Upload SAGE-GRPO outputs vbench2_launch

outputs/vbench2_launch/codex_dryrun/generation_manifest.json

@@ -0,0 +1,9 @@
+[
+  {
+    "index": 1,
+    "prompt": "A man is doing yoga.",
+    "seed": 930,
+    "sample_index": 0,
+    "dimension": "Human_Anatomy"
+  }
+]

outputs/vbench2_launch/codex_dryrun/logs/launch.log

@@ -0,0 +1,22 @@
+[2026-05-20 17:37:34] Run directory: /nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/codex_dryrun
+[2026-05-20 17:37:34] SAGE_DIR: /nfs/ywang29/Reward_finetuning/SAGE-GRPO
+[2026-05-20 17:37:34] VBENCH_DIR: /nfs/ywang29/Reward_finetuning/VideoX-Fun/VBench/VBench-2.0
+[2026-05-20 17:37:34] DIMENSIONS: Human_Anatomy
+[2026-05-20 17:37:34] EVAL_MODE: custom_input
+[2026-05-20 17:37:34] RUN_SETUP=false; assuming environments are already configured.
+[2026-05-20 17:37:34] RUN_SMOKE_TEST=false; skipping smoke tests.
+[2026-05-20 17:37:34] Preparing VBench2 prompt CSV and manifest.
+selected_dimensions: {'Human_Anatomy': 1}
+generation_rows: 1
+prompt_csv: /nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/codex_dryrun/vbench2_prompts.csv
+manifest: /nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/codex_dryrun/generation_manifest.json
+selected_info_json: /nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/codex_dryrun/vbench2_selected_info.json
+[2026-05-20 17:37:34] RUN_GENERATE=false; skipping generation.
+[2026-05-20 17:37:34] RUN_STAGE=false; skipping staging.
+[2026-05-20 17:37:34] RUN_VBENCH_EVAL=false; skipping real VBench2 metric scoring.
+[2026-05-20 17:37:34] Done.
+[2026-05-20 17:37:34] Prompt CSV: /nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/codex_dryrun/vbench2_prompts.csv
+[2026-05-20 17:37:34] Generated output: /nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/codex_dryrun/generated/raw_generated.mp4
+[2026-05-20 17:37:34] Staged videos: /nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/codex_dryrun/vbench2_staged_videos
+[2026-05-20 17:37:34] Eval output: /nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/codex_dryrun/vbench2_eval_results
+[2026-05-20 17:37:34] Log: /nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/codex_dryrun/logs/launch.log

outputs/vbench2_launch/codex_dryrun/vbench2_prompts.csv

@@ -0,0 +1,2 @@
+index,prompt,seed
+1,A man is doing yoga.,930

outputs/vbench2_launch/codex_dryrun/vbench2_selected_info.json

@@ -0,0 +1,8 @@
+[
+  {
+    "prompt_en": "A man is doing yoga.",
+    "dimension": [
+      "Human_Anatomy"
+    ]
+  }
+]

outputs/vbench2_launch/codex_smoke_20260520_183314/logs/launch.log

@@ -0,0 +1,67 @@
+[2026-05-20 18:33:14] Run directory: /nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/codex_smoke_20260520_183314
+[2026-05-20 18:33:14] SAGE_DIR: /nfs/ywang29/Reward_finetuning/SAGE-GRPO
+[2026-05-20 18:33:14] VBENCH_DIR: /nfs/ywang29/Reward_finetuning/VideoX-Fun/VBench/VBench-2.0
+[2026-05-20 18:33:14] DIMENSIONS: Human_Anatomy
+[2026-05-20 18:33:14] EVAL_MODE: custom_input
+[2026-05-20 18:33:14] RUN_SETUP=false; assuming environments are already configured.
+[2026-05-20 18:33:14] Running SAGE generation import smoke test.
+torch: 2.10.0+cu130
+torchvision: 0.25.0+cu130
+transformers: 4.57.1
+diffusers: 0.38.0
+decord: 0.6.0
+cv2: 4.13.0
+cuda_available: True
+cuda_version: 13.0
+[rank0]:[W520 18:33:28.642617685 ProcessGroupNCCL.cpp:1553] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+ffmpeg version 4.4.2-0ubuntu0.22.04.1 Copyright (c) 2000-2021 the FFmpeg developers
+[2026-05-20 18:33:28] Running VBench2 import and metadata smoke test.
+ffmpeg: /opt/miniconda3/envs/vbench2/lib/libncursesw.so.6: no version information available (required by /lib/x86_64-linux-gnu/libcaca.so.0)
+ffmpeg: /opt/miniconda3/envs/vbench2/lib/libncursesw.so.6: no version information available (required by /lib/x86_64-linux-gnu/libcaca.so.0)
+ffmpeg: /opt/miniconda3/envs/vbench2/lib/libtinfo.so.6: no version information available (required by /lib/x86_64-linux-gnu/libcaca.so.0)
+torch: 2.5.1+cu121
+torchvision: 0.20.1+cu121
+decord: 0.6.0
+cv2: 4.11.0
+mmcv: 2.2.0
+Traceback (most recent call last):
+  File "<stdin>", line 7, in <module>
+  File "/opt/miniconda3/envs/vbench2/lib/python3.10/importlib/__init__.py", line 126, in import_module
+    return _bootstrap._gcd_import(name[level:], package, level)
+  File "<frozen importlib._bootstrap>", line 1050, in _gcd_import
+  File "<frozen importlib._bootstrap>", line 1027, in _find_and_load
+  File "<frozen importlib._bootstrap>", line 1006, in _find_and_load_unlocked
+  File "<frozen importlib._bootstrap>", line 688, in _load_unlocked
+  File "<frozen importlib._bootstrap_external>", line 883, in exec_module
+  File "<frozen importlib._bootstrap>", line 241, in _call_with_frames_removed
+  File "/opt/miniconda3/envs/vbench2/lib/python3.10/site-packages/mmyolo/__init__.py", line 3, in <module>
+    import mmdet
+  File "/opt/miniconda3/envs/vbench2/lib/python3.10/site-packages/mmdet/__init__.py", line 17, in <module>
+    and mmcv_version < digit_version(mmcv_maximum_version)), \
+AssertionError: MMCV==2.2.0 is used but incompatible. Please install mmcv>=2.0.0rc4, <2.1.0.
+[ERROR] line 373: SMOKE_DIR="$smoke_dir" SMOKE_OUT="${RUN_DIR}/smoke_vbench2" python - <<'PY'
+import importlib
+import os
+import torch
+from pathlib import Path
+
+for name in ["torch", "torchvision", "decord", "cv2", "mmcv", "mmyolo", "gdown", "vbench2"]:
+    mod = importlib.import_module(name)
+    print(f"{name}: {getattr(mod, '__version__', 'ok')}")
+
+from vbench2 import VBench2
+
+out = Path(os.environ["SMOKE_OUT"])
+out.mkdir(parents=True, exist_ok=True)
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+bench = VBench2(device, "vbench2/VBench2_full_info.json", str(out))
+meta = bench.build_full_info_json(
+    os.environ["SMOKE_DIR"],
+    "smoke_metadata",
+    ["Human_Anatomy"],
+    mode="custom_input",
+)
+assert Path(meta).is_file(), meta
+print(f"vbench2_metadata_ok: {meta}")
+PY
+

outputs/vbench2_launch/codex_smoke_20260520_183314/smoke_videos/smoke prompt-0.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f62102820217073de045275bec77001d935956edb547054dab95294c47d90152
+size 2785

outputs/vbench2_launch/codex_smoke_20260520_183414/generation_manifest.json

@@ -0,0 +1,9 @@
+[
+  {
+    "index": 1,
+    "prompt": "A man is doing yoga.",
+    "seed": 930,
+    "sample_index": 0,
+    "dimension": "Human_Anatomy"
+  }
+]

outputs/vbench2_launch/codex_smoke_20260520_183414/logs/launch.log

@@ -0,0 +1,49 @@
+[2026-05-20 18:34:15] Run directory: /nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/codex_smoke_20260520_183414
+[2026-05-20 18:34:15] SAGE_DIR: /nfs/ywang29/Reward_finetuning/SAGE-GRPO
+[2026-05-20 18:34:15] VBENCH_DIR: /nfs/ywang29/Reward_finetuning/VideoX-Fun/VBench/VBench-2.0
+[2026-05-20 18:34:15] DIMENSIONS: Human_Anatomy
+[2026-05-20 18:34:15] EVAL_MODE: custom_input
+[2026-05-20 18:34:15] RUN_SETUP=false; assuming environments are already configured.
+[2026-05-20 18:34:15] Running SAGE generation import smoke test.
+torch: 2.10.0+cu130
+torchvision: 0.25.0+cu130
+transformers: 4.57.1
+diffusers: 0.38.0
+decord: 0.6.0
+cv2: 4.13.0
+cuda_available: True
+cuda_version: 13.0
+[rank0]:[W520 18:34:27.430478372 ProcessGroupNCCL.cpp:1553] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+ffmpeg version 4.4.2-0ubuntu0.22.04.1 Copyright (c) 2000-2021 the FFmpeg developers
+[2026-05-20 18:34:28] Running VBench2 import and metadata smoke test.
+ffmpeg: /opt/miniconda3/envs/vbench2/lib/libncursesw.so.6: no version information available (required by /lib/x86_64-linux-gnu/libcaca.so.0)
+ffmpeg: /opt/miniconda3/envs/vbench2/lib/libncursesw.so.6: no version information available (required by /lib/x86_64-linux-gnu/libcaca.so.0)
+ffmpeg: /opt/miniconda3/envs/vbench2/lib/libtinfo.so.6: no version information available (required by /lib/x86_64-linux-gnu/libcaca.so.0)
+torch: 2.5.1+cu121
+torchvision: 0.20.1+cu121
+decord: 0.6.0
+cv2: 4.11.0
+mmcv: 2.2.0
+mmyolo: 0.6.0
+gdown: 5.2.0
+ffmpeg: /opt/miniconda3/envs/vbench2/lib/libncursesw.so.6: no version information available (required by /lib/x86_64-linux-gnu/libcaca.so.0)
+ffmpeg: /opt/miniconda3/envs/vbench2/lib/libncursesw.so.6: no version information available (required by /lib/x86_64-linux-gnu/libcaca.so.0)
+ffmpeg: /opt/miniconda3/envs/vbench2/lib/libtinfo.so.6: no version information available (required by /lib/x86_64-linux-gnu/libcaca.so.0)
+vbench2: ok
+Evaluation meta data saved to /nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/codex_smoke_20260520_183414/smoke_vbench2/smoke_metadata_full_info.json
+vbench2_metadata_ok: /nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/codex_smoke_20260520_183414/smoke_vbench2/smoke_metadata_full_info.json
+[2026-05-20 18:34:33] Preparing VBench2 prompt CSV and manifest.
+selected_dimensions: {'Human_Anatomy': 1}
+generation_rows: 1
+prompt_csv: /nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/codex_smoke_20260520_183414/vbench2_prompts.csv
+manifest: /nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/codex_smoke_20260520_183414/generation_manifest.json
+selected_info_json: /nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/codex_smoke_20260520_183414/vbench2_selected_info.json
+[2026-05-20 18:34:33] RUN_GENERATE=false; skipping generation.
+[2026-05-20 18:34:33] RUN_STAGE=false; skipping staging.
+[2026-05-20 18:34:33] RUN_VBENCH_EVAL=false; skipping real VBench2 metric scoring.
+[2026-05-20 18:34:33] Done.
+[2026-05-20 18:34:33] Prompt CSV: /nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/codex_smoke_20260520_183414/vbench2_prompts.csv
+[2026-05-20 18:34:33] Generated output: /nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/codex_smoke_20260520_183414/generated/raw_generated.mp4
+[2026-05-20 18:34:33] Staged videos: /nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/codex_smoke_20260520_183414/vbench2_staged_videos
+[2026-05-20 18:34:33] Eval output: /nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/codex_smoke_20260520_183414/vbench2_eval_results
+[2026-05-20 18:34:33] Log: /nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/codex_smoke_20260520_183414/logs/launch.log

outputs/vbench2_launch/codex_smoke_20260520_183414/smoke_vbench2/smoke_metadata_full_info.json

@@ -0,0 +1,11 @@
+[
+    {
+        "prompt_en": "smoke prompt",
+        "dimension": [
+            "Human_Anatomy"
+        ],
+        "video_list": [
+            "/nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/codex_smoke_20260520_183414/smoke_videos/smoke prompt-0.mp4"
+        ]
+    }
+]

outputs/vbench2_launch/codex_smoke_20260520_183414/smoke_videos/smoke prompt-0.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f62102820217073de045275bec77001d935956edb547054dab95294c47d90152
+size 2785

outputs/vbench2_launch/codex_smoke_20260520_183414/vbench2_prompts.csv

@@ -0,0 +1,2 @@
+index,prompt,seed
+1,A man is doing yoga.,930

outputs/vbench2_launch/codex_smoke_20260520_183414/vbench2_selected_info.json

@@ -0,0 +1,8 @@
+[
+  {
+    "prompt_en": "A man is doing yoga.",
+    "dimension": [
+      "Human_Anatomy"
+    ]
+  }
+]

outputs/vbench2_launch/sage_vbench2_20260520_175550/logs/launch.log

@@ -0,0 +1,26 @@
+[2026-05-20 17:55:50] Run directory: /nfs/ywang29/Reward_finetuning/SAGE-GRPO/outputs/vbench2_launch/sage_vbench2_20260520_175550
+[2026-05-20 17:55:50] SAGE_DIR: /nfs/ywang29/Reward_finetuning/SAGE-GRPO
+[2026-05-20 17:55:50] VBENCH_DIR: /nfs/ywang29/Reward_finetuning/VideoX-Fun/VBench/VBench-2.0
+[2026-05-20 17:55:50] DIMENSIONS: Human_Anatomy
+[2026-05-20 17:55:50] EVAL_MODE: custom_input
+[2026-05-20 17:55:51] Creating SAGE generation conda env: sage_grpo_py312
+
+CondaToSNonInteractiveError: Terms of Service have not been accepted for the following channels. Please accept or remove them before proceeding:
+    - https://repo.anaconda.com/pkgs/main
+    - https://repo.anaconda.com/pkgs/r
+
+To accept these channels' Terms of Service, run the following commands:
+    conda tos accept --override-channels --channel https://repo.anaconda.com/pkgs/main
+    conda tos accept --override-channels --channel https://repo.anaconda.com/pkgs/r
+
+For information on safely removing channels from your conda configuration,
+please see the official documentation:
+
+    https://www.anaconda.com/docs/tools/working-with-conda/channels
+
+[ERROR] line 154: "$CONDA_EXE" "$@"
+[ERROR] line 156: ( if [ -n "${_CE_M:+x}" ] && [ -n "${_CE_CONDA:+x}" ]; then
+    "$CONDA_EXE" $_CE_M $_CE_CONDA "$@";
+else
+    "$CONDA_EXE" "$@";
+fi )