544: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2026-02-11 20:04:06.108023: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2026-02-11 20:04:07.025500: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Global seed set to 123
INFO:mainlogger:LatentVisualDiffusion: Running in v-prediction mode
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
AE working on z of shape (1, 4, 32, 32) = 4096 dimensions.
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
>>> model checkpoint loaded.
>>> Load pre-trained model ...
INFO:root:***** Configing Data *****
>>> unitree_z1_stackbox: 1 data samples loaded.
>>> unitree_z1_stackbox: data stats loaded.
>>> unitree_z1_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox: data stats loaded.
>>> unitree_z1_dual_arm_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox_v2: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox_v2: data stats loaded.
>>> unitree_z1_dual_arm_stackbox_v2: normalizer initiated.
>>> unitree_z1_dual_arm_cleanup_pencils: 1 data samples loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: data stats loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: normalizer initiated.
>>> unitree_g1_pack_camera: 1 data samples loaded.
>>> unitree_g1_pack_camera: data stats loaded.
>>> unitree_g1_pack_camera: normalizer initiated.
>>> Dataset is successfully loaded ...
>>> Generate 16 frames under each generation ...
DEBUG:h5py._conv:Creating converter from 3 to 5
DEBUG:PIL.PngImagePlugin:STREAM b'IHDR' 16 13
DEBUG:PIL.PngImagePlugin:STREAM b'pHYs' 41 9
DEBUG:PIL.PngImagePlugin:STREAM b'IDAT' 62 4096
0%| | 0/11 [00:00<?, ?it/s]>>> Step 0: generating actions ...
>>> Step 0: interacting with world model ...
DEBUG:PIL.Image:Importing BlpImagePlugin
DEBUG:PIL.Image:Importing BmpImagePlugin
DEBUG:PIL.Image:Importing BufrStubImagePlugin
DEBUG:PIL.Image:Importing CurImagePlugin
DEBUG:PIL.Image:Importing DcxImagePlugin
DEBUG:PIL.Image:Importing DdsImagePlugin
DEBUG:PIL.Image:Importing EpsImagePlugin
DEBUG:PIL.Image:Importing FitsImagePlugin
DEBUG:PIL.Image:Importing FitsStubImagePlugin
DEBUG:PIL.Image:Importing FliImagePlugin
DEBUG:PIL.Image:Importing FpxImagePlugin
DEBUG:PIL.Image:Image: failed to import FpxImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing FtexImagePlugin
DEBUG:PIL.Image:Importing GbrImagePlugin
DEBUG:PIL.Image:Importing GifImagePlugin
DEBUG:PIL.Image:Importing GribStubImagePlugin
DEBUG:PIL.Image:Importing Hdf5StubImagePlugin
DEBUG:PIL.Image:Importing IcnsImagePlugin
DEBUG:PIL.Image:Importing IcoImagePlugin
DEBUG:PIL.Image:Importing ImImagePlugin
DEBUG:PIL.Image:Importing ImtImagePlugin
DEBUG:PIL.Image:Importing IptcImagePlugin
DEBUG:PIL.Image:Importing JpegImagePlugin
DEBUG:PIL.Image:Importing Jpeg2KImagePlugin
DEBUG:PIL.Image:Importing McIdasImagePlugin
DEBUG:PIL.Image:Importing MicImagePlugin
DEBUG:PIL.Image:Image: failed to import MicImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing MpegImagePlugin
DEBUG:PIL.Image:Importing MpoImagePlugin
DEBUG:PIL.Image:Importing MspImagePlugin
DEBUG:PIL.Image:Importing PalmImagePlugin
DEBUG:PIL.Image:Importing PcdImagePlugin
DEBUG:PIL.Image:Importing PcxImagePlugin
DEBUG:PIL.Image:Importing PdfImagePlugin
DEBUG:PIL.Image:Importing PixarImagePlugin
DEBUG:PIL.Image:Importing PngImagePlugin
DEBUG:PIL.Image:Importing PpmImagePlugin
DEBUG:PIL.Image:Importing PsdImagePlugin
DEBUG:PIL.Image:Importing QoiImagePlugin
DEBUG:PIL.Image:Importing SgiImagePlugin
DEBUG:PIL.Image:Importing SpiderImagePlugin
DEBUG:PIL.Image:Importing SunImagePlugin
DEBUG:PIL.Image:Importing TgaImagePlugin
DEBUG:PIL.Image:Importing TiffImagePlugin
DEBUG:PIL.Image:Importing WebPImagePlugin
DEBUG:PIL.Image:Importing WmfImagePlugin
DEBUG:PIL.Image:Importing XbmImagePlugin
DEBUG:PIL.Image:Importing XpmImagePlugin
DEBUG:PIL.Image:Importing XVThumbImagePlugin
9%|▉ | 1/11 [01:14<12:22, 74.22s/it]
18%|█▊ | 2/11 [02:28<11:09, 74.33s/it]
27%|██▋ | 3/11 [03:42<09:54, 74.32s/it]
36%|███▋ | 4/11 [04:57<08:40, 74.32s/it]
45%|████▌ | 5/11 [06:11<07:25, 74.28s/it]
55%|█████▍ | 6/11 [07:25<06:10, 74.19s/it]
64%|██████▎ | 7/11 [08:39<04:56, 74.11s/it]
73%|███████▎ | 8/11 [09:53<03:42, 74.07s/it]
82%|████████▏ | 9/11 [11:07<02:28, 74.06s/it]
91%|█████████ | 10/11 [12:21<01:14, 74.01s/it]
100%|██████████| 11/11 [13:35<00:00, 73.98s/it]
100%|██████████| 11/11 [13:35<00:00, 74.12s/it]
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 1: generating actions ...
>>> Step 1: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 2: generating actions ...
>>> Step 2: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 3: generating actions ...
>>> Step 3: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 4: generating actions ...
>>> Step 4: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 5: generating actions ...
>>> Step 5: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 6: generating actions ...
>>> Step 6: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 7: generating actions ...
>>> Step 7: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 8: generating actions ...
>>> Step 8: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 9: generating actions ...
>>> Step 9: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 10: generating actions ...
>>> Step 10: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
real 15m13.215s
user 66m8.595s
sys 0m45.545s
2026-02-11 20:19:19.271045: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2026-02-11 20:19:19.320688: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2026-02-11 20:19:19.320734: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2026-02-11 20:19:19.322059: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2026-02-11 20:19:19.329606: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2026-02-11 20:19:20.248938: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Global seed set to 123
INFO:mainlogger:LatentVisualDiffusion: Running in v-prediction mode
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
AE working on z of shape (1, 4, 32, 32) = 4096 dimensions.
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
>>> model checkpoint loaded.
>>> Load pre-trained model ...
INFO:root:***** Configing Data *****
>>> unitree_z1_stackbox: 1 data samples loaded.
>>> unitree_z1_stackbox: data stats loaded.
>>> unitree_z1_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox: data stats loaded.
>>> unitree_z1_dual_arm_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox_v2: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox_v2: data stats loaded.
>>> unitree_z1_dual_arm_stackbox_v2: normalizer initiated.
>>> unitree_z1_dual_arm_cleanup_pencils: 1 data samples loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: data stats loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: normalizer initiated.
>>> unitree_g1_pack_camera: 1 data samples loaded.
>>> unitree_g1_pack_camera: data stats loaded.
>>> unitree_g1_pack_camera: normalizer initiated.
>>> Dataset is successfully loaded ...
>>> Generate 16 frames under each generation ...
DEBUG:h5py._conv:Creating converter from 3 to 5
DEBUG:PIL.PngImagePlugin:STREAM b'IHDR' 16 13
DEBUG:PIL.PngImagePlugin:STREAM b'pHYs' 41 9
DEBUG:PIL.PngImagePlugin:STREAM b'IDAT' 62 4096
0%| | 0/11 [00:00<?, ?it/s]>>> Step 0: generating actions ...
>>> Step 0: interacting with world model ...
DEBUG:PIL.Image:Importing BlpImagePlugin
DEBUG:PIL.Image:Importing BmpImagePlugin
DEBUG:PIL.Image:Importing BufrStubImagePlugin
DEBUG:PIL.Image:Importing CurImagePlugin
DEBUG:PIL.Image:Importing DcxImagePlugin
DEBUG:PIL.Image:Importing DdsImagePlugin
DEBUG:PIL.Image:Importing EpsImagePlugin
DEBUG:PIL.Image:Importing FitsImagePlugin
DEBUG:PIL.Image:Importing FitsStubImagePlugin
DEBUG:PIL.Image:Importing FliImagePlugin
DEBUG:PIL.Image:Importing FpxImagePlugin
DEBUG:PIL.Image:Image: failed to import FpxImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing FtexImagePlugin
DEBUG:PIL.Image:Importing GbrImagePlugin
DEBUG:PIL.Image:Importing GifImagePlugin
DEBUG:PIL.Image:Importing GribStubImagePlugin
DEBUG:PIL.Image:Importing Hdf5StubImagePlugin
DEBUG:PIL.Image:Importing IcnsImagePlugin
DEBUG:PIL.Image:Importing IcoImagePlugin
DEBUG:PIL.Image:Importing ImImagePlugin
DEBUG:PIL.Image:Importing ImtImagePlugin
DEBUG:PIL.Image:Importing IptcImagePlugin
DEBUG:PIL.Image:Importing JpegImagePlugin
DEBUG:PIL.Image:Importing Jpeg2KImagePlugin
DEBUG:PIL.Image:Importing McIdasImagePlugin
DEBUG:PIL.Image:Importing MicImagePlugin
DEBUG:PIL.Image:Image: failed to import MicImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing MpegImagePlugin
DEBUG:PIL.Image:Importing MpoImagePlugin
DEBUG:PIL.Image:Importing MspImagePlugin
DEBUG:PIL.Image:Importing PalmImagePlugin
DEBUG:PIL.Image:Importing PcdImagePlugin
DEBUG:PIL.Image:Importing PcxImagePlugin
DEBUG:PIL.Image:Importing PdfImagePlugin
DEBUG:PIL.Image:Importing PixarImagePlugin
DEBUG:PIL.Image:Importing PngImagePlugin
DEBUG:PIL.Image:Importing PpmImagePlugin
DEBUG:PIL.Image:Importing PsdImagePlugin
DEBUG:PIL.Image:Importing QoiImagePlugin
DEBUG:PIL.Image:Importing SgiImagePlugin
DEBUG:PIL.Image:Importing SpiderImagePlugin
DEBUG:PIL.Image:Importing SunImagePlugin
DEBUG:PIL.Image:Importing TgaImagePlugin
DEBUG:PIL.Image:Importing TiffImagePlugin
DEBUG:PIL.Image:Importing WebPImagePlugin
DEBUG:PIL.Image:Importing WmfImagePlugin
DEBUG:PIL.Image:Importing XbmImagePlugin
DEBUG:PIL.Image:Importing XpmImagePlugin
DEBUG:PIL.Image:Importing XVThumbImagePlugin
9%|▉ | 1/11 [01:14<12:22, 74.28s/it]
18%|█▊ | 2/11 [02:28<11:09, 74.38s/it]
27%|██▋ | 3/11 [03:43<09:55, 74.45s/it]
36%|███▋ | 4/11 [04:57<08:41, 74.43s/it]
45%|████▌ | 5/11 [06:11<07:25, 74.25s/it]
55%|█████▍ | 6/11 [07:26<06:11, 74.31s/it]
64%|██████▎ | 7/11 [08:40<04:57, 74.26s/it]
73%|███████▎ | 8/11 [09:54<03:43, 74.34s/it]
82%|████████▏ | 9/11 [11:08<02:28, 74.29s/it]
91%|█████████ | 10/11 [12:23<01:14, 74.26s/it]
100%|██████████| 11/11 [13:37<00:00, 74.39s/it]
100%|██████████| 11/11 [13:37<00:00, 74.34s/it]
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 1: generating actions ...
>>> Step 1: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 2: generating actions ...
>>> Step 2: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 3: generating actions ...
>>> Step 3: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 4: generating actions ...
>>> Step 4: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 5: generating actions ...
>>> Step 5: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 6: generating actions ...
>>> Step 6: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 7: generating actions ...
>>> Step 7: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 8: generating actions ...
>>> Step 8: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 9: generating actions ...
>>> Step 9: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 10: generating actions ...
>>> Step 10: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
real 15m15.291s
user 65m30.562s
sys 0m45.706s
2026-02-11 20:34:34.563818: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2026-02-11 20:34:34.613426: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2026-02-11 20:34:34.613485: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2026-02-11 20:34:34.614802: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2026-02-11 20:34:34.622286: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2026-02-11 20:34:35.540506: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Global seed set to 123
INFO:mainlogger:LatentVisualDiffusion: Running in v-prediction mode
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
AE working on z of shape (1, 4, 32, 32) = 4096 dimensions.
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
>>> model checkpoint loaded.
>>> Load pre-trained model ...
INFO:root:***** Configing Data *****
>>> unitree_z1_stackbox: 1 data samples loaded.
>>> unitree_z1_stackbox: data stats loaded.
>>> unitree_z1_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox: data stats loaded.
>>> unitree_z1_dual_arm_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox_v2: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox_v2: data stats loaded.
>>> unitree_z1_dual_arm_stackbox_v2: normalizer initiated.
>>> unitree_z1_dual_arm_cleanup_pencils: 1 data samples loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: data stats loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: normalizer initiated.
>>> unitree_g1_pack_camera: 1 data samples loaded.
>>> unitree_g1_pack_camera: data stats loaded.
>>> unitree_g1_pack_camera: normalizer initiated.
>>> Dataset is successfully loaded ...
>>> Generate 16 frames under each generation ...
DEBUG:h5py._conv:Creating converter from 3 to 5
DEBUG:PIL.PngImagePlugin:STREAM b'IHDR' 16 13
DEBUG:PIL.PngImagePlugin:STREAM b'pHYs' 41 9
DEBUG:PIL.PngImagePlugin:STREAM b'IDAT' 62 4096
0%| | 0/11 [00:00<?, ?it/s]>>> Step 0: generating actions ...
>>> Step 0: interacting with world model ...
DEBUG:PIL.Image:Importing BlpImagePlugin
DEBUG:PIL.Image:Importing BmpImagePlugin
DEBUG:PIL.Image:Importing BufrStubImagePlugin
DEBUG:PIL.Image:Importing CurImagePlugin
DEBUG:PIL.Image:Importing DcxImagePlugin
DEBUG:PIL.Image:Importing DdsImagePlugin
DEBUG:PIL.Image:Importing EpsImagePlugin
DEBUG:PIL.Image:Importing FitsImagePlugin
DEBUG:PIL.Image:Importing FitsStubImagePlugin
DEBUG:PIL.Image:Importing FliImagePlugin
DEBUG:PIL.Image:Importing FpxImagePlugin
DEBUG:PIL.Image:Image: failed to import FpxImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing FtexImagePlugin
DEBUG:PIL.Image:Importing GbrImagePlugin
DEBUG:PIL.Image:Importing GifImagePlugin
DEBUG:PIL.Image:Importing GribStubImagePlugin
DEBUG:PIL.Image:Importing Hdf5StubImagePlugin
DEBUG:PIL.Image:Importing IcnsImagePlugin
DEBUG:PIL.Image:Importing IcoImagePlugin
DEBUG:PIL.Image:Importing ImImagePlugin
DEBUG:PIL.Image:Importing ImtImagePlugin
DEBUG:PIL.Image:Importing IptcImagePlugin
DEBUG:PIL.Image:Importing JpegImagePlugin
DEBUG:PIL.Image:Importing Jpeg2KImagePlugin
DEBUG:PIL.Image:Importing McIdasImagePlugin
DEBUG:PIL.Image:Importing MicImagePlugin
DEBUG:PIL.Image:Image: failed to import MicImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing MpegImagePlugin
DEBUG:PIL.Image:Importing MpoImagePlugin
DEBUG:PIL.Image:Importing MspImagePlugin
DEBUG:PIL.Image:Importing PalmImagePlugin
DEBUG:PIL.Image:Importing PcdImagePlugin
DEBUG:PIL.Image:Importing PcxImagePlugin
DEBUG:PIL.Image:Importing PdfImagePlugin
DEBUG:PIL.Image:Importing PixarImagePlugin
DEBUG:PIL.Image:Importing PngImagePlugin
DEBUG:PIL.Image:Importing PpmImagePlugin
DEBUG:PIL.Image:Importing PsdImagePlugin
DEBUG:PIL.Image:Importing QoiImagePlugin
DEBUG:PIL.Image:Importing SgiImagePlugin
DEBUG:PIL.Image:Importing SpiderImagePlugin
DEBUG:PIL.Image:Importing SunImagePlugin
DEBUG:PIL.Image:Importing TgaImagePlugin
DEBUG:PIL.Image:Importing TiffImagePlugin
DEBUG:PIL.Image:Importing WebPImagePlugin
DEBUG:PIL.Image:Importing WmfImagePlugin
DEBUG:PIL.Image:Importing XbmImagePlugin
DEBUG:PIL.Image:Importing XpmImagePlugin
DEBUG:PIL.Image:Importing XVThumbImagePlugin
9%|▉ | 1/11 [01:14<12:25, 74.52s/it]
18%|█▊ | 2/11 [02:29<11:15, 75.00s/it]
27%|██▋ | 3/11 [03:44<09:59, 74.99s/it]
36%|███▋ | 4/11 [04:59<08:43, 74.74s/it]
45%|████▌ | 5/11 [06:13<07:26, 74.48s/it]
55%|█████▍ | 6/11 [07:27<06:12, 74.56s/it]
64%|██████▎ | 7/11 [08:42<04:57, 74.46s/it]
73%|███████▎ | 8/11 [09:56<03:43, 74.48s/it]
82%|████████▏ | 9/11 [11:10<02:28, 74.32s/it]
91%|█████████ | 10/11 [12:23<01:13, 73.94s/it]
100%|██████████| 11/11 [13:36<00:00, 73.64s/it]
100%|██████████| 11/11 [13:36<00:00, 74.25s/it]
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 1: generating actions ...
>>> Step 1: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 2: generating actions ...
>>> Step 2: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 3: generating actions ...
>>> Step 3: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 4: generating actions ...
>>> Step 4: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 5: generating actions ...
>>> Step 5: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 6: generating actions ...
>>> Step 6: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 7: generating actions ...
>>> Step 7: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 8: generating actions ...
>>> Step 8: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 9: generating actions ...
>>> Step 9: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 10: generating actions ...
>>> Step 10: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
real 15m13.379s
user 65m31.865s
sys 0m45.108s
2026-02-11 20:49:47.965949: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2026-02-11 20:49:48.015942: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2026-02-11 20:49:48.015997: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2026-02-11 20:49:48.017330: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2026-02-11 20:49:48.024854: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2026-02-11 20:49:48.943205: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Global seed set to 123
INFO:mainlogger:LatentVisualDiffusion: Running in v-prediction mode
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
AE working on z of shape (1, 4, 32, 32) = 4096 dimensions.
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
>>> model checkpoint loaded.
>>> Load pre-trained model ...
INFO:root:***** Configing Data *****
>>> unitree_z1_stackbox: 1 data samples loaded.
>>> unitree_z1_stackbox: data stats loaded.
>>> unitree_z1_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox: data stats loaded.
>>> unitree_z1_dual_arm_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox_v2: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox_v2: data stats loaded.
>>> unitree_z1_dual_arm_stackbox_v2: normalizer initiated.
>>> unitree_z1_dual_arm_cleanup_pencils: 1 data samples loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: data stats loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: normalizer initiated.
>>> unitree_g1_pack_camera: 1 data samples loaded.
>>> unitree_g1_pack_camera: data stats loaded.
>>> unitree_g1_pack_camera: normalizer initiated.
>>> Dataset is successfully loaded ...
>>> Generate 16 frames under each generation ...
DEBUG:h5py._conv:Creating converter from 3 to 5
DEBUG:PIL.PngImagePlugin:STREAM b'IHDR' 16 13
DEBUG:PIL.PngImagePlugin:STREAM b'pHYs' 41 9
DEBUG:PIL.PngImagePlugin:STREAM b'IDAT' 62 4096
0%| | 0/8 [00:00<?, ?it/s]>>> Step 0: generating actions ...
>>> Step 0: interacting with world model ...
DEBUG:PIL.Image:Importing BlpImagePlugin
DEBUG:PIL.Image:Importing BmpImagePlugin
DEBUG:PIL.Image:Importing BufrStubImagePlugin
DEBUG:PIL.Image:Importing CurImagePlugin
DEBUG:PIL.Image:Importing DcxImagePlugin
DEBUG:PIL.Image:Importing DdsImagePlugin
DEBUG:PIL.Image:Importing EpsImagePlugin
DEBUG:PIL.Image:Importing FitsImagePlugin
DEBUG:PIL.Image:Importing FitsStubImagePlugin
DEBUG:PIL.Image:Importing FliImagePlugin
DEBUG:PIL.Image:Importing FpxImagePlugin
DEBUG:PIL.Image:Image: failed to import FpxImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing FtexImagePlugin
DEBUG:PIL.Image:Importing GbrImagePlugin
DEBUG:PIL.Image:Importing GifImagePlugin
DEBUG:PIL.Image:Importing GribStubImagePlugin
DEBUG:PIL.Image:Importing Hdf5StubImagePlugin
DEBUG:PIL.Image:Importing IcnsImagePlugin
DEBUG:PIL.Image:Importing IcoImagePlugin
DEBUG:PIL.Image:Importing ImImagePlugin
DEBUG:PIL.Image:Importing ImtImagePlugin
DEBUG:PIL.Image:Importing IptcImagePlugin
DEBUG:PIL.Image:Importing JpegImagePlugin
DEBUG:PIL.Image:Importing Jpeg2KImagePlugin
DEBUG:PIL.Image:Importing McIdasImagePlugin
DEBUG:PIL.Image:Importing MicImagePlugin
DEBUG:PIL.Image:Image: failed to import MicImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing MpegImagePlugin
DEBUG:PIL.Image:Importing MpoImagePlugin
DEBUG:PIL.Image:Importing MspImagePlugin
DEBUG:PIL.Image:Importing PalmImagePlugin
DEBUG:PIL.Image:Importing PcdImagePlugin
DEBUG:PIL.Image:Importing PcxImagePlugin
DEBUG:PIL.Image:Importing PdfImagePlugin
DEBUG:PIL.Image:Importing PixarImagePlugin
DEBUG:PIL.Image:Importing PngImagePlugin
DEBUG:PIL.Image:Importing PpmImagePlugin
DEBUG:PIL.Image:Importing PsdImagePlugin
DEBUG:PIL.Image:Importing QoiImagePlugin
DEBUG:PIL.Image:Importing SgiImagePlugin
DEBUG:PIL.Image:Importing SpiderImagePlugin
DEBUG:PIL.Image:Importing SunImagePlugin
DEBUG:PIL.Image:Importing TgaImagePlugin
DEBUG:PIL.Image:Importing TiffImagePlugin
DEBUG:PIL.Image:Importing WebPImagePlugin
DEBUG:PIL.Image:Importing WmfImagePlugin
DEBUG:PIL.Image:Importing XbmImagePlugin
DEBUG:PIL.Image:Importing XpmImagePlugin
DEBUG:PIL.Image:Importing XVThumbImagePlugin
12%|█▎ | 1/8 [01:15<08:48, 75.51s/it]
25%|██▌ | 2/8 [02:30<07:32, 75.39s/it]
38%|███▊ | 3/8 [03:46<06:16, 75.35s/it]
50%|█████ | 4/8 [05:00<05:00, 75.01s/it]
62%|██████▎ | 5/8 [06:14<03:44, 74.68s/it]
75%|███████▌ | 6/8 [07:28<02:28, 74.40s/it]
88%|████████▊ | 7/8 [08:42<01:14, 74.19s/it]
100%|██████████| 8/8 [09:55<00:00, 73.95s/it]
100%|██████████| 8/8 [09:55<00:00, 74.47s/it]
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 1: generating actions ...
>>> Step 1: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 2: generating actions ...
>>> Step 2: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 3: generating actions ...
>>> Step 3: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 4: generating actions ...
>>> Step 4: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 5: generating actions ...
>>> Step 5: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 6: generating actions ...
>>> Step 6: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 7: generating actions ...
>>> Step 7: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
real 11m31.593s
user 48m36.594s
sys 0m41.776s
2026-02-11 21:01:19.535243: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2026-02-11 21:01:19.585230: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2026-02-11 21:01:19.585275: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2026-02-11 21:01:19.586600: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2026-02-11 21:01:19.594107: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2026-02-11 21:01:20.510688: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Global seed set to 123
INFO:mainlogger:LatentVisualDiffusion: Running in v-prediction mode
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
AE working on z of shape (1, 4, 32, 32) = 4096 dimensions.
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
>>> model checkpoint loaded.
>>> Load pre-trained model ...
INFO:root:***** Configing Data *****
>>> unitree_z1_stackbox: 1 data samples loaded.
>>> unitree_z1_stackbox: data stats loaded.
>>> unitree_z1_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox: data stats loaded.
>>> unitree_z1_dual_arm_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox_v2: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox_v2: data stats loaded.
>>> unitree_z1_dual_arm_stackbox_v2: normalizer initiated.
>>> unitree_z1_dual_arm_cleanup_pencils: 1 data samples loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: data stats loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: normalizer initiated.
>>> unitree_g1_pack_camera: 1 data samples loaded.
>>> unitree_g1_pack_camera: data stats loaded.
>>> unitree_g1_pack_camera: normalizer initiated.
>>> Dataset is successfully loaded ...
>>> Generate 16 frames under each generation ...
DEBUG:h5py._conv:Creating converter from 3 to 5
DEBUG:PIL.PngImagePlugin:STREAM b'IHDR' 16 13
DEBUG:PIL.PngImagePlugin:STREAM b'pHYs' 41 9
DEBUG:PIL.PngImagePlugin:STREAM b'IDAT' 62 4096
0%| | 0/8 [00:00<?, ?it/s]>>> Step 0: generating actions ...
>>> Step 0: interacting with world model ...
DEBUG:PIL.Image:Importing BlpImagePlugin
DEBUG:PIL.Image:Importing BmpImagePlugin
DEBUG:PIL.Image:Importing BufrStubImagePlugin
DEBUG:PIL.Image:Importing CurImagePlugin
DEBUG:PIL.Image:Importing DcxImagePlugin
DEBUG:PIL.Image:Importing DdsImagePlugin
DEBUG:PIL.Image:Importing EpsImagePlugin
DEBUG:PIL.Image:Importing FitsImagePlugin
DEBUG:PIL.Image:Importing FitsStubImagePlugin
DEBUG:PIL.Image:Importing FliImagePlugin
DEBUG:PIL.Image:Importing FpxImagePlugin
DEBUG:PIL.Image:Image: failed to import FpxImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing FtexImagePlugin
DEBUG:PIL.Image:Importing GbrImagePlugin
DEBUG:PIL.Image:Importing GifImagePlugin
DEBUG:PIL.Image:Importing GribStubImagePlugin
DEBUG:PIL.Image:Importing Hdf5StubImagePlugin
DEBUG:PIL.Image:Importing IcnsImagePlugin
DEBUG:PIL.Image:Importing IcoImagePlugin
DEBUG:PIL.Image:Importing ImImagePlugin
DEBUG:PIL.Image:Importing ImtImagePlugin
DEBUG:PIL.Image:Importing IptcImagePlugin
DEBUG:PIL.Image:Importing JpegImagePlugin
DEBUG:PIL.Image:Importing Jpeg2KImagePlugin
DEBUG:PIL.Image:Importing McIdasImagePlugin
DEBUG:PIL.Image:Importing MicImagePlugin
DEBUG:PIL.Image:Image: failed to import MicImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing MpegImagePlugin
DEBUG:PIL.Image:Importing MpoImagePlugin
DEBUG:PIL.Image:Importing MspImagePlugin
DEBUG:PIL.Image:Importing PalmImagePlugin
DEBUG:PIL.Image:Importing PcdImagePlugin
DEBUG:PIL.Image:Importing PcxImagePlugin
DEBUG:PIL.Image:Importing PdfImagePlugin
DEBUG:PIL.Image:Importing PixarImagePlugin
DEBUG:PIL.Image:Importing PngImagePlugin
DEBUG:PIL.Image:Importing PpmImagePlugin
DEBUG:PIL.Image:Importing PsdImagePlugin
DEBUG:PIL.Image:Importing QoiImagePlugin
DEBUG:PIL.Image:Importing SgiImagePlugin
DEBUG:PIL.Image:Importing SpiderImagePlugin
DEBUG:PIL.Image:Importing SunImagePlugin
DEBUG:PIL.Image:Importing TgaImagePlugin
DEBUG:PIL.Image:Importing TiffImagePlugin
DEBUG:PIL.Image:Importing WebPImagePlugin
DEBUG:PIL.Image:Importing WmfImagePlugin
DEBUG:PIL.Image:Importing XbmImagePlugin
DEBUG:PIL.Image:Importing XpmImagePlugin
DEBUG:PIL.Image:Importing XVThumbImagePlugin
12%|█▎ | 1/8 [01:16<08:54, 76.34s/it]
25%|██▌ | 2/8 [02:32<07:37, 76.28s/it]
38%|███▊ | 3/8 [03:48<06:21, 76.24s/it]
50%|█████ | 4/8 [05:04<05:04, 76.15s/it]
62%|██████▎ | 5/8 [06:21<03:48, 76.24s/it]
75%|███████▌ | 6/8 [07:36<02:32, 76.08s/it]
88%|████████▊ | 7/8 [08:52<01:15, 75.93s/it]
100%|██████████| 8/8 [10:09<00:00, 76.12s/it]
100%|██████████| 8/8 [10:09<00:00, 76.14s/it]
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 1: generating actions ...
>>> Step 1: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 2: generating actions ...
>>> Step 2: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 3: generating actions ...
>>> Step 3: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 4: generating actions ...
>>> Step 4: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 5: generating actions ...
>>> Step 5: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 6: generating actions ...
>>> Step 6: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 7: generating actions ...
>>> Step 7: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
real 11m45.279s
user 47m55.345s
sys 2m55.266s
2026-02-11 21:13:04.812376: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2026-02-11 21:13:04.862167: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2026-02-11 21:13:04.862223: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2026-02-11 21:13:04.863549: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2026-02-11 21:13:04.871078: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2026-02-11 21:13:05.785070: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Global seed set to 123
INFO:mainlogger:LatentVisualDiffusion: Running in v-prediction mode
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
AE working on z of shape (1, 4, 32, 32) = 4096 dimensions.
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
>>> model checkpoint loaded.
>>> Load pre-trained model ...
INFO:root:***** Configing Data *****
>>> unitree_z1_stackbox: 1 data samples loaded.
>>> unitree_z1_stackbox: data stats loaded.
>>> unitree_z1_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox: data stats loaded.
>>> unitree_z1_dual_arm_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox_v2: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox_v2: data stats loaded.
>>> unitree_z1_dual_arm_stackbox_v2: normalizer initiated.
>>> unitree_z1_dual_arm_cleanup_pencils: 1 data samples loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: data stats loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: normalizer initiated.
>>> unitree_g1_pack_camera: 1 data samples loaded.
>>> unitree_g1_pack_camera: data stats loaded.
>>> unitree_g1_pack_camera: normalizer initiated.
>>> Dataset is successfully loaded ...
>>> Generate 16 frames under each generation ...
DEBUG:h5py._conv:Creating converter from 3 to 5
DEBUG:PIL.PngImagePlugin:STREAM b'IHDR' 16 13
DEBUG:PIL.PngImagePlugin:STREAM b'pHYs' 41 9
DEBUG:PIL.PngImagePlugin:STREAM b'IDAT' 62 4096
0%| | 0/8 [00:00<?, ?it/s]>>> Step 0: generating actions ...
>>> Step 0: interacting with world model ...
DEBUG:PIL.Image:Importing BlpImagePlugin
DEBUG:PIL.Image:Importing BmpImagePlugin
DEBUG:PIL.Image:Importing BufrStubImagePlugin
DEBUG:PIL.Image:Importing CurImagePlugin
DEBUG:PIL.Image:Importing DcxImagePlugin
DEBUG:PIL.Image:Importing DdsImagePlugin
DEBUG:PIL.Image:Importing EpsImagePlugin
DEBUG:PIL.Image:Importing FitsImagePlugin
DEBUG:PIL.Image:Importing FitsStubImagePlugin
DEBUG:PIL.Image:Importing FliImagePlugin
DEBUG:PIL.Image:Importing FpxImagePlugin
DEBUG:PIL.Image:Image: failed to import FpxImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing FtexImagePlugin
DEBUG:PIL.Image:Importing GbrImagePlugin
DEBUG:PIL.Image:Importing GifImagePlugin
DEBUG:PIL.Image:Importing GribStubImagePlugin
DEBUG:PIL.Image:Importing Hdf5StubImagePlugin
DEBUG:PIL.Image:Importing IcnsImagePlugin
DEBUG:PIL.Image:Importing IcoImagePlugin
DEBUG:PIL.Image:Importing ImImagePlugin
DEBUG:PIL.Image:Importing ImtImagePlugin
DEBUG:PIL.Image:Importing IptcImagePlugin
DEBUG:PIL.Image:Importing JpegImagePlugin
DEBUG:PIL.Image:Importing Jpeg2KImagePlugin
DEBUG:PIL.Image:Importing McIdasImagePlugin
DEBUG:PIL.Image:Importing MicImagePlugin
DEBUG:PIL.Image:Image: failed to import MicImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing MpegImagePlugin
DEBUG:PIL.Image:Importing MpoImagePlugin
DEBUG:PIL.Image:Importing MspImagePlugin
DEBUG:PIL.Image:Importing PalmImagePlugin
DEBUG:PIL.Image:Importing PcdImagePlugin
DEBUG:PIL.Image:Importing PcxImagePlugin
DEBUG:PIL.Image:Importing PdfImagePlugin
DEBUG:PIL.Image:Importing PixarImagePlugin
DEBUG:PIL.Image:Importing PngImagePlugin
DEBUG:PIL.Image:Importing PpmImagePlugin
DEBUG:PIL.Image:Importing PsdImagePlugin
DEBUG:PIL.Image:Importing QoiImagePlugin
DEBUG:PIL.Image:Importing SgiImagePlugin
DEBUG:PIL.Image:Importing SpiderImagePlugin
DEBUG:PIL.Image:Importing SunImagePlugin
DEBUG:PIL.Image:Importing TgaImagePlugin
DEBUG:PIL.Image:Importing TiffImagePlugin
DEBUG:PIL.Image:Importing WebPImagePlugin
DEBUG:PIL.Image:Importing WmfImagePlugin
DEBUG:PIL.Image:Importing XbmImagePlugin
DEBUG:PIL.Image:Importing XpmImagePlugin
DEBUG:PIL.Image:Importing XVThumbImagePlugin
12%|█▎ | 1/8 [01:15<08:45, 75.11s/it]
25%|██▌ | 2/8 [02:30<07:31, 75.30s/it]
38%|███▊ | 3/8 [03:45<06:16, 75.32s/it]
50%|█████ | 4/8 [05:01<05:01, 75.29s/it]
62%|██████▎ | 5/8 [06:16<03:46, 75.38s/it]
75%|███████▌ | 6/8 [07:32<02:30, 75.48s/it]
88%|████████▊ | 7/8 [08:47<01:15, 75.39s/it]
100%|██████████| 8/8 [10:02<00:00, 75.30s/it]
100%|██████████| 8/8 [10:02<00:00, 75.33s/it]
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 1: generating actions ...
>>> Step 1: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 2: generating actions ...
>>> Step 2: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 3: generating actions ...
>>> Step 3: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 4: generating actions ...
>>> Step 4: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 5: generating actions ...
>>> Step 5: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 6: generating actions ...
>>> Step 6: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 7: generating actions ...
>>> Step 7: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
real 11m37.624s
user 48m24.273s
sys 0m41.670s
2026-02-11 21:24:42.443699: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2026-02-11 21:24:42.494143: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2026-02-11 21:24:42.494201: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2026-02-11 21:24:42.495506: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2026-02-11 21:24:42.503003: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2026-02-11 21:24:43.415898: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Global seed set to 123
INFO:mainlogger:LatentVisualDiffusion: Running in v-prediction mode
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
AE working on z of shape (1, 4, 32, 32) = 4096 dimensions.
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
>>> model checkpoint loaded.
>>> Load pre-trained model ...
INFO:root:***** Configing Data *****
>>> unitree_z1_stackbox: 1 data samples loaded.
>>> unitree_z1_stackbox: data stats loaded.
>>> unitree_z1_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox: data stats loaded.
>>> unitree_z1_dual_arm_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox_v2: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox_v2: data stats loaded.
>>> unitree_z1_dual_arm_stackbox_v2: normalizer initiated.
>>> unitree_z1_dual_arm_cleanup_pencils: 1 data samples loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: data stats loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: normalizer initiated.
>>> unitree_g1_pack_camera: 1 data samples loaded.
>>> unitree_g1_pack_camera: data stats loaded.
>>> unitree_g1_pack_camera: normalizer initiated.
>>> Dataset is successfully loaded ...
>>> Generate 16 frames under each generation ...
DEBUG:h5py._conv:Creating converter from 3 to 5
DEBUG:PIL.PngImagePlugin:STREAM b'IHDR' 16 13
DEBUG:PIL.PngImagePlugin:STREAM b'pHYs' 41 9
DEBUG:PIL.PngImagePlugin:STREAM b'IDAT' 62 4096
0%| | 0/8 [00:00<?, ?it/s]>>> Step 0: generating actions ...
>>> Step 0: interacting with world model ...
DEBUG:PIL.Image:Importing BlpImagePlugin
DEBUG:PIL.Image:Importing BmpImagePlugin
DEBUG:PIL.Image:Importing BufrStubImagePlugin
DEBUG:PIL.Image:Importing CurImagePlugin
DEBUG:PIL.Image:Importing DcxImagePlugin
DEBUG:PIL.Image:Importing DdsImagePlugin
DEBUG:PIL.Image:Importing EpsImagePlugin
DEBUG:PIL.Image:Importing FitsImagePlugin
DEBUG:PIL.Image:Importing FitsStubImagePlugin
DEBUG:PIL.Image:Importing FliImagePlugin
DEBUG:PIL.Image:Importing FpxImagePlugin
DEBUG:PIL.Image:Image: failed to import FpxImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing FtexImagePlugin
DEBUG:PIL.Image:Importing GbrImagePlugin
DEBUG:PIL.Image:Importing GifImagePlugin
DEBUG:PIL.Image:Importing GribStubImagePlugin
DEBUG:PIL.Image:Importing Hdf5StubImagePlugin
DEBUG:PIL.Image:Importing IcnsImagePlugin
DEBUG:PIL.Image:Importing IcoImagePlugin
DEBUG:PIL.Image:Importing ImImagePlugin
DEBUG:PIL.Image:Importing ImtImagePlugin
DEBUG:PIL.Image:Importing IptcImagePlugin
DEBUG:PIL.Image:Importing JpegImagePlugin
DEBUG:PIL.Image:Importing Jpeg2KImagePlugin
DEBUG:PIL.Image:Importing McIdasImagePlugin
DEBUG:PIL.Image:Importing MicImagePlugin
DEBUG:PIL.Image:Image: failed to import MicImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing MpegImagePlugin
DEBUG:PIL.Image:Importing MpoImagePlugin
DEBUG:PIL.Image:Importing MspImagePlugin
DEBUG:PIL.Image:Importing PalmImagePlugin
DEBUG:PIL.Image:Importing PcdImagePlugin
DEBUG:PIL.Image:Importing PcxImagePlugin
DEBUG:PIL.Image:Importing PdfImagePlugin
DEBUG:PIL.Image:Importing PixarImagePlugin
DEBUG:PIL.Image:Importing PngImagePlugin
DEBUG:PIL.Image:Importing PpmImagePlugin
DEBUG:PIL.Image:Importing PsdImagePlugin
DEBUG:PIL.Image:Importing QoiImagePlugin
DEBUG:PIL.Image:Importing SgiImagePlugin
DEBUG:PIL.Image:Importing SpiderImagePlugin
DEBUG:PIL.Image:Importing SunImagePlugin
DEBUG:PIL.Image:Importing TgaImagePlugin
DEBUG:PIL.Image:Importing TiffImagePlugin
DEBUG:PIL.Image:Importing WebPImagePlugin
DEBUG:PIL.Image:Importing WmfImagePlugin
DEBUG:PIL.Image:Importing XbmImagePlugin
DEBUG:PIL.Image:Importing XpmImagePlugin
DEBUG:PIL.Image:Importing XVThumbImagePlugin
12%|█▎ | 1/8 [01:15<08:46, 75.28s/it]
25%|██▌ | 2/8 [02:30<07:32, 75.34s/it]
38%|███▊ | 3/8 [03:45<06:15, 75.08s/it]
50%|█████ | 4/8 [04:59<04:58, 74.69s/it]
62%|██████▎ | 5/8 [06:13<03:43, 74.43s/it]
75%|███████▌ | 6/8 [07:27<02:28, 74.27s/it]
88%|████████▊ | 7/8 [08:41<01:14, 74.21s/it]
100%|██████████| 8/8 [09:55<00:00, 74.13s/it]
100%|██████████| 8/8 [09:55<00:00, 74.43s/it]
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 1: generating actions ...
>>> Step 1: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 2: generating actions ...
>>> Step 2: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 3: generating actions ...
>>> Step 3: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 4: generating actions ...
>>> Step 4: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 5: generating actions ...
>>> Step 5: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 6: generating actions ...
>>> Step 6: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 7: generating actions ...
>>> Step 7: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
real 11m32.306s
user 48m37.680s
sys 0m41.716s
2026-02-11 21:36:14.761055: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2026-02-11 21:36:14.811056: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2026-02-11 21:36:14.811115: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2026-02-11 21:36:14.812480: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2026-02-11 21:36:14.820115: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2026-02-11 21:36:15.736583: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Global seed set to 123
INFO:mainlogger:LatentVisualDiffusion: Running in v-prediction mode
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
AE working on z of shape (1, 4, 32, 32) = 4096 dimensions.
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
>>> model checkpoint loaded.
>>> Load pre-trained model ...
INFO:root:***** Configing Data *****
>>> unitree_z1_stackbox: 1 data samples loaded.
>>> unitree_z1_stackbox: data stats loaded.
>>> unitree_z1_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox: data stats loaded.
>>> unitree_z1_dual_arm_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox_v2: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox_v2: data stats loaded.
>>> unitree_z1_dual_arm_stackbox_v2: normalizer initiated.
>>> unitree_z1_dual_arm_cleanup_pencils: 1 data samples loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: data stats loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: normalizer initiated.
>>> unitree_g1_pack_camera: 1 data samples loaded.
>>> unitree_g1_pack_camera: data stats loaded.
>>> unitree_g1_pack_camera: normalizer initiated.
>>> Dataset is successfully loaded ...
>>> Generate 16 frames under each generation ...
DEBUG:h5py._conv:Creating converter from 3 to 5
DEBUG:PIL.PngImagePlugin:STREAM b'IHDR' 16 13
DEBUG:PIL.PngImagePlugin:STREAM b'pHYs' 41 9
DEBUG:PIL.PngImagePlugin:STREAM b'IDAT' 62 4096
0%| | 0/7 [00:00<?, ?it/s]>>> Step 0: generating actions ...
>>> Step 0: interacting with world model ...
DEBUG:PIL.Image:Importing BlpImagePlugin
DEBUG:PIL.Image:Importing BmpImagePlugin
DEBUG:PIL.Image:Importing BufrStubImagePlugin
DEBUG:PIL.Image:Importing CurImagePlugin
DEBUG:PIL.Image:Importing DcxImagePlugin
DEBUG:PIL.Image:Importing DdsImagePlugin
DEBUG:PIL.Image:Importing EpsImagePlugin
DEBUG:PIL.Image:Importing FitsImagePlugin
DEBUG:PIL.Image:Importing FitsStubImagePlugin
DEBUG:PIL.Image:Importing FliImagePlugin
DEBUG:PIL.Image:Importing FpxImagePlugin
DEBUG:PIL.Image:Image: failed to import FpxImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing FtexImagePlugin
DEBUG:PIL.Image:Importing GbrImagePlugin
DEBUG:PIL.Image:Importing GifImagePlugin
DEBUG:PIL.Image:Importing GribStubImagePlugin
DEBUG:PIL.Image:Importing Hdf5StubImagePlugin
DEBUG:PIL.Image:Importing IcnsImagePlugin
DEBUG:PIL.Image:Importing IcoImagePlugin
DEBUG:PIL.Image:Importing ImImagePlugin
DEBUG:PIL.Image:Importing ImtImagePlugin
DEBUG:PIL.Image:Importing IptcImagePlugin
DEBUG:PIL.Image:Importing JpegImagePlugin
DEBUG:PIL.Image:Importing Jpeg2KImagePlugin
DEBUG:PIL.Image:Importing McIdasImagePlugin
DEBUG:PIL.Image:Importing MicImagePlugin
DEBUG:PIL.Image:Image: failed to import MicImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing MpegImagePlugin
DEBUG:PIL.Image:Importing MpoImagePlugin
DEBUG:PIL.Image:Importing MspImagePlugin
DEBUG:PIL.Image:Importing PalmImagePlugin
DEBUG:PIL.Image:Importing PcdImagePlugin
DEBUG:PIL.Image:Importing PcxImagePlugin
DEBUG:PIL.Image:Importing PdfImagePlugin
DEBUG:PIL.Image:Importing PixarImagePlugin
DEBUG:PIL.Image:Importing PngImagePlugin
DEBUG:PIL.Image:Importing PpmImagePlugin
DEBUG:PIL.Image:Importing PsdImagePlugin
DEBUG:PIL.Image:Importing QoiImagePlugin
DEBUG:PIL.Image:Importing SgiImagePlugin
DEBUG:PIL.Image:Importing SpiderImagePlugin
DEBUG:PIL.Image:Importing SunImagePlugin
DEBUG:PIL.Image:Importing TgaImagePlugin
DEBUG:PIL.Image:Importing TiffImagePlugin
DEBUG:PIL.Image:Importing WebPImagePlugin
DEBUG:PIL.Image:Importing WmfImagePlugin
DEBUG:PIL.Image:Importing XbmImagePlugin
DEBUG:PIL.Image:Importing XpmImagePlugin
DEBUG:PIL.Image:Importing XVThumbImagePlugin
14%|█▍ | 1/7 [01:15<07:34, 75.70s/it]
29%|██▊ | 2/7 [02:31<06:18, 75.65s/it]
43%|████▎ | 3/7 [03:46<05:02, 75.52s/it]
57%|█████▋ | 4/7 [05:02<03:46, 75.47s/it]
71%|███████▏ | 5/7 [06:17<02:30, 75.40s/it]
86%|████████▌ | 6/7 [07:32<01:15, 75.37s/it]
100%|██████████| 7/7 [08:48<00:00, 75.38s/it]
100%|██████████| 7/7 [08:48<00:00, 75.44s/it]
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 1: generating actions ...
>>> Step 1: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 2: generating actions ...
>>> Step 2: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 3: generating actions ...
>>> Step 3: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 4: generating actions ...
>>> Step 4: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 5: generating actions ...
>>> Step 5: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 6: generating actions ...
>>> Step 6: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
real 10m26.533s
user 43m2.531s
sys 0m41.416s
2026-02-11 21:46:41.375935: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2026-02-11 21:46:41.426557: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2026-02-11 21:46:41.426614: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2026-02-11 21:46:41.427937: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2026-02-11 21:46:41.435507: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2026-02-11 21:46:42.361310: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Global seed set to 123
INFO:mainlogger:LatentVisualDiffusion: Running in v-prediction mode
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
AE working on z of shape (1, 4, 32, 32) = 4096 dimensions.
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
>>> model checkpoint loaded.
>>> Load pre-trained model ...
INFO:root:***** Configing Data *****
>>> unitree_z1_stackbox: 1 data samples loaded.
>>> unitree_z1_stackbox: data stats loaded.
>>> unitree_z1_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox: data stats loaded.
>>> unitree_z1_dual_arm_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox_v2: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox_v2: data stats loaded.
>>> unitree_z1_dual_arm_stackbox_v2: normalizer initiated.
>>> unitree_z1_dual_arm_cleanup_pencils: 1 data samples loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: data stats loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: normalizer initiated.
>>> unitree_g1_pack_camera: 1 data samples loaded.
>>> unitree_g1_pack_camera: data stats loaded.
>>> unitree_g1_pack_camera: normalizer initiated.
>>> Dataset is successfully loaded ...
>>> Generate 16 frames under each generation ...
DEBUG:h5py._conv:Creating converter from 3 to 5
DEBUG:PIL.PngImagePlugin:STREAM b'IHDR' 16 13
DEBUG:PIL.PngImagePlugin:STREAM b'pHYs' 41 9
DEBUG:PIL.PngImagePlugin:STREAM b'IDAT' 62 4096
0%| | 0/7 [00:00<?, ?it/s]>>> Step 0: generating actions ...
>>> Step 0: interacting with world model ...
DEBUG:PIL.Image:Importing BlpImagePlugin
DEBUG:PIL.Image:Importing BmpImagePlugin
DEBUG:PIL.Image:Importing BufrStubImagePlugin
DEBUG:PIL.Image:Importing CurImagePlugin
DEBUG:PIL.Image:Importing DcxImagePlugin
DEBUG:PIL.Image:Importing DdsImagePlugin
DEBUG:PIL.Image:Importing EpsImagePlugin
DEBUG:PIL.Image:Importing FitsImagePlugin
DEBUG:PIL.Image:Importing FitsStubImagePlugin
DEBUG:PIL.Image:Importing FliImagePlugin
DEBUG:PIL.Image:Importing FpxImagePlugin
DEBUG:PIL.Image:Image: failed to import FpxImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing FtexImagePlugin
DEBUG:PIL.Image:Importing GbrImagePlugin
DEBUG:PIL.Image:Importing GifImagePlugin
DEBUG:PIL.Image:Importing GribStubImagePlugin
DEBUG:PIL.Image:Importing Hdf5StubImagePlugin
DEBUG:PIL.Image:Importing IcnsImagePlugin
DEBUG:PIL.Image:Importing IcoImagePlugin
DEBUG:PIL.Image:Importing ImImagePlugin
DEBUG:PIL.Image:Importing ImtImagePlugin
DEBUG:PIL.Image:Importing IptcImagePlugin
DEBUG:PIL.Image:Importing JpegImagePlugin
DEBUG:PIL.Image:Importing Jpeg2KImagePlugin
DEBUG:PIL.Image:Importing McIdasImagePlugin
DEBUG:PIL.Image:Importing MicImagePlugin
DEBUG:PIL.Image:Image: failed to import MicImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing MpegImagePlugin
DEBUG:PIL.Image:Importing MpoImagePlugin
DEBUG:PIL.Image:Importing MspImagePlugin
DEBUG:PIL.Image:Importing PalmImagePlugin
DEBUG:PIL.Image:Importing PcdImagePlugin
DEBUG:PIL.Image:Importing PcxImagePlugin
DEBUG:PIL.Image:Importing PdfImagePlugin
DEBUG:PIL.Image:Importing PixarImagePlugin
DEBUG:PIL.Image:Importing PngImagePlugin
DEBUG:PIL.Image:Importing PpmImagePlugin
DEBUG:PIL.Image:Importing PsdImagePlugin
DEBUG:PIL.Image:Importing QoiImagePlugin
DEBUG:PIL.Image:Importing SgiImagePlugin
DEBUG:PIL.Image:Importing SpiderImagePlugin
DEBUG:PIL.Image:Importing SunImagePlugin
DEBUG:PIL.Image:Importing TgaImagePlugin
DEBUG:PIL.Image:Importing TiffImagePlugin
DEBUG:PIL.Image:Importing WebPImagePlugin
DEBUG:PIL.Image:Importing WmfImagePlugin
DEBUG:PIL.Image:Importing XbmImagePlugin
DEBUG:PIL.Image:Importing XpmImagePlugin
DEBUG:PIL.Image:Importing XVThumbImagePlugin
14%|█▍ | 1/7 [01:16<07:38, 76.39s/it]
29%|██▊ | 2/7 [02:33<06:23, 76.69s/it]
43%|████▎ | 3/7 [03:50<05:07, 76.87s/it]
57%|█████▋ | 4/7 [05:07<03:50, 76.91s/it]
71%|███████▏ | 5/7 [06:23<02:33, 76.80s/it]
86%|████████▌ | 6/7 [07:40<01:16, 76.77s/it]
100%|██████████| 7/7 [08:57<00:00, 76.85s/it]
100%|██████████| 7/7 [08:57<00:00, 76.81s/it]
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 1: generating actions ...
>>> Step 1: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 2: generating actions ...
>>> Step 2: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 3: generating actions ...
>>> Step 3: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 4: generating actions ...
>>> Step 4: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 5: generating actions ...
>>> Step 5: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 6: generating actions ...
>>> Step 6: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
real 10m36.315s
user 42m52.218s
sys 0m42.198s
2026-02-11 21:57:17.623993: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2026-02-11 21:57:17.673835: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2026-02-11 21:57:17.673891: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2026-02-11 21:57:17.675211: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2026-02-11 21:57:17.682716: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2026-02-11 21:57:18.593525: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Global seed set to 123
INFO:mainlogger:LatentVisualDiffusion: Running in v-prediction mode
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
AE working on z of shape (1, 4, 32, 32) = 4096 dimensions.
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
>>> model checkpoint loaded.
>>> Load pre-trained model ...
INFO:root:***** Configing Data *****
>>> unitree_z1_stackbox: 1 data samples loaded.
>>> unitree_z1_stackbox: data stats loaded.
>>> unitree_z1_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox: data stats loaded.
>>> unitree_z1_dual_arm_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox_v2: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox_v2: data stats loaded.
>>> unitree_z1_dual_arm_stackbox_v2: normalizer initiated.
>>> unitree_z1_dual_arm_cleanup_pencils: 1 data samples loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: data stats loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: normalizer initiated.
>>> unitree_g1_pack_camera: 1 data samples loaded.
>>> unitree_g1_pack_camera: data stats loaded.
>>> unitree_g1_pack_camera: normalizer initiated.
>>> Dataset is successfully loaded ...
>>> Generate 16 frames under each generation ...
DEBUG:h5py._conv:Creating converter from 3 to 5
DEBUG:PIL.PngImagePlugin:STREAM b'IHDR' 16 13
DEBUG:PIL.PngImagePlugin:STREAM b'pHYs' 41 9
DEBUG:PIL.PngImagePlugin:STREAM b'IDAT' 62 4096
0%| | 0/7 [00:00<?, ?it/s]>>> Step 0: generating actions ...
>>> Step 0: interacting with world model ...
DEBUG:PIL.Image:Importing BlpImagePlugin
DEBUG:PIL.Image:Importing BmpImagePlugin
DEBUG:PIL.Image:Importing BufrStubImagePlugin
DEBUG:PIL.Image:Importing CurImagePlugin
DEBUG:PIL.Image:Importing DcxImagePlugin
DEBUG:PIL.Image:Importing DdsImagePlugin
DEBUG:PIL.Image:Importing EpsImagePlugin
DEBUG:PIL.Image:Importing FitsImagePlugin
DEBUG:PIL.Image:Importing FitsStubImagePlugin
DEBUG:PIL.Image:Importing FliImagePlugin
DEBUG:PIL.Image:Importing FpxImagePlugin
DEBUG:PIL.Image:Image: failed to import FpxImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing FtexImagePlugin
DEBUG:PIL.Image:Importing GbrImagePlugin
DEBUG:PIL.Image:Importing GifImagePlugin
DEBUG:PIL.Image:Importing GribStubImagePlugin
DEBUG:PIL.Image:Importing Hdf5StubImagePlugin
DEBUG:PIL.Image:Importing IcnsImagePlugin
DEBUG:PIL.Image:Importing IcoImagePlugin
DEBUG:PIL.Image:Importing ImImagePlugin
DEBUG:PIL.Image:Importing ImtImagePlugin
DEBUG:PIL.Image:Importing IptcImagePlugin
DEBUG:PIL.Image:Importing JpegImagePlugin
DEBUG:PIL.Image:Importing Jpeg2KImagePlugin
DEBUG:PIL.Image:Importing McIdasImagePlugin
DEBUG:PIL.Image:Importing MicImagePlugin
DEBUG:PIL.Image:Image: failed to import MicImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing MpegImagePlugin
DEBUG:PIL.Image:Importing MpoImagePlugin
DEBUG:PIL.Image:Importing MspImagePlugin
DEBUG:PIL.Image:Importing PalmImagePlugin
DEBUG:PIL.Image:Importing PcdImagePlugin
DEBUG:PIL.Image:Importing PcxImagePlugin
DEBUG:PIL.Image:Importing PdfImagePlugin
DEBUG:PIL.Image:Importing PixarImagePlugin
DEBUG:PIL.Image:Importing PngImagePlugin
DEBUG:PIL.Image:Importing PpmImagePlugin
DEBUG:PIL.Image:Importing PsdImagePlugin
DEBUG:PIL.Image:Importing QoiImagePlugin
DEBUG:PIL.Image:Importing SgiImagePlugin
DEBUG:PIL.Image:Importing SpiderImagePlugin
DEBUG:PIL.Image:Importing SunImagePlugin
DEBUG:PIL.Image:Importing TgaImagePlugin
DEBUG:PIL.Image:Importing TiffImagePlugin
DEBUG:PIL.Image:Importing WebPImagePlugin
DEBUG:PIL.Image:Importing WmfImagePlugin
DEBUG:PIL.Image:Importing XbmImagePlugin
DEBUG:PIL.Image:Importing XpmImagePlugin
DEBUG:PIL.Image:Importing XVThumbImagePlugin
14%|█▍ | 1/7 [01:15<07:33, 75.59s/it]
29%|██▊ | 2/7 [02:31<06:17, 75.59s/it]
43%|████▎ | 3/7 [03:46<05:01, 75.44s/it]
57%|█████▋ | 4/7 [05:01<03:46, 75.39s/it]
71%|███████▏ | 5/7 [06:17<02:30, 75.35s/it]
86%|████████▌ | 6/7 [07:32<01:15, 75.32s/it]
100%|██████████| 7/7 [08:47<00:00, 75.24s/it]
100%|██████████| 7/7 [08:47<00:00, 75.34s/it]
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 1: generating actions ...
>>> Step 1: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 2: generating actions ...
>>> Step 2: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 3: generating actions ...
>>> Step 3: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 4: generating actions ...
>>> Step 4: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 5: generating actions ...
>>> Step 5: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 6: generating actions ...
>>> Step 6: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
real 10m25.792s
user 43m2.308s
sys 0m41.819s
2026-02-11 22:07:43.398736: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2026-02-11 22:07:43.448264: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2026-02-11 22:07:43.448321: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2026-02-11 22:07:43.449636: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2026-02-11 22:07:43.457127: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2026-02-11 22:07:44.370935: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Global seed set to 123
INFO:mainlogger:LatentVisualDiffusion: Running in v-prediction mode
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
AE working on z of shape (1, 4, 32, 32) = 4096 dimensions.
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
>>> model checkpoint loaded.
>>> Load pre-trained model ...
INFO:root:***** Configing Data *****
>>> unitree_z1_stackbox: 1 data samples loaded.
>>> unitree_z1_stackbox: data stats loaded.
>>> unitree_z1_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox: data stats loaded.
>>> unitree_z1_dual_arm_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox_v2: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox_v2: data stats loaded.
>>> unitree_z1_dual_arm_stackbox_v2: normalizer initiated.
>>> unitree_z1_dual_arm_cleanup_pencils: 1 data samples loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: data stats loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: normalizer initiated.
>>> unitree_g1_pack_camera: 1 data samples loaded.
>>> unitree_g1_pack_camera: data stats loaded.
>>> unitree_g1_pack_camera: normalizer initiated.
>>> Dataset is successfully loaded ...
>>> Generate 16 frames under each generation ...
DEBUG:h5py._conv:Creating converter from 3 to 5
DEBUG:PIL.PngImagePlugin:STREAM b'IHDR' 16 13
DEBUG:PIL.PngImagePlugin:STREAM b'pHYs' 41 9
DEBUG:PIL.PngImagePlugin:STREAM b'IDAT' 62 4096
0%| | 0/7 [00:00<?, ?it/s]>>> Step 0: generating actions ...
>>> Step 0: interacting with world model ...
DEBUG:PIL.Image:Importing BlpImagePlugin
DEBUG:PIL.Image:Importing BmpImagePlugin
DEBUG:PIL.Image:Importing BufrStubImagePlugin
DEBUG:PIL.Image:Importing CurImagePlugin
DEBUG:PIL.Image:Importing DcxImagePlugin
DEBUG:PIL.Image:Importing DdsImagePlugin
DEBUG:PIL.Image:Importing EpsImagePlugin
DEBUG:PIL.Image:Importing FitsImagePlugin
DEBUG:PIL.Image:Importing FitsStubImagePlugin
DEBUG:PIL.Image:Importing FliImagePlugin
DEBUG:PIL.Image:Importing FpxImagePlugin
DEBUG:PIL.Image:Image: failed to import FpxImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing FtexImagePlugin
DEBUG:PIL.Image:Importing GbrImagePlugin
DEBUG:PIL.Image:Importing GifImagePlugin
DEBUG:PIL.Image:Importing GribStubImagePlugin
DEBUG:PIL.Image:Importing Hdf5StubImagePlugin
DEBUG:PIL.Image:Importing IcnsImagePlugin
DEBUG:PIL.Image:Importing IcoImagePlugin
DEBUG:PIL.Image:Importing ImImagePlugin
DEBUG:PIL.Image:Importing ImtImagePlugin
DEBUG:PIL.Image:Importing IptcImagePlugin
DEBUG:PIL.Image:Importing JpegImagePlugin
DEBUG:PIL.Image:Importing Jpeg2KImagePlugin
DEBUG:PIL.Image:Importing McIdasImagePlugin
DEBUG:PIL.Image:Importing MicImagePlugin
DEBUG:PIL.Image:Image: failed to import MicImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing MpegImagePlugin
DEBUG:PIL.Image:Importing MpoImagePlugin
DEBUG:PIL.Image:Importing MspImagePlugin
DEBUG:PIL.Image:Importing PalmImagePlugin
DEBUG:PIL.Image:Importing PcdImagePlugin
DEBUG:PIL.Image:Importing PcxImagePlugin
DEBUG:PIL.Image:Importing PdfImagePlugin
DEBUG:PIL.Image:Importing PixarImagePlugin
DEBUG:PIL.Image:Importing PngImagePlugin
DEBUG:PIL.Image:Importing PpmImagePlugin
DEBUG:PIL.Image:Importing PsdImagePlugin
DEBUG:PIL.Image:Importing QoiImagePlugin
DEBUG:PIL.Image:Importing SgiImagePlugin
DEBUG:PIL.Image:Importing SpiderImagePlugin
DEBUG:PIL.Image:Importing SunImagePlugin
DEBUG:PIL.Image:Importing TgaImagePlugin
DEBUG:PIL.Image:Importing TiffImagePlugin
DEBUG:PIL.Image:Importing WebPImagePlugin
DEBUG:PIL.Image:Importing WmfImagePlugin
DEBUG:PIL.Image:Importing XbmImagePlugin
DEBUG:PIL.Image:Importing XpmImagePlugin
DEBUG:PIL.Image:Importing XVThumbImagePlugin
14%|█▍ | 1/7 [01:17<07:42, 77.04s/it]
29%|██▊ | 2/7 [02:33<06:24, 76.95s/it]
43%|████▎ | 3/7 [03:50<05:07, 76.87s/it]
57%|█████▋ | 4/7 [05:06<03:49, 76.59s/it]
71%|███████▏ | 5/7 [06:24<02:33, 76.82s/it]
86%|████████▌ | 6/7 [07:39<01:16, 76.43s/it]
100%|██████████| 7/7 [08:55<00:00, 76.06s/it]
100%|██████████| 7/7 [08:55<00:00, 76.44s/it]
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 1: generating actions ...
>>> Step 1: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 2: generating actions ...
>>> Step 2: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 3: generating actions ...
>>> Step 3: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 4: generating actions ...
>>> Step 4: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 5: generating actions ...
>>> Step 5: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 6: generating actions ...
>>> Step 6: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
real 10m33.965s
user 42m57.756s
sys 0m41.772s
2026-02-11 22:18:17.396072: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2026-02-11 22:18:17.446095: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2026-02-11 22:18:17.446154: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2026-02-11 22:18:17.447480: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2026-02-11 22:18:17.455025: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2026-02-11 22:18:18.367007: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Global seed set to 123
INFO:mainlogger:LatentVisualDiffusion: Running in v-prediction mode
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
AE working on z of shape (1, 4, 32, 32) = 4096 dimensions.
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
>>> model checkpoint loaded.
>>> Load pre-trained model ...
INFO:root:***** Configing Data *****
>>> unitree_z1_stackbox: 1 data samples loaded.
>>> unitree_z1_stackbox: data stats loaded.
>>> unitree_z1_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox: data stats loaded.
>>> unitree_z1_dual_arm_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox_v2: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox_v2: data stats loaded.
>>> unitree_z1_dual_arm_stackbox_v2: normalizer initiated.
>>> unitree_z1_dual_arm_cleanup_pencils: 1 data samples loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: data stats loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: normalizer initiated.
>>> unitree_g1_pack_camera: 1 data samples loaded.
>>> unitree_g1_pack_camera: data stats loaded.
>>> unitree_g1_pack_camera: normalizer initiated.
>>> Dataset is successfully loaded ...
>>> Generate 16 frames under each generation ...
DEBUG:h5py._conv:Creating converter from 3 to 5
DEBUG:PIL.PngImagePlugin:STREAM b'IHDR' 16 13
DEBUG:PIL.PngImagePlugin:STREAM b'pHYs' 41 9
DEBUG:PIL.PngImagePlugin:STREAM b'IDAT' 62 4096
0%| | 0/11 [00:00<?, ?it/s]>>> Step 0: generating actions ...
>>> Step 0: interacting with world model ...
DEBUG:PIL.Image:Importing BlpImagePlugin
DEBUG:PIL.Image:Importing BmpImagePlugin
DEBUG:PIL.Image:Importing BufrStubImagePlugin
DEBUG:PIL.Image:Importing CurImagePlugin
DEBUG:PIL.Image:Importing DcxImagePlugin
DEBUG:PIL.Image:Importing DdsImagePlugin
DEBUG:PIL.Image:Importing EpsImagePlugin
DEBUG:PIL.Image:Importing FitsImagePlugin
DEBUG:PIL.Image:Importing FitsStubImagePlugin
DEBUG:PIL.Image:Importing FliImagePlugin
DEBUG:PIL.Image:Importing FpxImagePlugin
DEBUG:PIL.Image:Image: failed to import FpxImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing FtexImagePlugin
DEBUG:PIL.Image:Importing GbrImagePlugin
DEBUG:PIL.Image:Importing GifImagePlugin
DEBUG:PIL.Image:Importing GribStubImagePlugin
DEBUG:PIL.Image:Importing Hdf5StubImagePlugin
DEBUG:PIL.Image:Importing IcnsImagePlugin
DEBUG:PIL.Image:Importing IcoImagePlugin
DEBUG:PIL.Image:Importing ImImagePlugin
DEBUG:PIL.Image:Importing ImtImagePlugin
DEBUG:PIL.Image:Importing IptcImagePlugin
DEBUG:PIL.Image:Importing JpegImagePlugin
DEBUG:PIL.Image:Importing Jpeg2KImagePlugin
DEBUG:PIL.Image:Importing McIdasImagePlugin
DEBUG:PIL.Image:Importing MicImagePlugin
DEBUG:PIL.Image:Image: failed to import MicImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing MpegImagePlugin
DEBUG:PIL.Image:Importing MpoImagePlugin
DEBUG:PIL.Image:Importing MspImagePlugin
DEBUG:PIL.Image:Importing PalmImagePlugin
DEBUG:PIL.Image:Importing PcdImagePlugin
DEBUG:PIL.Image:Importing PcxImagePlugin
DEBUG:PIL.Image:Importing PdfImagePlugin
DEBUG:PIL.Image:Importing PixarImagePlugin
DEBUG:PIL.Image:Importing PngImagePlugin
DEBUG:PIL.Image:Importing PpmImagePlugin
DEBUG:PIL.Image:Importing PsdImagePlugin
DEBUG:PIL.Image:Importing QoiImagePlugin
DEBUG:PIL.Image:Importing SgiImagePlugin
DEBUG:PIL.Image:Importing SpiderImagePlugin
DEBUG:PIL.Image:Importing SunImagePlugin
DEBUG:PIL.Image:Importing TgaImagePlugin
DEBUG:PIL.Image:Importing TiffImagePlugin
DEBUG:PIL.Image:Importing WebPImagePlugin
DEBUG:PIL.Image:Importing WmfImagePlugin
DEBUG:PIL.Image:Importing XbmImagePlugin
DEBUG:PIL.Image:Importing XpmImagePlugin
DEBUG:PIL.Image:Importing XVThumbImagePlugin
9%|▉ | 1/11 [01:15<12:35, 75.53s/it]
18%|█▊ | 2/11 [02:30<11:18, 75.39s/it]
27%|██▋ | 3/11 [03:46<10:03, 75.38s/it]
36%|███▋ | 4/11 [05:01<08:48, 75.47s/it]
45%|████▌ | 5/11 [06:16<07:31, 75.32s/it]
55%|█████▍ | 6/11 [07:31<06:15, 75.08s/it]
64%|██████▎ | 7/11 [08:46<05:00, 75.07s/it]
73%|███████▎ | 8/11 [10:00<03:44, 74.76s/it]
82%|████████▏ | 9/11 [11:15<02:29, 74.87s/it]
91%|█████████ | 10/11 [12:30<01:14, 74.79s/it]
100%|██████████| 11/11 [13:45<00:00, 74.80s/it]
100%|██████████| 11/11 [13:45<00:00, 75.02s/it]
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 1: generating actions ...
>>> Step 1: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 2: generating actions ...
>>> Step 2: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 3: generating actions ...
>>> Step 3: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 4: generating actions ...
>>> Step 4: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 5: generating actions ...
>>> Step 5: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 6: generating actions ...
>>> Step 6: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 7: generating actions ...
>>> Step 7: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 8: generating actions ...
>>> Step 8: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 9: generating actions ...
>>> Step 9: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 10: generating actions ...
>>> Step 10: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
real 15m24.880s
user 65m40.817s
sys 0m41.445s
2026-02-11 22:33:42.261398: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2026-02-11 22:33:42.310786: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2026-02-11 22:33:42.310845: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2026-02-11 22:33:42.312191: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2026-02-11 22:33:42.319738: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2026-02-11 22:33:43.232517: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Global seed set to 123
INFO:mainlogger:LatentVisualDiffusion: Running in v-prediction mode
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
AE working on z of shape (1, 4, 32, 32) = 4096 dimensions.
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
>>> model checkpoint loaded.
>>> Load pre-trained model ...
INFO:root:***** Configing Data *****
>>> unitree_z1_stackbox: 1 data samples loaded.
>>> unitree_z1_stackbox: data stats loaded.
>>> unitree_z1_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox: data stats loaded.
>>> unitree_z1_dual_arm_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox_v2: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox_v2: data stats loaded.
>>> unitree_z1_dual_arm_stackbox_v2: normalizer initiated.
>>> unitree_z1_dual_arm_cleanup_pencils: 1 data samples loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: data stats loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: normalizer initiated.
>>> unitree_g1_pack_camera: 1 data samples loaded.
>>> unitree_g1_pack_camera: data stats loaded.
>>> unitree_g1_pack_camera: normalizer initiated.
>>> Dataset is successfully loaded ...
>>> Generate 16 frames under each generation ...
DEBUG:h5py._conv:Creating converter from 3 to 5
DEBUG:PIL.PngImagePlugin:STREAM b'IHDR' 16 13
DEBUG:PIL.PngImagePlugin:STREAM b'pHYs' 41 9
DEBUG:PIL.PngImagePlugin:STREAM b'IDAT' 62 4096
0%| | 0/11 [00:00<?, ?it/s]>>> Step 0: generating actions ...
>>> Step 0: interacting with world model ...
DEBUG:PIL.Image:Importing BlpImagePlugin
DEBUG:PIL.Image:Importing BmpImagePlugin
DEBUG:PIL.Image:Importing BufrStubImagePlugin
DEBUG:PIL.Image:Importing CurImagePlugin
DEBUG:PIL.Image:Importing DcxImagePlugin
DEBUG:PIL.Image:Importing DdsImagePlugin
DEBUG:PIL.Image:Importing EpsImagePlugin
DEBUG:PIL.Image:Importing FitsImagePlugin
DEBUG:PIL.Image:Importing FitsStubImagePlugin
DEBUG:PIL.Image:Importing FliImagePlugin
DEBUG:PIL.Image:Importing FpxImagePlugin
DEBUG:PIL.Image:Image: failed to import FpxImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing FtexImagePlugin
DEBUG:PIL.Image:Importing GbrImagePlugin
DEBUG:PIL.Image:Importing GifImagePlugin
DEBUG:PIL.Image:Importing GribStubImagePlugin
DEBUG:PIL.Image:Importing Hdf5StubImagePlugin
DEBUG:PIL.Image:Importing IcnsImagePlugin
DEBUG:PIL.Image:Importing IcoImagePlugin
DEBUG:PIL.Image:Importing ImImagePlugin
DEBUG:PIL.Image:Importing ImtImagePlugin
DEBUG:PIL.Image:Importing IptcImagePlugin
DEBUG:PIL.Image:Importing JpegImagePlugin
DEBUG:PIL.Image:Importing Jpeg2KImagePlugin
DEBUG:PIL.Image:Importing McIdasImagePlugin
DEBUG:PIL.Image:Importing MicImagePlugin
DEBUG:PIL.Image:Image: failed to import MicImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing MpegImagePlugin
DEBUG:PIL.Image:Importing MpoImagePlugin
DEBUG:PIL.Image:Importing MspImagePlugin
DEBUG:PIL.Image:Importing PalmImagePlugin
DEBUG:PIL.Image:Importing PcdImagePlugin
DEBUG:PIL.Image:Importing PcxImagePlugin
DEBUG:PIL.Image:Importing PdfImagePlugin
DEBUG:PIL.Image:Importing PixarImagePlugin
DEBUG:PIL.Image:Importing PngImagePlugin
DEBUG:PIL.Image:Importing PpmImagePlugin
DEBUG:PIL.Image:Importing PsdImagePlugin
DEBUG:PIL.Image:Importing QoiImagePlugin
DEBUG:PIL.Image:Importing SgiImagePlugin
DEBUG:PIL.Image:Importing SpiderImagePlugin
DEBUG:PIL.Image:Importing SunImagePlugin
DEBUG:PIL.Image:Importing TgaImagePlugin
DEBUG:PIL.Image:Importing TiffImagePlugin
DEBUG:PIL.Image:Importing WebPImagePlugin
DEBUG:PIL.Image:Importing WmfImagePlugin
DEBUG:PIL.Image:Importing XbmImagePlugin
DEBUG:PIL.Image:Importing XpmImagePlugin
DEBUG:PIL.Image:Importing XVThumbImagePlugin
9%|▉ | 1/11 [01:13<12:18, 73.90s/it]
18%|█▊ | 2/11 [02:27<11:05, 73.99s/it]
27%|██▋ | 3/11 [03:41<09:50, 73.86s/it]
36%|███▋ | 4/11 [04:55<08:35, 73.70s/it]
45%|████▌ | 5/11 [06:08<07:20, 73.48s/it]
55%|█████▍ | 6/11 [07:21<06:06, 73.39s/it]
64%|██████▎ | 7/11 [08:34<04:53, 73.28s/it]
73%|███████▎ | 8/11 [09:47<03:39, 73.11s/it]
82%|████████▏ | 9/11 [11:00<02:26, 73.21s/it]
91%|█████████ | 10/11 [12:14<01:13, 73.49s/it]
100%|██████████| 11/11 [13:28<00:00, 73.55s/it]
100%|██████████| 11/11 [13:28<00:00, 73.50s/it]
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 1: generating actions ...
>>> Step 1: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 2: generating actions ...
>>> Step 2: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 3: generating actions ...
>>> Step 3: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 4: generating actions ...
>>> Step 4: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 5: generating actions ...
>>> Step 5: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 6: generating actions ...
>>> Step 6: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 7: generating actions ...
>>> Step 7: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 8: generating actions ...
>>> Step 8: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 9: generating actions ...
>>> Step 9: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 10: generating actions ...
>>> Step 10: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
real 15m7.472s
user 65m58.841s
sys 0m41.599s
2026-02-11 22:48:49.761688: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2026-02-11 22:48:49.811395: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2026-02-11 22:48:49.811456: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2026-02-11 22:48:49.812798: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2026-02-11 22:48:49.820307: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2026-02-11 22:48:50.732941: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Global seed set to 123
INFO:mainlogger:LatentVisualDiffusion: Running in v-prediction mode
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
AE working on z of shape (1, 4, 32, 32) = 4096 dimensions.
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
>>> model checkpoint loaded.
>>> Load pre-trained model ...
INFO:root:***** Configing Data *****
>>> unitree_z1_stackbox: 1 data samples loaded.
>>> unitree_z1_stackbox: data stats loaded.
>>> unitree_z1_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox: data stats loaded.
>>> unitree_z1_dual_arm_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox_v2: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox_v2: data stats loaded.
>>> unitree_z1_dual_arm_stackbox_v2: normalizer initiated.
>>> unitree_z1_dual_arm_cleanup_pencils: 1 data samples loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: data stats loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: normalizer initiated.
>>> unitree_g1_pack_camera: 1 data samples loaded.
>>> unitree_g1_pack_camera: data stats loaded.
>>> unitree_g1_pack_camera: normalizer initiated.
>>> Dataset is successfully loaded ...
>>> Generate 16 frames under each generation ...
DEBUG:h5py._conv:Creating converter from 3 to 5
DEBUG:PIL.PngImagePlugin:STREAM b'IHDR' 16 13
DEBUG:PIL.PngImagePlugin:STREAM b'pHYs' 41 9
DEBUG:PIL.PngImagePlugin:STREAM b'IDAT' 62 4096
0%| | 0/11 [00:00<?, ?it/s]>>> Step 0: generating actions ...
>>> Step 0: interacting with world model ...
DEBUG:PIL.Image:Importing BlpImagePlugin
DEBUG:PIL.Image:Importing BmpImagePlugin
DEBUG:PIL.Image:Importing BufrStubImagePlugin
DEBUG:PIL.Image:Importing CurImagePlugin
DEBUG:PIL.Image:Importing DcxImagePlugin
DEBUG:PIL.Image:Importing DdsImagePlugin
DEBUG:PIL.Image:Importing EpsImagePlugin
DEBUG:PIL.Image:Importing FitsImagePlugin
DEBUG:PIL.Image:Importing FitsStubImagePlugin
DEBUG:PIL.Image:Importing FliImagePlugin
DEBUG:PIL.Image:Importing FpxImagePlugin
DEBUG:PIL.Image:Image: failed to import FpxImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing FtexImagePlugin
DEBUG:PIL.Image:Importing GbrImagePlugin
DEBUG:PIL.Image:Importing GifImagePlugin
DEBUG:PIL.Image:Importing GribStubImagePlugin
DEBUG:PIL.Image:Importing Hdf5StubImagePlugin
DEBUG:PIL.Image:Importing IcnsImagePlugin
DEBUG:PIL.Image:Importing IcoImagePlugin
DEBUG:PIL.Image:Importing ImImagePlugin
DEBUG:PIL.Image:Importing ImtImagePlugin
DEBUG:PIL.Image:Importing IptcImagePlugin
DEBUG:PIL.Image:Importing JpegImagePlugin
DEBUG:PIL.Image:Importing Jpeg2KImagePlugin
DEBUG:PIL.Image:Importing McIdasImagePlugin
DEBUG:PIL.Image:Importing MicImagePlugin
DEBUG:PIL.Image:Image: failed to import MicImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing MpegImagePlugin
DEBUG:PIL.Image:Importing MpoImagePlugin
DEBUG:PIL.Image:Importing MspImagePlugin
DEBUG:PIL.Image:Importing PalmImagePlugin
DEBUG:PIL.Image:Importing PcdImagePlugin
DEBUG:PIL.Image:Importing PcxImagePlugin
DEBUG:PIL.Image:Importing PdfImagePlugin
DEBUG:PIL.Image:Importing PixarImagePlugin
DEBUG:PIL.Image:Importing PngImagePlugin
DEBUG:PIL.Image:Importing PpmImagePlugin
DEBUG:PIL.Image:Importing PsdImagePlugin
DEBUG:PIL.Image:Importing QoiImagePlugin
DEBUG:PIL.Image:Importing SgiImagePlugin
DEBUG:PIL.Image:Importing SpiderImagePlugin
DEBUG:PIL.Image:Importing SunImagePlugin
DEBUG:PIL.Image:Importing TgaImagePlugin
DEBUG:PIL.Image:Importing TiffImagePlugin
DEBUG:PIL.Image:Importing WebPImagePlugin
DEBUG:PIL.Image:Importing WmfImagePlugin
DEBUG:PIL.Image:Importing XbmImagePlugin
DEBUG:PIL.Image:Importing XpmImagePlugin
DEBUG:PIL.Image:Importing XVThumbImagePlugin
9%|▉ | 1/11 [01:14<12:29, 74.99s/it]
18%|█▊ | 2/11 [02:30<11:18, 75.44s/it]
27%|██▋ | 3/11 [03:46<10:03, 75.49s/it]
36%|███▋ | 4/11 [05:01<08:47, 75.30s/it]
45%|████▌ | 5/11 [06:15<07:30, 75.02s/it]
55%|█████▍ | 6/11 [07:30<06:14, 74.84s/it]
64%|██████▎ | 7/11 [08:44<04:59, 74.79s/it]
73%|███████▎ | 8/11 [10:00<03:44, 74.94s/it]
82%|████████▏ | 9/11 [11:15<02:29, 74.89s/it]
91%|█████████ | 10/11 [12:30<01:14, 74.95s/it]
100%|██████████| 11/11 [13:45<00:00, 74.95s/it]
100%|██████████| 11/11 [13:45<00:00, 75.01s/it]
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 1: generating actions ...
>>> Step 1: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 2: generating actions ...
>>> Step 2: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 3: generating actions ...
>>> Step 3: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 4: generating actions ...
>>> Step 4: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 5: generating actions ...
>>> Step 5: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 6: generating actions ...
>>> Step 6: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 7: generating actions ...
>>> Step 7: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 8: generating actions ...
>>> Step 8: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 9: generating actions ...
>>> Step 9: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 10: generating actions ...
>>> Step 10: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
real 15m26.006s
user 65m35.774s
sys 0m44.636s
2026-02-11 23:04:15.762959: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2026-02-11 23:04:15.814243: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2026-02-11 23:04:15.814301: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2026-02-11 23:04:15.815653: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2026-02-11 23:04:15.823287: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2026-02-11 23:04:16.742609: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Global seed set to 123
INFO:mainlogger:LatentVisualDiffusion: Running in v-prediction mode
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
AE working on z of shape (1, 4, 32, 32) = 4096 dimensions.
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
>>> model checkpoint loaded.
>>> Load pre-trained model ...
INFO:root:***** Configing Data *****
>>> unitree_z1_stackbox: 1 data samples loaded.
>>> unitree_z1_stackbox: data stats loaded.
>>> unitree_z1_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox: data stats loaded.
>>> unitree_z1_dual_arm_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox_v2: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox_v2: data stats loaded.
>>> unitree_z1_dual_arm_stackbox_v2: normalizer initiated.
>>> unitree_z1_dual_arm_cleanup_pencils: 1 data samples loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: data stats loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: normalizer initiated.
>>> unitree_g1_pack_camera: 1 data samples loaded.
>>> unitree_g1_pack_camera: data stats loaded.
>>> unitree_g1_pack_camera: normalizer initiated.
>>> Dataset is successfully loaded ...
>>> Generate 16 frames under each generation ...
DEBUG:h5py._conv:Creating converter from 3 to 5
DEBUG:PIL.PngImagePlugin:STREAM b'IHDR' 16 13
DEBUG:PIL.PngImagePlugin:STREAM b'pHYs' 41 9
DEBUG:PIL.PngImagePlugin:STREAM b'IDAT' 62 4096
0%| | 0/11 [00:00<?, ?it/s]>>> Step 0: generating actions ...
>>> Step 0: interacting with world model ...
DEBUG:PIL.Image:Importing BlpImagePlugin
DEBUG:PIL.Image:Importing BmpImagePlugin
DEBUG:PIL.Image:Importing BufrStubImagePlugin
DEBUG:PIL.Image:Importing CurImagePlugin
DEBUG:PIL.Image:Importing DcxImagePlugin
DEBUG:PIL.Image:Importing DdsImagePlugin
DEBUG:PIL.Image:Importing EpsImagePlugin
DEBUG:PIL.Image:Importing FitsImagePlugin
DEBUG:PIL.Image:Importing FitsStubImagePlugin
DEBUG:PIL.Image:Importing FliImagePlugin
DEBUG:PIL.Image:Importing FpxImagePlugin
DEBUG:PIL.Image:Image: failed to import FpxImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing FtexImagePlugin
DEBUG:PIL.Image:Importing GbrImagePlugin
DEBUG:PIL.Image:Importing GifImagePlugin
DEBUG:PIL.Image:Importing GribStubImagePlugin
DEBUG:PIL.Image:Importing Hdf5StubImagePlugin
DEBUG:PIL.Image:Importing IcnsImagePlugin
DEBUG:PIL.Image:Importing IcoImagePlugin
DEBUG:PIL.Image:Importing ImImagePlugin
DEBUG:PIL.Image:Importing ImtImagePlugin
DEBUG:PIL.Image:Importing IptcImagePlugin
DEBUG:PIL.Image:Importing JpegImagePlugin
DEBUG:PIL.Image:Importing Jpeg2KImagePlugin
DEBUG:PIL.Image:Importing McIdasImagePlugin
DEBUG:PIL.Image:Importing MicImagePlugin
DEBUG:PIL.Image:Image: failed to import MicImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing MpegImagePlugin
DEBUG:PIL.Image:Importing MpoImagePlugin
DEBUG:PIL.Image:Importing MspImagePlugin
DEBUG:PIL.Image:Importing PalmImagePlugin
DEBUG:PIL.Image:Importing PcdImagePlugin
DEBUG:PIL.Image:Importing PcxImagePlugin
DEBUG:PIL.Image:Importing PdfImagePlugin
DEBUG:PIL.Image:Importing PixarImagePlugin
DEBUG:PIL.Image:Importing PngImagePlugin
DEBUG:PIL.Image:Importing PpmImagePlugin
DEBUG:PIL.Image:Importing PsdImagePlugin
DEBUG:PIL.Image:Importing QoiImagePlugin
DEBUG:PIL.Image:Importing SgiImagePlugin
DEBUG:PIL.Image:Importing SpiderImagePlugin
DEBUG:PIL.Image:Importing SunImagePlugin
DEBUG:PIL.Image:Importing TgaImagePlugin
DEBUG:PIL.Image:Importing TiffImagePlugin
DEBUG:PIL.Image:Importing WebPImagePlugin
DEBUG:PIL.Image:Importing WmfImagePlugin
DEBUG:PIL.Image:Importing XbmImagePlugin
DEBUG:PIL.Image:Importing XpmImagePlugin
DEBUG:PIL.Image:Importing XVThumbImagePlugin
9%|▉ | 1/11 [01:14<12:25, 74.55s/it]
18%|█▊ | 2/11 [02:29<11:13, 74.83s/it]
27%|██▋ | 3/11 [03:44<09:58, 74.77s/it]
36%|███▋ | 4/11 [04:59<08:43, 74.78s/it]
45%|████▌ | 5/11 [06:13<07:28, 74.83s/it]
55%|█████▍ | 6/11 [07:28<06:13, 74.62s/it]
64%|██████▎ | 7/11 [08:42<04:57, 74.37s/it]
73%|███████▎ | 8/11 [09:56<03:42, 74.28s/it]
82%|████████▏ | 9/11 [11:10<02:28, 74.26s/it]
91%|█████████ | 10/11 [12:24<01:14, 74.31s/it]
100%|██████████| 11/11 [13:39<00:00, 74.38s/it]
100%|██████████| 11/11 [13:39<00:00, 74.48s/it]
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 1: generating actions ...
>>> Step 1: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 2: generating actions ...
>>> Step 2: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 3: generating actions ...
>>> Step 3: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 4: generating actions ...
>>> Step 4: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 5: generating actions ...
>>> Step 5: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 6: generating actions ...
>>> Step 6: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 7: generating actions ...
>>> Step 7: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 8: generating actions ...
>>> Step 8: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 9: generating actions ...
>>> Step 9: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 10: generating actions ...
>>> Step 10: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
real 15m20.607s
user 66m10.462s
sys 0m45.655s
2026-02-11 23:19:36.475817: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2026-02-11 23:19:36.525118: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2026-02-11 23:19:36.525172: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2026-02-11 23:19:36.526479: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2026-02-11 23:19:36.533981: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2026-02-11 23:19:37.461985: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Global seed set to 123
INFO:mainlogger:LatentVisualDiffusion: Running in v-prediction mode
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
AE working on z of shape (1, 4, 32, 32) = 4096 dimensions.
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
>>> model checkpoint loaded.
>>> Load pre-trained model ...
INFO:root:***** Configing Data *****
>>> unitree_z1_stackbox: 1 data samples loaded.
>>> unitree_z1_stackbox: data stats loaded.
>>> unitree_z1_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox: data stats loaded.
>>> unitree_z1_dual_arm_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox_v2: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox_v2: data stats loaded.
>>> unitree_z1_dual_arm_stackbox_v2: normalizer initiated.
>>> unitree_z1_dual_arm_cleanup_pencils: 1 data samples loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: data stats loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: normalizer initiated.
>>> unitree_g1_pack_camera: 1 data samples loaded.
>>> unitree_g1_pack_camera: data stats loaded.
>>> unitree_g1_pack_camera: normalizer initiated.
>>> Dataset is successfully loaded ...
>>> Generate 16 frames under each generation ...
DEBUG:h5py._conv:Creating converter from 3 to 5
DEBUG:PIL.PngImagePlugin:STREAM b'IHDR' 16 13
DEBUG:PIL.PngImagePlugin:STREAM b'pHYs' 41 9
DEBUG:PIL.PngImagePlugin:STREAM b'IDAT' 62 4096
0%| | 0/12 [00:00<?, ?it/s]>>> Step 0: generating actions ...
>>> Step 0: interacting with world model ...
DEBUG:PIL.Image:Importing BlpImagePlugin
DEBUG:PIL.Image:Importing BmpImagePlugin
DEBUG:PIL.Image:Importing BufrStubImagePlugin
DEBUG:PIL.Image:Importing CurImagePlugin
DEBUG:PIL.Image:Importing DcxImagePlugin
DEBUG:PIL.Image:Importing DdsImagePlugin
DEBUG:PIL.Image:Importing EpsImagePlugin
DEBUG:PIL.Image:Importing FitsImagePlugin
DEBUG:PIL.Image:Importing FitsStubImagePlugin
DEBUG:PIL.Image:Importing FliImagePlugin
DEBUG:PIL.Image:Importing FpxImagePlugin
DEBUG:PIL.Image:Image: failed to import FpxImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing FtexImagePlugin
DEBUG:PIL.Image:Importing GbrImagePlugin
DEBUG:PIL.Image:Importing GifImagePlugin
DEBUG:PIL.Image:Importing GribStubImagePlugin
DEBUG:PIL.Image:Importing Hdf5StubImagePlugin
DEBUG:PIL.Image:Importing IcnsImagePlugin
DEBUG:PIL.Image:Importing IcoImagePlugin
DEBUG:PIL.Image:Importing ImImagePlugin
DEBUG:PIL.Image:Importing ImtImagePlugin
DEBUG:PIL.Image:Importing IptcImagePlugin
DEBUG:PIL.Image:Importing JpegImagePlugin
DEBUG:PIL.Image:Importing Jpeg2KImagePlugin
DEBUG:PIL.Image:Importing McIdasImagePlugin
DEBUG:PIL.Image:Importing MicImagePlugin
DEBUG:PIL.Image:Image: failed to import MicImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing MpegImagePlugin
DEBUG:PIL.Image:Importing MpoImagePlugin
DEBUG:PIL.Image:Importing MspImagePlugin
DEBUG:PIL.Image:Importing PalmImagePlugin
DEBUG:PIL.Image:Importing PcdImagePlugin
DEBUG:PIL.Image:Importing PcxImagePlugin
DEBUG:PIL.Image:Importing PdfImagePlugin
DEBUG:PIL.Image:Importing PixarImagePlugin
DEBUG:PIL.Image:Importing PngImagePlugin
DEBUG:PIL.Image:Importing PpmImagePlugin
DEBUG:PIL.Image:Importing PsdImagePlugin
DEBUG:PIL.Image:Importing QoiImagePlugin
DEBUG:PIL.Image:Importing SgiImagePlugin
DEBUG:PIL.Image:Importing SpiderImagePlugin
DEBUG:PIL.Image:Importing SunImagePlugin
DEBUG:PIL.Image:Importing TgaImagePlugin
DEBUG:PIL.Image:Importing TiffImagePlugin
DEBUG:PIL.Image:Importing WebPImagePlugin
DEBUG:PIL.Image:Importing WmfImagePlugin
DEBUG:PIL.Image:Importing XbmImagePlugin
DEBUG:PIL.Image:Importing XpmImagePlugin
DEBUG:PIL.Image:Importing XVThumbImagePlugin
8%|▊ | 1/12 [01:14<13:38, 74.37s/it]
17%|█▋ | 2/12 [02:28<12:21, 74.13s/it]
25%|██▌ | 3/12 [03:42<11:05, 73.97s/it]
33%|███▎ | 4/12 [04:55<09:51, 73.92s/it]
42%|████▏ | 5/12 [06:09<08:35, 73.70s/it]
50%|█████ | 6/12 [07:22<07:21, 73.58s/it]
58%|█████▊ | 7/12 [08:35<06:07, 73.49s/it]
67%|██████▋ | 8/12 [09:49<04:53, 73.46s/it]
75%|███████▌ | 9/12 [11:02<03:40, 73.42s/it]
83%|████████▎ | 10/12 [12:16<02:26, 73.44s/it]
92%|█████████▏| 11/12 [13:29<01:13, 73.40s/it]
100%|██████████| 12/12 [14:42<00:00, 73.32s/it]
100%|██████████| 12/12 [14:42<00:00, 73.55s/it]
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 1: generating actions ...
>>> Step 1: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 2: generating actions ...
>>> Step 2: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 3: generating actions ...
>>> Step 3: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 4: generating actions ...
>>> Step 4: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 5: generating actions ...
>>> Step 5: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 6: generating actions ...
>>> Step 6: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 7: generating actions ...
>>> Step 7: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 8: generating actions ...
>>> Step 8: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 9: generating actions ...
>>> Step 9: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 10: generating actions ...
>>> Step 10: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 11: generating actions ...
>>> Step 11: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
real 16m18.361s
user 71m2.856s
sys 0m44.193s
2026-02-11 23:35:52.961572: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2026-02-11 23:35:53.011562: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2026-02-11 23:35:53.011622: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2026-02-11 23:35:53.012961: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2026-02-11 23:35:53.020520: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2026-02-11 23:35:53.953177: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Global seed set to 123
INFO:mainlogger:LatentVisualDiffusion: Running in v-prediction mode
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
AE working on z of shape (1, 4, 32, 32) = 4096 dimensions.
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
>>> model checkpoint loaded.
>>> Load pre-trained model ...
INFO:root:***** Configing Data *****
>>> unitree_z1_stackbox: 1 data samples loaded.
>>> unitree_z1_stackbox: data stats loaded.
>>> unitree_z1_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox: data stats loaded.
>>> unitree_z1_dual_arm_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox_v2: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox_v2: data stats loaded.
>>> unitree_z1_dual_arm_stackbox_v2: normalizer initiated.
>>> unitree_z1_dual_arm_cleanup_pencils: 1 data samples loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: data stats loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: normalizer initiated.
>>> unitree_g1_pack_camera: 1 data samples loaded.
>>> unitree_g1_pack_camera: data stats loaded.
>>> unitree_g1_pack_camera: normalizer initiated.
>>> Dataset is successfully loaded ...
>>> Generate 16 frames under each generation ...
DEBUG:h5py._conv:Creating converter from 3 to 5
DEBUG:PIL.PngImagePlugin:STREAM b'IHDR' 16 13
DEBUG:PIL.PngImagePlugin:STREAM b'pHYs' 41 9
DEBUG:PIL.PngImagePlugin:STREAM b'IDAT' 62 4096
0%| | 0/12 [00:00<?, ?it/s]>>> Step 0: generating actions ...
>>> Step 0: interacting with world model ...
DEBUG:PIL.Image:Importing BlpImagePlugin
DEBUG:PIL.Image:Importing BmpImagePlugin
DEBUG:PIL.Image:Importing BufrStubImagePlugin
DEBUG:PIL.Image:Importing CurImagePlugin
DEBUG:PIL.Image:Importing DcxImagePlugin
DEBUG:PIL.Image:Importing DdsImagePlugin
DEBUG:PIL.Image:Importing EpsImagePlugin
DEBUG:PIL.Image:Importing FitsImagePlugin
DEBUG:PIL.Image:Importing FitsStubImagePlugin
DEBUG:PIL.Image:Importing FliImagePlugin
DEBUG:PIL.Image:Importing FpxImagePlugin
DEBUG:PIL.Image:Image: failed to import FpxImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing FtexImagePlugin
DEBUG:PIL.Image:Importing GbrImagePlugin
DEBUG:PIL.Image:Importing GifImagePlugin
DEBUG:PIL.Image:Importing GribStubImagePlugin
DEBUG:PIL.Image:Importing Hdf5StubImagePlugin
DEBUG:PIL.Image:Importing IcnsImagePlugin
DEBUG:PIL.Image:Importing IcoImagePlugin
DEBUG:PIL.Image:Importing ImImagePlugin
DEBUG:PIL.Image:Importing ImtImagePlugin
DEBUG:PIL.Image:Importing IptcImagePlugin
DEBUG:PIL.Image:Importing JpegImagePlugin
DEBUG:PIL.Image:Importing Jpeg2KImagePlugin
DEBUG:PIL.Image:Importing McIdasImagePlugin
DEBUG:PIL.Image:Importing MicImagePlugin
DEBUG:PIL.Image:Image: failed to import MicImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing MpegImagePlugin
DEBUG:PIL.Image:Importing MpoImagePlugin
DEBUG:PIL.Image:Importing MspImagePlugin
DEBUG:PIL.Image:Importing PalmImagePlugin
DEBUG:PIL.Image:Importing PcdImagePlugin
DEBUG:PIL.Image:Importing PcxImagePlugin
DEBUG:PIL.Image:Importing PdfImagePlugin
DEBUG:PIL.Image:Importing PixarImagePlugin
DEBUG:PIL.Image:Importing PngImagePlugin
DEBUG:PIL.Image:Importing PpmImagePlugin
DEBUG:PIL.Image:Importing PsdImagePlugin
DEBUG:PIL.Image:Importing QoiImagePlugin
DEBUG:PIL.Image:Importing SgiImagePlugin
DEBUG:PIL.Image:Importing SpiderImagePlugin
DEBUG:PIL.Image:Importing SunImagePlugin
DEBUG:PIL.Image:Importing TgaImagePlugin
DEBUG:PIL.Image:Importing TiffImagePlugin
DEBUG:PIL.Image:Importing WebPImagePlugin
DEBUG:PIL.Image:Importing WmfImagePlugin
DEBUG:PIL.Image:Importing XbmImagePlugin
DEBUG:PIL.Image:Importing XpmImagePlugin
DEBUG:PIL.Image:Importing XVThumbImagePlugin
8%|▊ | 1/12 [01:13<13:33, 73.92s/it]
17%|█▋ | 2/12 [02:28<12:22, 74.25s/it]
25%|██▌ | 3/12 [03:41<11:01, 73.53s/it]
33%|███▎ | 4/12 [04:53<09:45, 73.17s/it]
42%|████▏ | 5/12 [06:06<08:30, 72.96s/it]
50%|█████ | 6/12 [07:18<07:16, 72.80s/it]
58%|█████▊ | 7/12 [08:31<06:03, 72.71s/it]
67%|██████▋ | 8/12 [09:43<04:50, 72.64s/it]
75%|███████▌ | 9/12 [10:56<03:37, 72.61s/it]
83%|████████▎ | 10/12 [12:08<02:25, 72.56s/it]
92%|█████████▏| 11/12 [13:21<01:12, 72.53s/it]
100%|██████████| 12/12 [14:33<00:00, 72.48s/it]
100%|██████████| 12/12 [14:33<00:00, 72.80s/it]
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 1: generating actions ...
>>> Step 1: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 2: generating actions ...
>>> Step 2: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 3: generating actions ...
>>> Step 3: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 4: generating actions ...
>>> Step 4: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 5: generating actions ...
>>> Step 5: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 6: generating actions ...
>>> Step 6: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 7: generating actions ...
>>> Step 7: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 8: generating actions ...
>>> Step 8: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 9: generating actions ...
>>> Step 9: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 10: generating actions ...
>>> Step 10: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 11: generating actions ...
>>> Step 11: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
real 16m6.503s
user 71m18.296s
sys 0m44.031s
2026-02-11 23:51:59.440508: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2026-02-11 23:51:59.489849: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2026-02-11 23:51:59.489897: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2026-02-11 23:51:59.491194: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2026-02-11 23:51:59.498659: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2026-02-11 23:52:00.422248: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Global seed set to 123
INFO:mainlogger:LatentVisualDiffusion: Running in v-prediction mode
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
AE working on z of shape (1, 4, 32, 32) = 4096 dimensions.
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
>>> model checkpoint loaded.
>>> Load pre-trained model ...
INFO:root:***** Configing Data *****
>>> unitree_z1_stackbox: 1 data samples loaded.
>>> unitree_z1_stackbox: data stats loaded.
>>> unitree_z1_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox: data stats loaded.
>>> unitree_z1_dual_arm_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox_v2: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox_v2: data stats loaded.
>>> unitree_z1_dual_arm_stackbox_v2: normalizer initiated.
>>> unitree_z1_dual_arm_cleanup_pencils: 1 data samples loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: data stats loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: normalizer initiated.
>>> unitree_g1_pack_camera: 1 data samples loaded.
>>> unitree_g1_pack_camera: data stats loaded.
>>> unitree_g1_pack_camera: normalizer initiated.
>>> Dataset is successfully loaded ...
>>> Generate 16 frames under each generation ...
DEBUG:h5py._conv:Creating converter from 3 to 5
DEBUG:PIL.PngImagePlugin:STREAM b'IHDR' 16 13
DEBUG:PIL.PngImagePlugin:STREAM b'pHYs' 41 9
DEBUG:PIL.PngImagePlugin:STREAM b'IDAT' 62 4096
0%| | 0/12 [00:00<?, ?it/s]>>> Step 0: generating actions ...
>>> Step 0: interacting with world model ...
DEBUG:PIL.Image:Importing BlpImagePlugin
DEBUG:PIL.Image:Importing BmpImagePlugin
DEBUG:PIL.Image:Importing BufrStubImagePlugin
DEBUG:PIL.Image:Importing CurImagePlugin
DEBUG:PIL.Image:Importing DcxImagePlugin
DEBUG:PIL.Image:Importing DdsImagePlugin
DEBUG:PIL.Image:Importing EpsImagePlugin
DEBUG:PIL.Image:Importing FitsImagePlugin
DEBUG:PIL.Image:Importing FitsStubImagePlugin
DEBUG:PIL.Image:Importing FliImagePlugin
DEBUG:PIL.Image:Importing FpxImagePlugin
DEBUG:PIL.Image:Image: failed to import FpxImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing FtexImagePlugin
DEBUG:PIL.Image:Importing GbrImagePlugin
DEBUG:PIL.Image:Importing GifImagePlugin
DEBUG:PIL.Image:Importing GribStubImagePlugin
DEBUG:PIL.Image:Importing Hdf5StubImagePlugin
DEBUG:PIL.Image:Importing IcnsImagePlugin
DEBUG:PIL.Image:Importing IcoImagePlugin
DEBUG:PIL.Image:Importing ImImagePlugin
DEBUG:PIL.Image:Importing ImtImagePlugin
DEBUG:PIL.Image:Importing IptcImagePlugin
DEBUG:PIL.Image:Importing JpegImagePlugin
DEBUG:PIL.Image:Importing Jpeg2KImagePlugin
DEBUG:PIL.Image:Importing McIdasImagePlugin
DEBUG:PIL.Image:Importing MicImagePlugin
DEBUG:PIL.Image:Image: failed to import MicImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing MpegImagePlugin
DEBUG:PIL.Image:Importing MpoImagePlugin
DEBUG:PIL.Image:Importing MspImagePlugin
DEBUG:PIL.Image:Importing PalmImagePlugin
DEBUG:PIL.Image:Importing PcdImagePlugin
DEBUG:PIL.Image:Importing PcxImagePlugin
DEBUG:PIL.Image:Importing PdfImagePlugin
DEBUG:PIL.Image:Importing PixarImagePlugin
DEBUG:PIL.Image:Importing PngImagePlugin
DEBUG:PIL.Image:Importing PpmImagePlugin
DEBUG:PIL.Image:Importing PsdImagePlugin
DEBUG:PIL.Image:Importing QoiImagePlugin
DEBUG:PIL.Image:Importing SgiImagePlugin
DEBUG:PIL.Image:Importing SpiderImagePlugin
DEBUG:PIL.Image:Importing SunImagePlugin
DEBUG:PIL.Image:Importing TgaImagePlugin
DEBUG:PIL.Image:Importing TiffImagePlugin
DEBUG:PIL.Image:Importing WebPImagePlugin
DEBUG:PIL.Image:Importing WmfImagePlugin
DEBUG:PIL.Image:Importing XbmImagePlugin
DEBUG:PIL.Image:Importing XpmImagePlugin
DEBUG:PIL.Image:Importing XVThumbImagePlugin
8%|▊ | 1/12 [01:14<13:34, 74.06s/it]
17%|█▋ | 2/12 [02:27<12:17, 73.74s/it]
25%|██▌ | 3/12 [03:40<10:59, 73.23s/it]
33%|███▎ | 4/12 [04:52<09:43, 72.94s/it]
42%|████▏ | 5/12 [06:05<08:29, 72.81s/it]
50%|█████ | 6/12 [07:17<07:16, 72.74s/it]
58%|█████▊ | 7/12 [08:30<06:03, 72.68s/it]
67%|██████▋ | 8/12 [09:42<04:50, 72.63s/it]
75%|███████▌ | 9/12 [10:55<03:38, 72.72s/it]
83%|████████▎ | 10/12 [12:09<02:25, 72.86s/it]
92%|█████████▏| 11/12 [13:21<01:12, 72.75s/it]
100%|██████████| 12/12 [14:33<00:00, 72.66s/it]
100%|██████████| 12/12 [14:33<00:00, 72.83s/it]
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 1: generating actions ...
>>> Step 1: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 2: generating actions ...
>>> Step 2: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 3: generating actions ...
>>> Step 3: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 4: generating actions ...
>>> Step 4: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 5: generating actions ...
>>> Step 5: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 6: generating actions ...
>>> Step 6: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 7: generating actions ...
>>> Step 7: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 8: generating actions ...
>>> Step 8: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 9: generating actions ...
>>> Step 9: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 10: generating actions ...
>>> Step 10: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 11: generating actions ...
>>> Step 11: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
real 16m6.056s
user 71m23.029s
sys 0m43.382s
2026-02-12 00:08:05.532356: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2026-02-12 00:08:05.582134: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2026-02-12 00:08:05.582180: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2026-02-12 00:08:05.583503: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2026-02-12 00:08:05.591049: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2026-02-12 00:08:06.523218: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Global seed set to 123
INFO:mainlogger:LatentVisualDiffusion: Running in v-prediction mode
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
AE working on z of shape (1, 4, 32, 32) = 4096 dimensions.
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
INFO:root:Loaded ViT-H-14 model config.
INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
>>> model checkpoint loaded.
>>> Load pre-trained model ...
INFO:root:***** Configing Data *****
>>> unitree_z1_stackbox: 1 data samples loaded.
>>> unitree_z1_stackbox: data stats loaded.
>>> unitree_z1_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox: data stats loaded.
>>> unitree_z1_dual_arm_stackbox: normalizer initiated.
>>> unitree_z1_dual_arm_stackbox_v2: 1 data samples loaded.
>>> unitree_z1_dual_arm_stackbox_v2: data stats loaded.
>>> unitree_z1_dual_arm_stackbox_v2: normalizer initiated.
>>> unitree_z1_dual_arm_cleanup_pencils: 1 data samples loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: data stats loaded.
>>> unitree_z1_dual_arm_cleanup_pencils: normalizer initiated.
>>> unitree_g1_pack_camera: 1 data samples loaded.
>>> unitree_g1_pack_camera: data stats loaded.
>>> unitree_g1_pack_camera: normalizer initiated.
>>> Dataset is successfully loaded ...
>>> Generate 16 frames under each generation ...
DEBUG:h5py._conv:Creating converter from 3 to 5
DEBUG:PIL.PngImagePlugin:STREAM b'IHDR' 16 13
DEBUG:PIL.PngImagePlugin:STREAM b'pHYs' 41 9
DEBUG:PIL.PngImagePlugin:STREAM b'IDAT' 62 4096
0%| | 0/12 [00:00<?, ?it/s]>>> Step 0: generating actions ...
>>> Step 0: interacting with world model ...
DEBUG:PIL.Image:Importing BlpImagePlugin
DEBUG:PIL.Image:Importing BmpImagePlugin
DEBUG:PIL.Image:Importing BufrStubImagePlugin
DEBUG:PIL.Image:Importing CurImagePlugin
DEBUG:PIL.Image:Importing DcxImagePlugin
DEBUG:PIL.Image:Importing DdsImagePlugin
DEBUG:PIL.Image:Importing EpsImagePlugin
DEBUG:PIL.Image:Importing FitsImagePlugin
DEBUG:PIL.Image:Importing FitsStubImagePlugin
DEBUG:PIL.Image:Importing FliImagePlugin
DEBUG:PIL.Image:Importing FpxImagePlugin
DEBUG:PIL.Image:Image: failed to import FpxImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing FtexImagePlugin
DEBUG:PIL.Image:Importing GbrImagePlugin
DEBUG:PIL.Image:Importing GifImagePlugin
DEBUG:PIL.Image:Importing GribStubImagePlugin
DEBUG:PIL.Image:Importing Hdf5StubImagePlugin
DEBUG:PIL.Image:Importing IcnsImagePlugin
DEBUG:PIL.Image:Importing IcoImagePlugin
DEBUG:PIL.Image:Importing ImImagePlugin
DEBUG:PIL.Image:Importing ImtImagePlugin
DEBUG:PIL.Image:Importing IptcImagePlugin
DEBUG:PIL.Image:Importing JpegImagePlugin
DEBUG:PIL.Image:Importing Jpeg2KImagePlugin
DEBUG:PIL.Image:Importing McIdasImagePlugin
DEBUG:PIL.Image:Importing MicImagePlugin
DEBUG:PIL.Image:Image: failed to import MicImagePlugin: No module named 'olefile'
DEBUG:PIL.Image:Importing MpegImagePlugin
DEBUG:PIL.Image:Importing MpoImagePlugin
DEBUG:PIL.Image:Importing MspImagePlugin
DEBUG:PIL.Image:Importing PalmImagePlugin
DEBUG:PIL.Image:Importing PcdImagePlugin
DEBUG:PIL.Image:Importing PcxImagePlugin
DEBUG:PIL.Image:Importing PdfImagePlugin
DEBUG:PIL.Image:Importing PixarImagePlugin
DEBUG:PIL.Image:Importing PngImagePlugin
DEBUG:PIL.Image:Importing PpmImagePlugin
DEBUG:PIL.Image:Importing PsdImagePlugin
DEBUG:PIL.Image:Importing QoiImagePlugin
DEBUG:PIL.Image:Importing SgiImagePlugin
DEBUG:PIL.Image:Importing SpiderImagePlugin
DEBUG:PIL.Image:Importing SunImagePlugin
DEBUG:PIL.Image:Importing TgaImagePlugin
DEBUG:PIL.Image:Importing TiffImagePlugin
DEBUG:PIL.Image:Importing WebPImagePlugin
DEBUG:PIL.Image:Importing WmfImagePlugin
DEBUG:PIL.Image:Importing XbmImagePlugin
DEBUG:PIL.Image:Importing XpmImagePlugin
DEBUG:PIL.Image:Importing XVThumbImagePlugin
8%|▊ | 1/12 [01:13<13:30, 73.69s/it]
17%|█▋ | 2/12 [02:26<12:12, 73.23s/it]
25%|██▌ | 3/12 [03:40<11:02, 73.59s/it]
33%|███▎ | 4/12 [04:54<09:50, 73.76s/it]
42%|████▏ | 5/12 [06:07<08:34, 73.51s/it]
50%|█████ | 6/12 [07:20<07:19, 73.19s/it]
58%|█████▊ | 7/12 [08:32<06:04, 72.98s/it]
67%|██████▋ | 8/12 [09:45<04:51, 72.85s/it]
75%|███████▌ | 9/12 [10:57<03:38, 72.76s/it]
83%|████████▎ | 10/12 [12:10<02:25, 72.69s/it]
92%|█████████▏| 11/12 [13:23<01:12, 72.66s/it]
100%|██████████| 12/12 [14:35<00:00, 72.63s/it]
100%|██████████| 12/12 [14:35<00:00, 72.97s/it]
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 1: generating actions ...
>>> Step 1: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 2: generating actions ...
>>> Step 2: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 3: generating actions ...
>>> Step 3: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 4: generating actions ...
>>> Step 4: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 5: generating actions ...
>>> Step 5: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 6: generating actions ...
>>> Step 6: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 7: generating actions ...
>>> Step 7: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 8: generating actions ...
>>> Step 8: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 9: generating actions ...
>>> Step 9: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 10: generating actions ...
>>> Step 10: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
>>> Step 11: generating actions ...
>>> Step 11: interacting with world model ...
>>>>>>>>>>>>>>>>>>>>>>>>
real 16m7.783s
user 71m30.579s
sys 0m41.570s