Add files using upload-large-folder tool

wandb/run-20260124_022454-nlxwwxfq/files/output.log

@@ -0,0 +1,44 @@
+I0124 02:24:55.881326 140556193482560 train_g2rpo_sd_merge.py:465]
+allow_tf32: true
+logdir: logs
+mixed_precision: bf16
+num_checkpoint_limit: 5
+num_epochs: 300
+pretrained:
+  model: ./data/StableDiffusion
+  revision: main
+prompt_fn: imagenet_animals
+prompt_fn_kwargs: {}
+resume_from: ''
+reward_fn: hpsv2
+run_name: 2026.01.24_02.24.53
+sample:
+  batch_size: 1
+  eta: 1.0
+  guidance_scale: 5.0
+  num_batches_per_epoch: 2
+  num_steps: 50
+save_freq: 20
+seed: 42
+train:
+  adam_beta1: 0.9
+  adam_beta2: 0.999
+  adam_epsilon: 1.0e-08
+  adam_weight_decay: 0.0001
+  adv_clip_max: 5
+  batch_size: 1
+  cfg: true
+  clip_range: 0.0001
+  gradient_accumulation_steps: 1
+  learning_rate: 1.0e-05
+  max_grad_norm: 1.0
+  num_inner_epochs: 1
+  timestep_fraction: 1.0
+  use_8bit_adam: false
+use_lora: false
+
+Loading pipeline components...: 100%|███████████████████████████████████████████████████████████████████████████████████████████| 7/7 [00:02<00:00,  2.67it/s]
+/home/zsj/anaconda3/envs/g2rpo/lib/python3.10/site-packages/timm/models/layers/__init__.py:48: FutureWarning: Importing from timm.models.layers is deprecated, please import via timm.layers
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+I0124 02:24:59.144597 140556193482560 factory.py:159] Loaded ViT-H-14 model config.
+I0124 02:25:04.967288 140556193482560 factory.py:207] Loading pretrained ViT-H-14 weights (./data/hps/open_clip_pytorch_model.bin).

wandb/run-20260124_022454-nlxwwxfq/files/requirements.txt

@@ -0,0 +1,189 @@
+scipy==1.13.0
+regex==2024.9.11
+sentencepiece==0.2.0
+six==1.16.0
+anyio==4.11.0
+nvidia-cuda-nvrtc-cu12==12.6.77
+scikit-video==1.1.11
+platformdirs==4.5.0
+mypy==1.11.1
+ruff==0.6.5
+charset-normalizer==3.4.4
+torch==2.9.0+cu126
+av==13.1.0
+pillow==10.2.0
+gpustat==1.1.1
+torchvision==0.24.0+cu126
+multidict==6.7.0
+torchmetrics==1.5.1
+aiohttp==3.13.1
+transformers==4.46.1
+decord==0.6.0
+wcwidth==0.2.14
+sphinx-lint==1.0.0
+nvidia-cuda-runtime-cu12==12.6.77
+pytz==2025.2
+codespell==2.3.0
+hpsv2==1.2.0
+mypy_extensions==1.1.0
+numpy==1.26.3
+omegaconf==2.3.0
+Markdown==3.9
+tzdata==2025.2
+pandas==2.2.3
+pytorch-lightning==2.4.0
+aiosignal==1.4.0
+aiohappyeyeballs==2.6.1
+python-dateutil==2.9.0.post0
+seaborn==0.13.2
+beautifulsoup4==4.12.3
+isort==5.13.2
+httpx==0.28.1
+certifi==2025.10.5
+ml_collections==1.1.0
+nvidia-cudnn-cu12==9.10.2.21
+hf-xet==1.2.0
+requests==2.31.0
+inflect==6.0.4
+iniconfig==2.1.0
+braceexpand==0.1.7
+h5py==3.12.1
+wandb==0.18.5
+protobuf==3.20.3
+ninja==1.13.0
+kiwisolver==1.4.9
+networkx==3.3
+packaging==25.0
+fvcore==0.1.5.post20221221
+pyparsing==3.2.5
+starlette==0.41.3
+frozenlist==1.8.0
+docker-pycreds==0.4.0
+Werkzeug==3.1.3
+MarkupSafe==2.1.5
+einops==0.8.0
+sentry-sdk==2.42.0
+PyYAML==6.0.1
+nvidia-nccl-cu12==2.27.5
+datasets==4.3.0
+polib==1.2.0
+safetensors==0.6.2
+async-timeout==5.0.1
+setproctitle==1.3.7
+clint==0.5.1
+matplotlib==3.9.2
+propcache==0.4.1
+termcolor==3.1.0
+antlr4-python3-runtime==4.9.3
+cycler==0.12.1
+fastvideo==1.2.0
+toml==0.10.2
+xxhash==3.6.0
+wheel==0.44.0
+albumentations==1.4.20
+fastapi==0.115.3
+nvidia-cufft-cu12==11.3.0.4
+yarl==1.22.0
+psutil==7.1.0
+tensorboard-data-server==0.7.2
+pydantic==2.9.2
+nvidia-nvtx-cu12==12.6.77
+portalocker==3.2.0
+triton==3.5.0
+annotated-types==0.7.0
+proglog==0.1.12
+nvidia-cusparselt-cu12==0.7.1
+yapf==0.32.0
+Jinja2==3.1.6
+types-requests==2.32.4.20250913
+lightning-utilities==0.15.2
+grpcio==1.75.1
+uvicorn==0.32.0
+typing_extensions==4.15.0
+nvidia-nvjitlink-cu12==12.6.85
+watch==0.2.7
+moviepy==1.0.3
+timm==1.0.11
+pytest-split==0.8.0
+gdown==5.2.0
+types-setuptools==80.9.0.20250822
+nvidia-cusolver-cu12==11.7.1.2
+types-PyYAML==6.0.12.20250915
+pip==25.2
+qwen-vl-utils==0.0.14
+soupsieve==2.8
+zipp==3.23.0
+flash_attn==2.8.3
+yacs==0.1.8
+diffusers==0.32.0
+pluggy==1.6.0
+opencv-python-headless==4.11.0.86
+mpmath==1.3.0
+test_tube==0.7.5
+stringzilla==4.2.1
+fonttools==4.60.1
+nvidia-ml-py==13.580.82
+parameterized==0.9.0
+loguru==0.7.3
+tabulate==0.9.0
+idna==3.6
+iopath==0.1.10
+decorator==4.4.2
+nvidia-cufile-cu12==1.11.1.6
+threadpoolctl==3.6.0
+pyarrow==21.0.0
+httpcore==1.0.9
+hydra-core==1.3.2
+multiprocess==0.70.16
+contourpy==1.3.2
+clip==1.0
+tqdm==4.66.5
+open_clip_torch==3.2.0
+accelerate==1.0.1
+gitdb==4.0.12
+importlib_metadata==8.7.0
+nvidia-cublas-cu12==12.6.4.1
+h11==0.16.0
+filelock==3.19.1
+liger_kernel==0.4.1
+click==8.3.0
+urllib3==2.2.0
+imageio-ffmpeg==0.5.1
+setuptools==80.9.0
+joblib==1.5.2
+tensorboard==2.20.0
+attrs==25.4.0
+future==1.0.0
+albucore==0.0.19
+fsspec==2025.9.0
+sympy==1.14.0
+eval_type_backport==0.2.2
+pydantic_core==2.23.4
+sniffio==1.3.1
+nvidia-nvshmem-cu12==3.3.20
+exceptiongroup==1.3.0
+smmap==5.0.2
+tomli==2.0.2
+ftfy==6.3.0
+dill==0.4.0
+pytest==7.2.0
+PySocks==1.7.1
+nvidia-curand-cu12==10.3.7.77
+tokenizers==0.20.1
+args==0.1.0
+fairscale==0.4.13
+peft==0.13.2
+webdataset==1.0.2
+huggingface-hub==0.26.1
+GitPython==3.1.45
+pytorchvideo==0.1.5
+scikit-learn==1.5.2
+bitsandbytes==0.48.1
+nvidia-cusparse-cu12==12.5.4.2
+nvidia-cuda-cupti-cu12==12.6.80
+imageio==2.36.0
+pydub==0.25.1
+image-reward==1.5
+absl-py==2.3.1
+blessed==1.22.0
+torchdiffeq==0.2.4

wandb/run-20260124_023610-n1ooz9bl/logs/debug-core.log

@@ -0,0 +1,8 @@
+{"time":"2026-01-24T02:36:09.151602379+08:00","level":"INFO","msg":"started logging, with flags","port-filename":"/tmp/tmpaso2kes5/port-612439.txt","pid":612439,"debug":false,"disable-analytics":false}
+{"time":"2026-01-24T02:36:09.151632657+08:00","level":"INFO","msg":"FeatureState","shutdownOnParentExitEnabled":false}
+{"time":"2026-01-24T02:36:09.15256438+08:00","level":"INFO","msg":"server is running","addr":{"IP":"127.0.0.1","Port":42423,"Zone":""}}
+{"time":"2026-01-24T02:36:09.152689129+08:00","level":"INFO","msg":"Will exit if parent process dies.","ppid":612439}
+{"time":"2026-01-24T02:36:09.342363565+08:00","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"127.0.0.1:38120"}
+{"time":"2026-01-24T02:36:10.159795478+08:00","level":"INFO","msg":"handleInformInit: received","streamId":"n1ooz9bl","id":"127.0.0.1:38120"}
+{"time":"2026-01-24T02:36:10.27267688+08:00","level":"INFO","msg":"handleInformInit: stream started","streamId":"n1ooz9bl","id":"127.0.0.1:38120"}
+{"time":"2026-01-24T02:37:22.867050941+08:00","level":"INFO","msg":"Parent process exited, terminating service process."}

wandb/run-20260124_023610-n1ooz9bl/logs/debug-internal.log

@@ -0,0 +1,8 @@
+{"time":"2026-01-24T02:36:10.160103717+08:00","level":"INFO","msg":"using version","core version":"0.18.5"}
+{"time":"2026-01-24T02:36:10.160137584+08:00","level":"INFO","msg":"created symlink","path":"/data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/wandb/run-20260124_023610-n1ooz9bl/logs/debug-core.log"}
+{"time":"2026-01-24T02:36:10.272633973+08:00","level":"INFO","msg":"created new stream","id":"n1ooz9bl"}
+{"time":"2026-01-24T02:36:10.272673149+08:00","level":"INFO","msg":"stream: started","id":"n1ooz9bl"}
+{"time":"2026-01-24T02:36:10.272719319+08:00","level":"INFO","msg":"sender: started","stream_id":"n1ooz9bl"}
+{"time":"2026-01-24T02:36:10.27275806+08:00","level":"INFO","msg":"handler: started","stream_id":{"value":"n1ooz9bl"}}
+{"time":"2026-01-24T02:36:10.272722646+08:00","level":"INFO","msg":"writer: Do: started","stream_id":{"value":"n1ooz9bl"}}
+{"time":"2026-01-24T02:36:11.031290755+08:00","level":"INFO","msg":"Starting system monitor"}

wandb/run-20260124_110518-tfmkls9a/files/config.yaml

@@ -0,0 +1,87 @@
+_wandb:
+    value:
+        cli_version: 0.18.5
+        m: []
+        python_version: 3.10.19
+        t:
+            "1":
+                - 1
+                - 11
+                - 41
+                - 49
+                - 55
+                - 71
+                - 83
+                - 98
+            "2":
+                - 1
+                - 11
+                - 41
+                - 49
+                - 55
+                - 63
+                - 71
+                - 83
+                - 98
+            "3":
+                - 13
+                - 23
+                - 55
+            "4": 3.10.19
+            "5": 0.18.5
+            "6": 4.46.1
+            "8":
+                - 5
+            "12": 0.18.5
+            "13": linux-x86_64
+allow_tf32:
+    value: true
+logdir:
+    value: logs
+mixed_precision:
+    value: bf16
+num_checkpoint_limit:
+    value: 5
+num_epochs:
+    value: 300
+pretrained:
+    value:
+        model: ./data/StableDiffusion
+        revision: main
+prompt_fn:
+    value: imagenet_animals
+resume_from:
+    value: ""
+reward_fn:
+    value: hpsv2
+run_name:
+    value: 2026.01.24_11.05.16
+sample:
+    value:
+        batch_size: 1
+        eta: 1
+        guidance_scale: 5
+        num_batches_per_epoch: 2
+        num_steps: 50
+save_freq:
+    value: 20
+seed:
+    value: 42
+train:
+    value:
+        adam_beta1: 0.9
+        adam_beta2: 0.999
+        adam_epsilon: 1e-08
+        adam_weight_decay: 0.0001
+        adv_clip_max: 5
+        batch_size: 1
+        cfg: true
+        clip_range: 0.0001
+        gradient_accumulation_steps: 1
+        learning_rate: 1e-05
+        max_grad_norm: 1
+        num_inner_epochs: 1
+        timestep_fraction: 1
+        use_8bit_adam: false
+use_lora:
+    value: false

wandb/run-20260124_110518-tfmkls9a/files/output.log

@@ -0,0 +1,70 @@
+I0124 11:05:19.230831 135505156663104 train_g2rpo_sd_merge.py:478]
+allow_tf32: true
+logdir: logs
+mixed_precision: bf16
+num_checkpoint_limit: 5
+num_epochs: 300
+pretrained:
+  model: ./data/StableDiffusion
+  revision: main
+prompt_fn: imagenet_animals
+prompt_fn_kwargs: {}
+resume_from: ''
+reward_fn: hpsv2
+run_name: 2026.01.24_11.05.16
+sample:
+  batch_size: 1
+  eta: 1.0
+  guidance_scale: 5.0
+  num_batches_per_epoch: 2
+  num_steps: 50
+save_freq: 20
+seed: 42
+train:
+  adam_beta1: 0.9
+  adam_beta2: 0.999
+  adam_epsilon: 1.0e-08
+  adam_weight_decay: 0.0001
+  adv_clip_max: 5
+  batch_size: 1
+  cfg: true
+  clip_range: 0.0001
+  gradient_accumulation_steps: 1
+  learning_rate: 1.0e-05
+  max_grad_norm: 1.0
+  num_inner_epochs: 1
+  timestep_fraction: 1.0
+  use_8bit_adam: false
+use_lora: false
+
+Loading pipeline components...: 100%|███████████████████████████████████████████████████████████████████████████████████████████| 7/7 [00:02<00:00,  2.65it/s]
+/home/zsj/anaconda3/envs/g2rpo/lib/python3.10/site-packages/timm/models/layers/__init__.py:48: FutureWarning: Importing from timm.models.layers is deprecated, please import via timm.layers
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+I0124 11:05:22.495319 135505156663104 factory.py:159] Loaded ViT-H-14 model config.
+I0124 11:05:27.175477 135505156663104 factory.py:207] Loading pretrained ViT-H-14 weights (./data/hps/open_clip_pytorch_model.bin).
+I0124 11:05:32.379540 135505156663104 train_g2rpo_sd_merge.py:670] ***** Running E-GRPO (G2RPO) Training for Stable Diffusion *****
+I0124 11:05:32.380349 135505156663104 train_g2rpo_sd_merge.py:671]   Num Epochs = 300
+I0124 11:05:32.380448 135505156663104 train_g2rpo_sd_merge.py:672]   Num generations per prompt = 4
+I0124 11:05:32.380522 135505156663104 train_g2rpo_sd_merge.py:673]   Eta step list = [0, 1, 2, 3, 4, 5, 6, 7]
+I0124 11:05:32.380584 135505156663104 train_g2rpo_sd_merge.py:674]   Eta step merge list = [1, 1, 1, 2, 2, 2, 3, 3]
+I0124 11:05:32.380647 135505156663104 train_g2rpo_sd_merge.py:675]   Granular list = [1]
+Traceback (most recent call last):
+  File "/data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/fastvideo/train_g2rpo_sd_merge.py", line 1001, in <module>
+    app.run(main)
+  File "/home/zsj/anaconda3/envs/g2rpo/lib/python3.10/site-packages/absl/app.py", line 316, in run
+    _run_main(main, args)
+  File "/home/zsj/anaconda3/envs/g2rpo/lib/python3.10/site-packages/absl/app.py", line 261, in _run_main
+    sys.exit(main(argv))
+  File "/data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/fastvideo/train_g2rpo_sd_merge.py", line 786, in main
+    (eval_image[0].cpu().permute(1, 2, 0).numpy() * 255).astype(np.uint8)
+TypeError: Got unsupported ScalarType BFloat16
+[rank0]: Traceback (most recent call last):
+[rank0]:   File "/data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/fastvideo/train_g2rpo_sd_merge.py", line 1001, in <module>
+[rank0]:     app.run(main)
+[rank0]:   File "/home/zsj/anaconda3/envs/g2rpo/lib/python3.10/site-packages/absl/app.py", line 316, in run
+[rank0]:     _run_main(main, args)
+[rank0]:   File "/home/zsj/anaconda3/envs/g2rpo/lib/python3.10/site-packages/absl/app.py", line 261, in _run_main
+[rank0]:     sys.exit(main(argv))
+[rank0]:   File "/data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/fastvideo/train_g2rpo_sd_merge.py", line 786, in main
+[rank0]:     (eval_image[0].cpu().permute(1, 2, 0).numpy() * 255).astype(np.uint8)
+[rank0]: TypeError: Got unsupported ScalarType BFloat16

wandb/run-20260124_110518-tfmkls9a/files/requirements.txt

@@ -0,0 +1,189 @@
+scipy==1.13.0
+regex==2024.9.11
+sentencepiece==0.2.0
+six==1.16.0
+anyio==4.11.0
+nvidia-cuda-nvrtc-cu12==12.6.77
+scikit-video==1.1.11
+platformdirs==4.5.0
+mypy==1.11.1
+ruff==0.6.5
+charset-normalizer==3.4.4
+torch==2.9.0+cu126
+av==13.1.0
+pillow==10.2.0
+gpustat==1.1.1
+torchvision==0.24.0+cu126
+multidict==6.7.0
+torchmetrics==1.5.1
+aiohttp==3.13.1
+transformers==4.46.1
+decord==0.6.0
+wcwidth==0.2.14
+sphinx-lint==1.0.0
+nvidia-cuda-runtime-cu12==12.6.77
+pytz==2025.2
+codespell==2.3.0
+hpsv2==1.2.0
+mypy_extensions==1.1.0
+numpy==1.26.3
+omegaconf==2.3.0
+Markdown==3.9
+tzdata==2025.2
+pandas==2.2.3
+pytorch-lightning==2.4.0
+aiosignal==1.4.0
+aiohappyeyeballs==2.6.1
+python-dateutil==2.9.0.post0
+seaborn==0.13.2
+beautifulsoup4==4.12.3
+isort==5.13.2
+httpx==0.28.1
+certifi==2025.10.5
+ml_collections==1.1.0
+nvidia-cudnn-cu12==9.10.2.21
+hf-xet==1.2.0
+requests==2.31.0
+inflect==6.0.4
+iniconfig==2.1.0
+braceexpand==0.1.7
+h5py==3.12.1
+wandb==0.18.5
+protobuf==3.20.3
+ninja==1.13.0
+kiwisolver==1.4.9
+networkx==3.3
+packaging==25.0
+fvcore==0.1.5.post20221221
+pyparsing==3.2.5
+starlette==0.41.3
+frozenlist==1.8.0
+docker-pycreds==0.4.0
+Werkzeug==3.1.3
+MarkupSafe==2.1.5
+einops==0.8.0
+sentry-sdk==2.42.0
+PyYAML==6.0.1
+nvidia-nccl-cu12==2.27.5
+datasets==4.3.0
+polib==1.2.0
+safetensors==0.6.2
+async-timeout==5.0.1
+setproctitle==1.3.7
+clint==0.5.1
+matplotlib==3.9.2
+propcache==0.4.1
+termcolor==3.1.0
+antlr4-python3-runtime==4.9.3
+cycler==0.12.1
+fastvideo==1.2.0
+toml==0.10.2
+xxhash==3.6.0
+wheel==0.44.0
+albumentations==1.4.20
+fastapi==0.115.3
+nvidia-cufft-cu12==11.3.0.4
+yarl==1.22.0
+psutil==7.1.0
+tensorboard-data-server==0.7.2
+pydantic==2.9.2
+nvidia-nvtx-cu12==12.6.77
+portalocker==3.2.0
+triton==3.5.0
+annotated-types==0.7.0
+proglog==0.1.12
+nvidia-cusparselt-cu12==0.7.1
+yapf==0.32.0
+Jinja2==3.1.6
+types-requests==2.32.4.20250913
+lightning-utilities==0.15.2
+grpcio==1.75.1
+uvicorn==0.32.0
+typing_extensions==4.15.0
+nvidia-nvjitlink-cu12==12.6.85
+watch==0.2.7
+moviepy==1.0.3
+timm==1.0.11
+pytest-split==0.8.0
+gdown==5.2.0
+types-setuptools==80.9.0.20250822
+nvidia-cusolver-cu12==11.7.1.2
+types-PyYAML==6.0.12.20250915
+pip==25.2
+qwen-vl-utils==0.0.14
+soupsieve==2.8
+zipp==3.23.0
+flash_attn==2.8.3
+yacs==0.1.8
+diffusers==0.32.0
+pluggy==1.6.0
+opencv-python-headless==4.11.0.86
+mpmath==1.3.0
+test_tube==0.7.5
+stringzilla==4.2.1
+fonttools==4.60.1
+nvidia-ml-py==13.580.82
+parameterized==0.9.0
+loguru==0.7.3
+tabulate==0.9.0
+idna==3.6
+iopath==0.1.10
+decorator==4.4.2
+nvidia-cufile-cu12==1.11.1.6
+threadpoolctl==3.6.0
+pyarrow==21.0.0
+httpcore==1.0.9
+hydra-core==1.3.2
+multiprocess==0.70.16
+contourpy==1.3.2
+clip==1.0
+tqdm==4.66.5
+open_clip_torch==3.2.0
+accelerate==1.0.1
+gitdb==4.0.12
+importlib_metadata==8.7.0
+nvidia-cublas-cu12==12.6.4.1
+h11==0.16.0
+filelock==3.19.1
+liger_kernel==0.4.1
+click==8.3.0
+urllib3==2.2.0
+imageio-ffmpeg==0.5.1
+setuptools==80.9.0
+joblib==1.5.2
+tensorboard==2.20.0
+attrs==25.4.0
+future==1.0.0
+albucore==0.0.19
+fsspec==2025.9.0
+sympy==1.14.0
+eval_type_backport==0.2.2
+pydantic_core==2.23.4
+sniffio==1.3.1
+nvidia-nvshmem-cu12==3.3.20
+exceptiongroup==1.3.0
+smmap==5.0.2
+tomli==2.0.2
+ftfy==6.3.0
+dill==0.4.0
+pytest==7.2.0
+PySocks==1.7.1
+nvidia-curand-cu12==10.3.7.77
+tokenizers==0.20.1
+args==0.1.0
+fairscale==0.4.13
+peft==0.13.2
+webdataset==1.0.2
+huggingface-hub==0.26.1
+GitPython==3.1.45
+pytorchvideo==0.1.5
+scikit-learn==1.5.2
+bitsandbytes==0.48.1
+nvidia-cusparse-cu12==12.5.4.2
+nvidia-cuda-cupti-cu12==12.6.80
+imageio==2.36.0
+pydub==0.25.1
+image-reward==1.5
+absl-py==2.3.1
+blessed==1.22.0
+torchdiffeq==0.2.4

wandb/run-20260124_110518-tfmkls9a/files/wandb-metadata.json

@@ -0,0 +1,96 @@
+{
+  "os":  "Linux-6.8.0-85-generic-x86_64-with-glibc2.35",
+  "python":  "3.10.19",
+  "startedAt":  "2026-01-24T03:05:18.186036Z",
+  "args":  [
+    "--config",
+    "fastvideo/config_sd/base.py",
+    "--eta_step_list",
+    "0,1,2,3,4,5,6,7",
+    "--eta_step_merge_list",
+    "1,1,1,2,2,2,3,3",
+    "--granular_list",
+    "1",
+    "--num_generations",
+    "4",
+    "--eta",
+    "1.0",
+    "--init_same_noise"
+  ],
+  "program":  "/data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/fastvideo/train_g2rpo_sd_merge.py",
+  "codePath":  "fastvideo/train_g2rpo_sd_merge.py",
+  "email":  "zhangemail1428@163.com",
+  "root":  "/data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code",
+  "host":  "abc",
+  "username":  "zsj",
+  "executable":  "/home/zsj/anaconda3/envs/g2rpo/bin/python",
+  "codePathLocal":  "fastvideo/train_g2rpo_sd_merge.py",
+  "cpu_count":  48,
+  "cpu_count_logical":  96,
+  "gpu":  "NVIDIA RTX 5880 Ada Generation",
+  "gpu_count":  8,
+  "disk":  {
+    "/":  {
+      "total":  "1006773899264",
+      "used":  "812154195968"
+    }
+  },
+  "memory":  {
+    "total":  "540697260032"
+  },
+  "cpu":  {
+    "count":  48,
+    "countLogical":  96
+  },
+  "gpu_nvidia":  [
+    {
+      "name":  "NVIDIA RTX 5880 Ada Generation",
+      "memoryTotal":  "51527024640",
+      "cudaCores":  14080,
+      "architecture":  "Ada"
+    },
+    {
+      "name":  "NVIDIA RTX 5880 Ada Generation",
+      "memoryTotal":  "51527024640",
+      "cudaCores":  14080,
+      "architecture":  "Ada"
+    },
+    {
+      "name":  "NVIDIA RTX 5880 Ada Generation",
+      "memoryTotal":  "51527024640",
+      "cudaCores":  14080,
+      "architecture":  "Ada"
+    },
+    {
+      "name":  "NVIDIA RTX 5880 Ada Generation",
+      "memoryTotal":  "51527024640",
+      "cudaCores":  14080,
+      "architecture":  "Ada"
+    },
+    {
+      "name":  "NVIDIA RTX 5880 Ada Generation",
+      "memoryTotal":  "51527024640",
+      "cudaCores":  14080,
+      "architecture":  "Ada"
+    },
+    {
+      "name":  "NVIDIA RTX 5880 Ada Generation",
+      "memoryTotal":  "51527024640",
+      "cudaCores":  14080,
+      "architecture":  "Ada"
+    },
+    {
+      "name":  "NVIDIA RTX 5880 Ada Generation",
+      "memoryTotal":  "51527024640",
+      "cudaCores":  14080,
+      "architecture":  "Ada"
+    },
+    {
+      "name":  "NVIDIA RTX 5880 Ada Generation",
+      "memoryTotal":  "51527024640",
+      "cudaCores":  14080,
+      "architecture":  "Ada"
+    }
+  ],
+  "cudaVersion":  "12.9"
+}

wandb/run-20260124_110518-tfmkls9a/files/wandb-summary.json

@@ -0,0 +1 @@
+{"_wandb":{"runtime":15}}

wandb/run-20260124_110518-tfmkls9a/logs/debug-core.log

@@ -0,0 +1,12 @@
+{"time":"2026-01-24T11:05:16.967516588+08:00","level":"INFO","msg":"started logging, with flags","port-filename":"/tmp/tmpfnk0hjfx/port-697635.txt","pid":697635,"debug":false,"disable-analytics":false}
+{"time":"2026-01-24T11:05:16.967548024+08:00","level":"INFO","msg":"FeatureState","shutdownOnParentExitEnabled":false}
+{"time":"2026-01-24T11:05:16.96847478+08:00","level":"INFO","msg":"server is running","addr":{"IP":"127.0.0.1","Port":41747,"Zone":""}}
+{"time":"2026-01-24T11:05:16.968614144+08:00","level":"INFO","msg":"Will exit if parent process dies.","ppid":697635}
+{"time":"2026-01-24T11:05:17.157369334+08:00","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"127.0.0.1:39640"}
+{"time":"2026-01-24T11:05:18.189748725+08:00","level":"INFO","msg":"handleInformInit: received","streamId":"tfmkls9a","id":"127.0.0.1:39640"}
+{"time":"2026-01-24T11:05:18.307722037+08:00","level":"INFO","msg":"handleInformInit: stream started","streamId":"tfmkls9a","id":"127.0.0.1:39640"}
+{"time":"2026-01-24T11:05:34.023368216+08:00","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"127.0.0.1:39640"}
+{"time":"2026-01-24T11:05:34.023526749+08:00","level":"INFO","msg":"connection: Close: initiating connection closure","id":"127.0.0.1:39640"}
+{"time":"2026-01-24T11:05:34.023604687+08:00","level":"INFO","msg":"server is shutting down"}
+{"time":"2026-01-24T11:05:34.023755657+08:00","level":"INFO","msg":"connection: Close: connection successfully closed","id":"127.0.0.1:39640"}
+{"time":"2026-01-24T11:05:35.735660702+08:00","level":"INFO","msg":"Parent process exited, terminating service process."}

wandb/run-20260124_110518-tfmkls9a/logs/debug-internal.log

@@ -0,0 +1,15 @@
+{"time":"2026-01-24T11:05:18.189933418+08:00","level":"INFO","msg":"using version","core version":"0.18.5"}
+{"time":"2026-01-24T11:05:18.189953315+08:00","level":"INFO","msg":"created symlink","path":"/data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/wandb/run-20260124_110518-tfmkls9a/logs/debug-core.log"}
+{"time":"2026-01-24T11:05:18.307674596+08:00","level":"INFO","msg":"created new stream","id":"tfmkls9a"}
+{"time":"2026-01-24T11:05:18.307717317+08:00","level":"INFO","msg":"stream: started","id":"tfmkls9a"}
+{"time":"2026-01-24T11:05:18.307888685+08:00","level":"INFO","msg":"sender: started","stream_id":"tfmkls9a"}
+{"time":"2026-01-24T11:05:18.30782203+08:00","level":"INFO","msg":"writer: Do: started","stream_id":{"value":"tfmkls9a"}}
+{"time":"2026-01-24T11:05:18.307910699+08:00","level":"INFO","msg":"handler: started","stream_id":{"value":"tfmkls9a"}}
+{"time":"2026-01-24T11:05:19.08652137+08:00","level":"INFO","msg":"Starting system monitor"}
+{"time":"2026-01-24T11:05:34.023538308+08:00","level":"INFO","msg":"stream: closing","id":"tfmkls9a"}
+{"time":"2026-01-24T11:05:34.023603272+08:00","level":"INFO","msg":"Stopping system monitor"}
+{"time":"2026-01-24T11:05:34.025002208+08:00","level":"INFO","msg":"Stopped system monitor"}
+{"time":"2026-01-24T11:05:34.69650131+08:00","level":"WARN","msg":"No job ingredients found, not creating job artifact"}
+{"time":"2026-01-24T11:05:34.696532166+08:00","level":"WARN","msg":"No source type found, not creating job artifact"}
+{"time":"2026-01-24T11:05:34.696542852+08:00","level":"INFO","msg":"sender: sendDefer: no job artifact to save"}
+{"time":"2026-01-24T11:05:35.716432442+08:00","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}

wandb/run-20260124_110518-tfmkls9a/logs/debug.log

@@ -0,0 +1,27 @@
+2026-01-24 11:05:18,182 INFO    MainThread:697635 [wandb_setup.py:_flush():79] Current SDK version is 0.18.5
+2026-01-24 11:05:18,182 INFO    MainThread:697635 [wandb_setup.py:_flush():79] Configure stats pid to 697635
+2026-01-24 11:05:18,182 INFO    MainThread:697635 [wandb_setup.py:_flush():79] Loading settings from /home/zsj/.config/wandb/settings
+2026-01-24 11:05:18,182 INFO    MainThread:697635 [wandb_setup.py:_flush():79] Loading settings from /data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/wandb/settings
+2026-01-24 11:05:18,182 INFO    MainThread:697635 [wandb_setup.py:_flush():79] Loading settings from environment variables: {}
+2026-01-24 11:05:18,182 INFO    MainThread:697635 [wandb_setup.py:_flush():79] Applying setup settings: {'mode': None, '_disable_service': None}
+2026-01-24 11:05:18,182 INFO    MainThread:697635 [wandb_setup.py:_flush():79] Inferring run settings from compute environment: {'program_relpath': 'fastvideo/train_g2rpo_sd_merge.py', 'program_abspath': '/data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/fastvideo/train_g2rpo_sd_merge.py', 'program': '/data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/fastvideo/train_g2rpo_sd_merge.py'}
+2026-01-24 11:05:18,182 INFO    MainThread:697635 [wandb_setup.py:_flush():79] Applying login settings: {}
+2026-01-24 11:05:18,182 INFO    MainThread:697635 [wandb_init.py:_log_setup():534] Logging user logs to /data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/wandb/run-20260124_110518-tfmkls9a/logs/debug.log
+2026-01-24 11:05:18,182 INFO    MainThread:697635 [wandb_init.py:_log_setup():535] Logging internal logs to /data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/wandb/run-20260124_110518-tfmkls9a/logs/debug-internal.log
+2026-01-24 11:05:18,182 INFO    MainThread:697635 [wandb_init.py:init():621] calling init triggers
+2026-01-24 11:05:18,182 INFO    MainThread:697635 [wandb_init.py:init():628] wandb.init called with sweep_config: {}
+config: {}
+2026-01-24 11:05:18,182 INFO    MainThread:697635 [wandb_init.py:init():671] starting backend
+2026-01-24 11:05:18,182 INFO    MainThread:697635 [wandb_init.py:init():675] sending inform_init request
+2026-01-24 11:05:18,185 INFO    MainThread:697635 [backend.py:_multiprocessing_setup():104] multiprocessing start_methods=fork,spawn,forkserver, using: spawn
+2026-01-24 11:05:18,185 INFO    MainThread:697635 [wandb_init.py:init():688] backend started and connected
+2026-01-24 11:05:18,190 INFO    MainThread:697635 [wandb_init.py:init():783] updated telemetry
+2026-01-24 11:05:18,190 INFO    MainThread:697635 [wandb_init.py:init():816] communicating run to backend with 90.0 second timeout
+2026-01-24 11:05:19,077 INFO    MainThread:697635 [wandb_init.py:init():867] starting run threads in backend
+2026-01-24 11:05:19,226 INFO    MainThread:697635 [wandb_run.py:_console_start():2463] atexit reg
+2026-01-24 11:05:19,226 INFO    MainThread:697635 [wandb_run.py:_redirect():2311] redirect: wrap_raw
+2026-01-24 11:05:19,227 INFO    MainThread:697635 [wandb_run.py:_redirect():2376] Wrapping output streams.
+2026-01-24 11:05:19,227 INFO    MainThread:697635 [wandb_run.py:_redirect():2401] Redirects installed.
+2026-01-24 11:05:19,228 INFO    MainThread:697635 [wandb_init.py:init():911] run started, returning control to user process
+2026-01-24 11:05:19,229 INFO    MainThread:697635 [wandb_run.py:_config_callback():1390] config_cb None None {'allow_tf32': True, 'logdir': 'logs', 'mixed_precision': 'bf16', 'num_checkpoint_limit': 5, 'num_epochs': 300, 'pretrained': {'model': './data/StableDiffusion', 'revision': 'main'}, 'prompt_fn': 'imagenet_animals', 'prompt_fn_kwargs': {}, 'resume_from': '', 'reward_fn': 'hpsv2', 'run_name': '2026.01.24_11.05.16', 'sample': {'batch_size': 1, 'eta': 1.0, 'guidance_scale': 5.0, 'num_batches_per_epoch': 2, 'num_steps': 50}, 'save_freq': 20, 'seed': 42, 'train': {'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'adam_weight_decay': 0.0001, 'adv_clip_max': 5, 'batch_size': 1, 'cfg': True, 'clip_range': 0.0001, 'gradient_accumulation_steps': 1, 'learning_rate': 1e-05, 'max_grad_norm': 1.0, 'num_inner_epochs': 1, 'timestep_fraction': 1.0, 'use_8bit_adam': False}, 'use_lora': False}
+2026-01-24 11:05:34,023 WARNING MsgRouterThr:697635 [router.py:message_loop():77] message_loop has been closed

wandb/run-20260124_141237-3a51ocuw/files/config.yaml

@@ -0,0 +1,138 @@
+_wandb:
+    value:
+        cli_version: 0.18.5
+        m: []
+        python_version: 3.10.19
+        t:
+            "1":
+                - 1
+                - 11
+                - 41
+                - 49
+                - 55
+                - 63
+                - 71
+                - 83
+                - 98
+            "2":
+                - 1
+                - 11
+                - 41
+                - 49
+                - 55
+                - 63
+                - 71
+                - 83
+                - 98
+            "3":
+                - 16
+                - 23
+                - 55
+            "4": 3.10.19
+            "5": 0.18.5
+            "6": 4.46.1
+            "8":
+                - 5
+            "12": 0.18.5
+            "13": linux-x86_64
+adv_clip_max:
+    value: 5
+cfg:
+    value: 0
+checkpointing_steps:
+    value: 50
+clip_range:
+    value: 0.0001
+data_json_path:
+    value: ./data/rl_embeddings/videos2caption.json
+dataloader_num_workers:
+    value: 4
+eta:
+    value: 0.7
+eta_step_list:
+    value:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+eta_step_merge_list:
+    value:
+        - 1
+        - 1
+        - 1
+        - 2
+        - 2
+        - 2
+        - 3
+        - 3
+fsdp_sharding_startegy:
+    value: full
+gradient_checkpointing:
+    value: false
+granular_list:
+    value:
+        - 1
+h:
+    value: 1024
+hps_clip_path:
+    value: ./data/hps/open_clip_pytorch_model.bin
+hps_path:
+    value: ./data/hps/HPS_v2.1_compressed.pt
+init_same_noise:
+    value: true
+learning_rate:
+    value: 2e-06
+lr_num_cycles:
+    value: 1
+lr_power:
+    value: 1
+lr_scheduler:
+    value: constant_with_warmup
+lr_warmup_steps:
+    value: 0
+max_grad_norm:
+    value: 1
+max_train_steps:
+    value: 301
+mixed_precision:
+    value: bf16
+num_generations:
+    value: 12
+num_latent_t:
+    value: 1
+output_dir:
+    value: ./output/g2rpo_qwenimage
+pretrained_model_name_or_path:
+    value: ./data/QwenImage
+sampler_seed:
+    value: 1223627
+sampling_steps:
+    value: 16
+seed:
+    value: 42
+selective_checkpointing:
+    value: 1
+shift:
+    value: 3
+sp_size:
+    value: 1
+t:
+    value: 1
+train_batch_size:
+    value: 1
+train_sp_batch_size:
+    value: 1
+use_cpu_offload:
+    value: false
+use_hpsv2:
+    value: true
+use_hpsv3:
+    value: false
+w:
+    value: 1024
+weight_decay:
+    value: 0.0001

wandb/run-20260124_141237-3a51ocuw/files/output.log

@@ -0,0 +1,56 @@
+***** Running E-GRPO (G2RPO) Training for QwenImage *****
+  Num examples = 50000
+  Dataloader size = 8334
+  Eta step list = [0, 1, 2, 3, 4, 5, 6, 7]
+  Eta step merge list = [1, 1, 1, 2, 2, 2, 3, 3]
+  Granular list = [1]
+  Total training parameters per FSDP shard = 3.405066888 B
+Steps:   0%|                                                                   | 0/100000 [00:00<?, ?it/s]Traceback (most recent call last):
+  File "/data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/fastvideo/train_g2rpo_qwenimage_merge.py", line 952, in <module>
+    main(args)
+  File "/data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/fastvideo/train_g2rpo_qwenimage_merge.py", line 834, in main
+    for step, (prompt_embeds, prompt_attention_masks, caption, original_length) in enumerate(train_dataloader):
+  File "/home/zsj/anaconda3/envs/g2rpo/lib/python3.10/site-packages/torch/utils/data/dataloader.py", line 732, in __next__
+    data = self._next_data()
+  File "/home/zsj/anaconda3/envs/g2rpo/lib/python3.10/site-packages/torch/utils/data/dataloader.py", line 1506, in _next_data
+    return self._process_data(data, worker_id)
+  File "/home/zsj/anaconda3/envs/g2rpo/lib/python3.10/site-packages/torch/utils/data/dataloader.py", line 1541, in _process_data
+    data.reraise()
+  File "/home/zsj/anaconda3/envs/g2rpo/lib/python3.10/site-packages/torch/_utils.py", line 769, in reraise
+    raise exception
+KeyError: Caught KeyError in DataLoader worker process 0.
+Original Traceback (most recent call last):
+  File "/home/zsj/anaconda3/envs/g2rpo/lib/python3.10/site-packages/torch/utils/data/_utils/worker.py", line 349, in _worker_loop
+    data = fetcher.fetch(index)  # type: ignore[possibly-undefined]
+  File "/home/zsj/anaconda3/envs/g2rpo/lib/python3.10/site-packages/torch/utils/data/_utils/fetch.py", line 52, in fetch
+    data = [self.dataset[idx] for idx in possibly_batched_index]
+  File "/home/zsj/anaconda3/envs/g2rpo/lib/python3.10/site-packages/torch/utils/data/_utils/fetch.py", line 52, in <listcomp>
+    data = [self.dataset[idx] for idx in possibly_batched_index]
+  File "/data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/fastvideo/dataset/latent_qwenimage_rl_datasets.py", line 50, in __getitem__
+    prompt_attention_mask_file = self.data_anno[idx]["prompt_attention_mask"]
+KeyError: 'prompt_attention_mask'
+
+[rank0]: Traceback (most recent call last):
+[rank0]:   File "/data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/fastvideo/train_g2rpo_qwenimage_merge.py", line 952, in <module>
+[rank0]:     main(args)
+[rank0]:   File "/data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/fastvideo/train_g2rpo_qwenimage_merge.py", line 834, in main
+[rank0]:     for step, (prompt_embeds, prompt_attention_masks, caption, original_length) in enumerate(train_dataloader):
+[rank0]:   File "/home/zsj/anaconda3/envs/g2rpo/lib/python3.10/site-packages/torch/utils/data/dataloader.py", line 732, in __next__
+[rank0]:     data = self._next_data()
+[rank0]:   File "/home/zsj/anaconda3/envs/g2rpo/lib/python3.10/site-packages/torch/utils/data/dataloader.py", line 1506, in _next_data
+[rank0]:     return self._process_data(data, worker_id)
+[rank0]:   File "/home/zsj/anaconda3/envs/g2rpo/lib/python3.10/site-packages/torch/utils/data/dataloader.py", line 1541, in _process_data
+[rank0]:     data.reraise()
+[rank0]:   File "/home/zsj/anaconda3/envs/g2rpo/lib/python3.10/site-packages/torch/_utils.py", line 769, in reraise
+[rank0]:     raise exception
+[rank0]: KeyError: Caught KeyError in DataLoader worker process 0.
+[rank0]: Original Traceback (most recent call last):
+[rank0]:   File "/home/zsj/anaconda3/envs/g2rpo/lib/python3.10/site-packages/torch/utils/data/_utils/worker.py", line 349, in _worker_loop
+[rank0]:     data = fetcher.fetch(index)  # type: ignore[possibly-undefined]
+[rank0]:   File "/home/zsj/anaconda3/envs/g2rpo/lib/python3.10/site-packages/torch/utils/data/_utils/fetch.py", line 52, in fetch
+[rank0]:     data = [self.dataset[idx] for idx in possibly_batched_index]
+[rank0]:   File "/home/zsj/anaconda3/envs/g2rpo/lib/python3.10/site-packages/torch/utils/data/_utils/fetch.py", line 52, in <listcomp>
+[rank0]:     data = [self.dataset[idx] for idx in possibly_batched_index]
+[rank0]:   File "/data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/fastvideo/dataset/latent_qwenimage_rl_datasets.py", line 50, in __getitem__
+[rank0]:     prompt_attention_mask_file = self.data_anno[idx]["prompt_attention_mask"]
+[rank0]: KeyError: 'prompt_attention_mask'

wandb/run-20260124_141237-3a51ocuw/files/requirements.txt

@@ -0,0 +1,191 @@
+scipy==1.13.0
+regex==2024.9.11
+sentencepiece==0.2.0
+six==1.16.0
+anyio==4.11.0
+nvidia-cuda-nvrtc-cu12==12.6.77
+scikit-video==1.1.11
+platformdirs==4.5.0
+mypy==1.11.1
+ruff==0.6.5
+charset-normalizer==3.4.4
+torch==2.9.0+cu126
+av==13.1.0
+pillow==10.2.0
+gpustat==1.1.1
+torchvision==0.24.0+cu126
+multidict==6.7.0
+torchmetrics==1.5.1
+aiohttp==3.13.1
+transformers==4.46.1
+decord==0.6.0
+wcwidth==0.2.14
+sphinx-lint==1.0.0
+nvidia-cuda-runtime-cu12==12.6.77
+pytz==2025.2
+codespell==2.3.0
+hpsv2==1.2.0
+mypy_extensions==1.1.0
+numpy==1.26.3
+omegaconf==2.3.0
+Markdown==3.9
+tzdata==2025.2
+pandas==2.2.3
+pytorch-lightning==2.4.0
+aiosignal==1.4.0
+aiohappyeyeballs==2.6.1
+python-dateutil==2.9.0.post0
+seaborn==0.13.2
+beautifulsoup4==4.12.3
+isort==5.13.2
+httpx==0.28.1
+certifi==2025.10.5
+ml_collections==1.1.0
+nvidia-cudnn-cu12==9.10.2.21
+peft==0.18.1
+hf-xet==1.2.0
+requests==2.31.0
+inflect==6.0.4
+iniconfig==2.1.0
+braceexpand==0.1.7
+h5py==3.12.1
+wandb==0.18.5
+protobuf==3.20.3
+ninja==1.13.0
+kiwisolver==1.4.9
+networkx==3.3
+packaging==25.0
+fvcore==0.1.5.post20221221
+pyparsing==3.2.5
+starlette==0.41.3
+frozenlist==1.8.0
+docker-pycreds==0.4.0
+Werkzeug==3.1.3
+MarkupSafe==2.1.5
+shellingham==1.5.4
+einops==0.8.0
+sentry-sdk==2.42.0
+PyYAML==6.0.1
+nvidia-nccl-cu12==2.27.5
+datasets==4.3.0
+polib==1.2.0
+safetensors==0.6.2
+async-timeout==5.0.1
+setproctitle==1.3.7
+clint==0.5.1
+matplotlib==3.9.2
+propcache==0.4.1
+termcolor==3.1.0
+antlr4-python3-runtime==4.9.3
+cycler==0.12.1
+fastvideo==1.2.0
+toml==0.10.2
+xxhash==3.6.0
+wheel==0.44.0
+albumentations==1.4.20
+fastapi==0.115.3
+nvidia-cufft-cu12==11.3.0.4
+yarl==1.22.0
+psutil==7.1.0
+tensorboard-data-server==0.7.2
+huggingface-hub==0.36.0
+pydantic==2.9.2
+nvidia-nvtx-cu12==12.6.77
+portalocker==3.2.0
+triton==3.5.0
+annotated-types==0.7.0
+proglog==0.1.12
+nvidia-cusparselt-cu12==0.7.1
+yapf==0.32.0
+Jinja2==3.1.6
+types-requests==2.32.4.20250913
+lightning-utilities==0.15.2
+grpcio==1.75.1
+uvicorn==0.32.0
+typing_extensions==4.15.0
+nvidia-nvjitlink-cu12==12.6.85
+watch==0.2.7
+moviepy==1.0.3
+timm==1.0.11
+pytest-split==0.8.0
+gdown==5.2.0
+types-setuptools==80.9.0.20250822
+nvidia-cusolver-cu12==11.7.1.2
+types-PyYAML==6.0.12.20250915
+pip==25.2
+typer-slim==0.21.1
+qwen-vl-utils==0.0.14
+soupsieve==2.8
+zipp==3.23.0
+flash_attn==2.8.3
+yacs==0.1.8
+pluggy==1.6.0
+opencv-python-headless==4.11.0.86
+mpmath==1.3.0
+test_tube==0.7.5
+stringzilla==4.2.1
+fonttools==4.60.1
+nvidia-ml-py==13.580.82
+parameterized==0.9.0
+loguru==0.7.3
+diffusers==0.36.0
+tabulate==0.9.0
+idna==3.6
+iopath==0.1.10
+decorator==4.4.2
+nvidia-cufile-cu12==1.11.1.6
+threadpoolctl==3.6.0
+pyarrow==21.0.0
+httpcore==1.0.9
+hydra-core==1.3.2
+multiprocess==0.70.16
+contourpy==1.3.2
+clip==1.0
+tqdm==4.66.5
+open_clip_torch==3.2.0
+accelerate==1.0.1
+gitdb==4.0.12
+importlib_metadata==8.7.0
+nvidia-cublas-cu12==12.6.4.1
+h11==0.16.0
+filelock==3.19.1
+liger_kernel==0.4.1
+click==8.3.0
+urllib3==2.2.0
+imageio-ffmpeg==0.5.1
+setuptools==80.9.0
+joblib==1.5.2
+tensorboard==2.20.0
+attrs==25.4.0
+future==1.0.0
+albucore==0.0.19
+fsspec==2025.9.0
+sympy==1.14.0
+eval_type_backport==0.2.2
+pydantic_core==2.23.4
+sniffio==1.3.1
+nvidia-nvshmem-cu12==3.3.20
+exceptiongroup==1.3.0
+smmap==5.0.2
+tomli==2.0.2
+ftfy==6.3.0
+dill==0.4.0
+pytest==7.2.0
+PySocks==1.7.1
+nvidia-curand-cu12==10.3.7.77
+tokenizers==0.20.1
+args==0.1.0
+fairscale==0.4.13
+webdataset==1.0.2
+GitPython==3.1.45
+pytorchvideo==0.1.5
+scikit-learn==1.5.2
+bitsandbytes==0.48.1
+nvidia-cusparse-cu12==12.5.4.2
+nvidia-cuda-cupti-cu12==12.6.80
+imageio==2.36.0
+pydub==0.25.1
+image-reward==1.5
+absl-py==2.3.1
+blessed==1.22.0
+torchdiffeq==0.2.4

wandb/run-20260124_141237-3a51ocuw/files/wandb-metadata.json

@@ -0,0 +1,137 @@
+{
+  "os":  "Linux-6.8.0-86-generic-x86_64-with-glibc2.35",
+  "python":  "3.10.19",
+  "startedAt":  "2026-01-24T06:12:37.513847Z",
+  "args":  [
+    "--pretrained_model_name_or_path",
+    "./data/QwenImage",
+    "--data_json_path",
+    "./data/rl_embeddings/videos2caption.json",
+    "--output_dir",
+    "./output/g2rpo_qwenimage",
+    "--hps_path",
+    "./data/hps/HPS_v2.1_compressed.pt",
+    "--hps_clip_path",
+    "./data/hps/open_clip_pytorch_model.bin",
+    "--h",
+    "1024",
+    "--w",
+    "1024",
+    "--sampling_steps",
+    "16",
+    "--eta",
+    "0.7",
+    "--shift",
+    "3.0",
+    "--num_generations",
+    "12",
+    "--learning_rate",
+    "2e-6",
+    "--max_train_steps",
+    "301",
+    "--checkpointing_steps",
+    "50",
+    "--eta_step_list",
+    "0",
+    "1",
+    "2",
+    "3",
+    "4",
+    "5",
+    "6",
+    "7",
+    "--eta_step_merge_list",
+    "1",
+    "1",
+    "1",
+    "2",
+    "2",
+    "2",
+    "3",
+    "3",
+    "--granular_list",
+    "1",
+    "--init_same_noise",
+    "--clip_range",
+    "1e-4",
+    "--adv_clip_max",
+    "5.0",
+    "--use_hpsv2"
+  ],
+  "program":  "/data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/fastvideo/train_g2rpo_qwenimage_merge.py",
+  "codePath":  "fastvideo/train_g2rpo_qwenimage_merge.py",
+  "email":  "zhangemail1428@163.com",
+  "root":  "/data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code",
+  "host":  "abc",
+  "username":  "zsj",
+  "executable":  "/home/zsj/anaconda3/envs/g2rpo/bin/python",
+  "codePathLocal":  "fastvideo/train_g2rpo_qwenimage_merge.py",
+  "cpu_count":  48,
+  "cpu_count_logical":  96,
+  "gpu":  "NVIDIA RTX 5880 Ada Generation",
+  "gpu_count":  8,
+  "disk":  {
+    "/":  {
+      "total":  "1006773899264",
+      "used":  "803118862336"
+    }
+  },
+  "memory":  {
+    "total":  "540697153536"
+  },
+  "cpu":  {
+    "count":  48,
+    "countLogical":  96
+  },
+  "gpu_nvidia":  [
+    {
+      "name":  "NVIDIA RTX 5880 Ada Generation",
+      "memoryTotal":  "51527024640",
+      "cudaCores":  14080,
+      "architecture":  "Ada"
+    },
+    {
+      "name":  "NVIDIA RTX 5880 Ada Generation",
+      "memoryTotal":  "51527024640",
+      "cudaCores":  14080,
+      "architecture":  "Ada"
+    },
+    {
+      "name":  "NVIDIA RTX 5880 Ada Generation",
+      "memoryTotal":  "51527024640",
+      "cudaCores":  14080,
+      "architecture":  "Ada"
+    },
+    {
+      "name":  "NVIDIA RTX 5880 Ada Generation",
+      "memoryTotal":  "51527024640",
+      "cudaCores":  14080,
+      "architecture":  "Ada"
+    },
+    {
+      "name":  "NVIDIA RTX 5880 Ada Generation",
+      "memoryTotal":  "51527024640",
+      "cudaCores":  14080,
+      "architecture":  "Ada"
+    },
+    {
+      "name":  "NVIDIA RTX 5880 Ada Generation",
+      "memoryTotal":  "51527024640",
+      "cudaCores":  14080,
+      "architecture":  "Ada"
+    },
+    {
+      "name":  "NVIDIA RTX 5880 Ada Generation",
+      "memoryTotal":  "51527024640",
+      "cudaCores":  14080,
+      "architecture":  "Ada"
+    },
+    {
+      "name":  "NVIDIA RTX 5880 Ada Generation",
+      "memoryTotal":  "51527024640",
+      "cudaCores":  14080,
+      "architecture":  "Ada"
+    }
+  ],
+  "cudaVersion":  "12.9"
+}

wandb/run-20260124_141237-3a51ocuw/files/wandb-summary.json

@@ -0,0 +1 @@
+{"_wandb":{"runtime":1}}

wandb/run-20260124_141237-3a51ocuw/logs/debug-core.log

@@ -0,0 +1,15 @@
+{"time":"2026-01-24T14:12:36.403578289+08:00","level":"INFO","msg":"started logging, with flags","port-filename":"/tmp/tmp_5qzahaq/port-28129.txt","pid":28129,"debug":false,"disable-analytics":false}
+{"time":"2026-01-24T14:12:36.403616892+08:00","level":"INFO","msg":"FeatureState","shutdownOnParentExitEnabled":false}
+{"time":"2026-01-24T14:12:36.404894071+08:00","level":"INFO","msg":"Will exit if parent process dies.","ppid":28129}
+{"time":"2026-01-24T14:12:36.404999637+08:00","level":"INFO","msg":"server is running","addr":{"IP":"127.0.0.1","Port":38405,"Zone":""}}
+{"time":"2026-01-24T14:12:36.572515476+08:00","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"127.0.0.1:41386"}
+{"time":"2026-01-24T14:12:37.508533932+08:00","level":"INFO","msg":"handleInformInit: received","streamId":"3a51ocuw","id":"127.0.0.1:41386"}
+{"time":"2026-01-24T14:12:37.63058458+08:00","level":"INFO","msg":"handleInformInit: stream started","streamId":"3a51ocuw","id":"127.0.0.1:41386"}
+{"time":"2026-01-24T14:12:38.683918448+08:00","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"127.0.0.1:41386"}
+{"time":"2026-01-24T14:12:38.684070643+08:00","level":"INFO","msg":"connection: Close: initiating connection closure","id":"127.0.0.1:41386"}
+{"time":"2026-01-24T14:12:38.684177673+08:00","level":"INFO","msg":"server is shutting down"}
+{"time":"2026-01-24T14:12:38.684241348+08:00","level":"INFO","msg":"connection: Close: connection successfully closed","id":"127.0.0.1:41386"}
+{"time":"2026-01-24T14:12:38.772325053+08:00","level":"ERROR","msg":"processOutgoingData: flush error","error":"write tcp 127.0.0.1:38405->127.0.0.1:41386: use of closed network connection","id":"127.0.0.1:41386"}
+{"time":"2026-01-24T14:12:40.353470819+08:00","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"127.0.0.1:41386"}
+{"time":"2026-01-24T14:12:40.353514215+08:00","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"127.0.0.1:41386"}
+{"time":"2026-01-24T14:12:40.353538791+08:00","level":"INFO","msg":"server is closed"}

wandb/run-20260124_141237-3a51ocuw/logs/debug-internal.log

@@ -0,0 +1,19 @@
+{"time":"2026-01-24T14:12:37.509015822+08:00","level":"INFO","msg":"using version","core version":"0.18.5"}
+{"time":"2026-01-24T14:12:37.509057782+08:00","level":"INFO","msg":"created symlink","path":"/data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/wandb/run-20260124_141237-3a51ocuw/logs/debug-core.log"}
+{"time":"2026-01-24T14:12:37.630545922+08:00","level":"INFO","msg":"created new stream","id":"3a51ocuw"}
+{"time":"2026-01-24T14:12:37.630581136+08:00","level":"INFO","msg":"stream: started","id":"3a51ocuw"}
+{"time":"2026-01-24T14:12:37.630856728+08:00","level":"INFO","msg":"writer: Do: started","stream_id":{"value":"3a51ocuw"}}
+{"time":"2026-01-24T14:12:37.630945854+08:00","level":"INFO","msg":"sender: started","stream_id":"3a51ocuw"}
+{"time":"2026-01-24T14:12:37.630815002+08:00","level":"INFO","msg":"handler: started","stream_id":{"value":"3a51ocuw"}}
+{"time":"2026-01-24T14:12:38.30268329+08:00","level":"INFO","msg":"Starting system monitor"}
+{"time":"2026-01-24T14:12:38.684097398+08:00","level":"INFO","msg":"stream: closing","id":"3a51ocuw"}
+{"time":"2026-01-24T14:12:38.684210233+08:00","level":"INFO","msg":"Stopping system monitor"}
+{"time":"2026-01-24T14:12:38.685194374+08:00","level":"INFO","msg":"Stopped system monitor"}
+{"time":"2026-01-24T14:12:39.357336767+08:00","level":"WARN","msg":"No job ingredients found, not creating job artifact"}
+{"time":"2026-01-24T14:12:39.357370894+08:00","level":"WARN","msg":"No source type found, not creating job artifact"}
+{"time":"2026-01-24T14:12:39.357389716+08:00","level":"INFO","msg":"sender: sendDefer: no job artifact to save"}
+{"time":"2026-01-24T14:12:40.006428963+08:00","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2026-01-24T14:12:40.353157928+08:00","level":"INFO","msg":"handler: closed","stream_id":{"value":"3a51ocuw"}}
+{"time":"2026-01-24T14:12:40.353244646+08:00","level":"INFO","msg":"writer: Close: closed","stream_id":{"value":"3a51ocuw"}}
+{"time":"2026-01-24T14:12:40.353290403+08:00","level":"INFO","msg":"sender: closed","stream_id":"3a51ocuw"}
+{"time":"2026-01-24T14:12:40.353339352+08:00","level":"INFO","msg":"stream: closed","id":"3a51ocuw"}

wandb/run-20260124_141237-3a51ocuw/logs/debug.log

@@ -0,0 +1,26 @@
+2026-01-24 14:12:37,499 INFO    MainThread:28129 [wandb_setup.py:_flush():79] Current SDK version is 0.18.5
+2026-01-24 14:12:37,500 INFO    MainThread:28129 [wandb_setup.py:_flush():79] Configure stats pid to 28129
+2026-01-24 14:12:37,500 INFO    MainThread:28129 [wandb_setup.py:_flush():79] Loading settings from /home/zsj/.config/wandb/settings
+2026-01-24 14:12:37,500 INFO    MainThread:28129 [wandb_setup.py:_flush():79] Loading settings from /data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/wandb/settings
+2026-01-24 14:12:37,500 INFO    MainThread:28129 [wandb_setup.py:_flush():79] Loading settings from environment variables: {}
+2026-01-24 14:12:37,500 INFO    MainThread:28129 [wandb_setup.py:_flush():79] Applying setup settings: {'mode': None, '_disable_service': None}
+2026-01-24 14:12:37,500 INFO    MainThread:28129 [wandb_setup.py:_flush():79] Inferring run settings from compute environment: {'program_relpath': 'fastvideo/train_g2rpo_qwenimage_merge.py', 'program_abspath': '/data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/fastvideo/train_g2rpo_qwenimage_merge.py', 'program': '/data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/fastvideo/train_g2rpo_qwenimage_merge.py'}
+2026-01-24 14:12:37,500 INFO    MainThread:28129 [wandb_setup.py:_flush():79] Applying login settings: {}
+2026-01-24 14:12:37,500 INFO    MainThread:28129 [wandb_init.py:_log_setup():534] Logging user logs to /data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/wandb/run-20260124_141237-3a51ocuw/logs/debug.log
+2026-01-24 14:12:37,501 INFO    MainThread:28129 [wandb_init.py:_log_setup():535] Logging internal logs to /data1/zsj/SceneDPO/Rebuttal/E-GRPO/scoure_code/wandb/run-20260124_141237-3a51ocuw/logs/debug-internal.log
+2026-01-24 14:12:37,501 INFO    MainThread:28129 [wandb_init.py:init():621] calling init triggers
+2026-01-24 14:12:37,501 INFO    MainThread:28129 [wandb_init.py:init():628] wandb.init called with sweep_config: {}
+config: {'data_json_path': './data/rl_embeddings/videos2caption.json', 'dataloader_num_workers': 4, 'train_batch_size': 1, 'num_latent_t': 1, 'pretrained_model_name_or_path': './data/QwenImage', 'hps_path': './data/hps/HPS_v2.1_compressed.pt', 'hps_clip_path': './data/hps/open_clip_pytorch_model.bin', 'cfg': 0.0, 'seed': 42, 'output_dir': './output/g2rpo_qwenimage', 'checkpointing_steps': 50, 'max_train_steps': 301, 'learning_rate': 2e-06, 'lr_warmup_steps': 0, 'max_grad_norm': 1.0, 'gradient_checkpointing': False, 'selective_checkpointing': 1.0, 'mixed_precision': 'bf16', 'use_cpu_offload': False, 'sp_size': 1, 'train_sp_batch_size': 1, 'fsdp_sharding_startegy': 'full', 'lr_scheduler': 'constant_with_warmup', 'lr_num_cycles': 1, 'lr_power': 1.0, 'weight_decay': 0.0001, 'h': 1024, 'w': 1024, 't': 1, 'sampling_steps': 16, 'eta': 0.7, 'sampler_seed': 1223627, 'num_generations': 12, 'shift': 3.0, 'init_same_noise': True, 'clip_range': 0.0001, 'adv_clip_max': 5.0, 'eta_step_list': [0, 1, 2, 3, 4, 5, 6, 7], 'eta_step_merge_list': [1, 1, 1, 2, 2, 2, 3, 3], 'granular_list': [1], 'use_hpsv2': True, 'use_hpsv3': False}
+2026-01-24 14:12:37,501 INFO    MainThread:28129 [wandb_init.py:init():671] starting backend
+2026-01-24 14:12:37,501 INFO    MainThread:28129 [wandb_init.py:init():675] sending inform_init request
+2026-01-24 14:12:37,506 INFO    MainThread:28129 [backend.py:_multiprocessing_setup():104] multiprocessing start_methods=fork,spawn,forkserver, using: spawn
+2026-01-24 14:12:37,509 INFO    MainThread:28129 [wandb_init.py:init():688] backend started and connected
+2026-01-24 14:12:37,676 INFO    MainThread:28129 [wandb_init.py:init():783] updated telemetry
+2026-01-24 14:12:37,678 INFO    MainThread:28129 [wandb_init.py:init():816] communicating run to backend with 90.0 second timeout
+2026-01-24 14:12:38,296 INFO    MainThread:28129 [wandb_init.py:init():867] starting run threads in backend
+2026-01-24 14:12:38,502 INFO    MainThread:28129 [wandb_run.py:_console_start():2463] atexit reg
+2026-01-24 14:12:38,503 INFO    MainThread:28129 [wandb_run.py:_redirect():2311] redirect: wrap_raw
+2026-01-24 14:12:38,504 INFO    MainThread:28129 [wandb_run.py:_redirect():2376] Wrapping output streams.
+2026-01-24 14:12:38,505 INFO    MainThread:28129 [wandb_run.py:_redirect():2401] Redirects installed.
+2026-01-24 14:12:38,512 INFO    MainThread:28129 [wandb_init.py:init():911] run started, returning control to user process
+2026-01-24 14:12:38,684 WARNING MsgRouterThr:28129 [router.py:message_loop():77] message_loop has been closed