debug.log

[05-13 16:34:17|DEBUG|packages/cosmos-oss/cosmos_oss/init.py:176:init_output_dir] Flags(internal=False, smoke=False, verbose=False, experimental_checkpoints=False)
[05-13 16:34:23|INFO|cosmos_predict2/_src/imaginaire/utils/checkpoint_db.py:187:path] Downloading checkpoint nvidia/Cosmos-Guardrail1(9c7b7da4-2d95-45bb-9cb8-2eed954e9736)
[05-13 16:34:23|INFO|cosmos_predict2/_src/imaginaire/utils/checkpoint_db.py:134:_download] Downloading checkpoint from Hugging Face with {'repo_id': 'nvidia/Cosmos-Guardrail1', 'repo_type': 'model', 'revision': 'd6d4bfa899a71454a700907664f3e88f503950cf', 'allow_patterns': ['*']}
[05-13 16:34:38|DEBUG|cosmos_predict2/_src/imaginaire/utils/env_parsers/env_parser.py:86:get_val_dict] getting val dict of CredentialEnvParser
[05-13 16:34:38|DEBUG|cosmos_predict2/_src/imaginaire/utils/env_parsers/env_parser.py:86:get_val_dict] getting val dict of CredentialEnvParser
[05-13 16:34:44|INFO|cosmos_predict2/_src/imaginaire/visualize/video.py:31:<module>] No module named 'ffmpegcv'
[05-13 16:34:44|CRITICAL|cosmos_predict2/_src/predict2/checkpointer/dcp.py:188:<module>] for the back comptiable pytorch! New DefaultLoadPlanner class is created.
[05-13 16:34:44|DEBUG|cosmos_predict2/inference.py:31:__init__] SetupArguments(output_dir=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results') model='2B/post-trained' checkpoint_path='/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/checkpoints/checkpoint/iter_3600/model_ema_bf16.pt' experiment='Stage-c_pt_4-Index-2-Size-2B-Res-720-Fps-16-Note-rf_with_edm_ckpt' config_file='cosmos_predict2/_src/predict2/configs/video2world/config.py' context_parallel_size=1 offload_diffusion_model=False offload_tokenizer=False offload_text_encoder=False disable_guardrails=True offload_guardrail_models=True keep_going=True profile=False)
[05-13 16:34:44|INFO|cosmos_predict2/_src/predict2/datasets/augmentor_provider.py:94:augmentor_register] registering video_basic_augmentor_v1...
[05-13 16:34:44|INFO|cosmos_predict2/_src/predict2/datasets/augmentor_provider.py:94:augmentor_register] registering video_basic_augmentor_v2...
[05-13 16:34:44|INFO|cosmos_predict2/_src/predict2/datasets/augmentor_provider.py:94:augmentor_register] registering noframedrop_nocameramove_video_augmentor_v1...
[05-13 16:34:44|INFO|cosmos_predict2/_src/predict2/datasets/augmentor_provider.py:94:augmentor_register] registering nocameramove_video_augmentor_v1...
[05-13 16:34:44|INFO|cosmos_predict2/_src/predict2/datasets/augmentor_provider.py:94:augmentor_register] registering image_basic_augmentor...
[05-13 16:34:44|INFO|cosmos_predict2/_src/predict2/datasets/augmentor_provider.py:94:augmentor_register] registering image_basic_augmentor_without_embeddings...
[05-13 16:34:45|DEBUG|cosmos_predict2/_src/imaginaire/attention/utils/safe_log.py:35:debug] Flash Attention v3 is not supported because the Python package was not found.
[05-13 16:34:45|DEBUG|cosmos_predict2/_src/imaginaire/attention/utils/safe_log.py:35:debug] NATTEN Attention is not supported due to insufficient NATTEN version natten.__version__='0.21.0', expected at least NATTEN_MIN_RELEASE_VERSION=[0, 21, 5], or NATTEN_MIN_DEV_VERSION=([0, 21, 5], 9).
[05-13 16:34:45|CRITICAL|cosmos_predict2/_src/imaginaire/utils/config_helper.py:192:import_all_modules_from_package] Reloading all modules from package cosmos_predict2._src.predict2.configs.video2world.experiment
[05-13 16:34:45|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2._src.predict2.configs.video2world.experiment.reason_embeddings
[05-13 16:34:45|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2._src.predict2.configs.video2world.experiment.reason_embeddings.model_14b_reason_1p1
[05-13 16:34:46|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2._src.predict2.configs.video2world.experiment.reason_embeddings.model_14b_reason_1p1_rectified_flow
[05-13 16:34:46|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2._src.predict2.configs.video2world.experiment.reason_embeddings.model_2B_reason_1p1
[05-13 16:34:46|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2._src.predict2.configs.video2world.experiment.reason_embeddings.model_2B_reason_1p1_rectified_flow
[05-13 16:34:46|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2._src.predict2.configs.video2world.experiment.reason_embeddings.model_7b_reason_1p1
[05-13 16:34:46|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2._src.predict2.configs.video2world.experiment.reason_embeddings.stage3_14B_index_3
[05-13 16:34:46|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2._src.predict2.configs.video2world.experiment.reason_embeddings.stage3_2B
[05-13 16:34:46|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2._src.predict2.configs.video2world.experiment.resume_text2world
[05-13 16:34:46|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2._src.predict2.configs.video2world.experiment.resume_text2world.sparse_14B
[05-13 16:34:47|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2._src.predict2.configs.video2world.experiment.resume_text2world.sparse_2B
[05-13 16:34:47|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2._src.predict2.configs.video2world.experiment.resume_text2world.stage3
[05-13 16:34:47|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2._src.predict2.configs.video2world.experiment.resume_text2world.stage3_2B_ablation
[05-13 16:34:48|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2._src.predict2.configs.video2world.experiment.resume_text2world.state3_14B_index_3
[05-13 16:34:48|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2._src.predict2.configs.video2world.experiment.specialized_model
[05-13 16:34:48|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2._src.predict2.configs.video2world.experiment.specialized_model.SFT_14B_RF
[05-13 16:34:48|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2._src.predict2.configs.video2world.experiment.specialized_model.SFT_2B
[05-13 16:34:48|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2._src.predict2.configs.video2world.experiment.specialized_model.SFT_2B_RF
[05-13 16:34:48|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_face
[05-13 16:34:48|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_crowded
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_high_motion
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_robotics
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_av
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_physical_ai
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_4k_cooldown
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_face_only2
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_crowded_only2
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_high_motion_only2
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_robotics_only2
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_av_only2
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_physical_ai_only2
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_4k_cooldown_only2
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_face_high_sigma
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_crowded_high_sigma
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_high_motion_high_sigma
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_robotics_high_sigma
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_av_high_sigma
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_physical_ai_high_sigma
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_4k_cooldown_high_sigma
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_face_high_sigma2
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_crowded_high_sigma2
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-1-Size-2B-Res-720-Fps-16-Note-rf_robotics_high_sigma2
[05-13 16:34:49|INFO|cosmos_predict2/_src/predict2/configs/video2world/experiment/specialized_model/SFT_2B_RF.py:853:<module>] Storing Stage-c_pt_4-Index-2-Size-2B-Res-720-Fps-16-Note-rf_with_edm_ckpt
[05-13 16:34:49|CRITICAL|cosmos_predict2/_src/imaginaire/utils/config_helper.py:192:import_all_modules_from_package] Reloading all modules from package cosmos_predict2.experiments
[05-13 16:34:49|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2.experiments.base
[05-13 16:34:49|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2.experiments.base.action
[05-13 16:34:49|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2.experiments.base.cosmos_nemo_assets
[05-13 16:34:49|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2.experiments.base.cosmos_nemo_assets_lora
[05-13 16:34:49|DEBUG|cosmos_predict2/_src/imaginaire/utils/config_helper.py:210:import_modules_recursively] Reloading module cosmos_predict2.experiments.base.groot
[05-13 16:34:51|INFO|cosmos_predict2/_src/predict2/utils/model_loader.py:91:load_model_from_checkpoint] Overriding config checkpoint path with: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/checkpoints/checkpoint/iter_3600/model_ema_bf16.pt
[05-13 16:34:51|INFO|cosmos_predict2/_src/imaginaire/utils/misc.py:152:set_random_seed] Using random seed 0.
[05-13 16:34:51|INFO|cosmos_predict2/_src/predict2/utils/model_loader.py:111:load_model_from_checkpoint] Loading model from /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/checkpoints/checkpoint/iter_3600/model_ema_bf16.pt
[05-13 16:34:51|WARNING|cosmos_predict2/_src/predict2/models/text2world_model_rectified_flow.py:132:__init__] DiffusionModel: precision torch.bfloat16
[05-13 16:34:51|INFO|cosmos_predict2/_src/imaginaire/utils/checkpoint_db.py:183:path] Using local checkpoint Wan2.1/vae(685afcaa-4de2-42fe-b7b9-69f7a2dee4d8) from /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/checkpoints/nvidia/Cosmos-Predict2.5-2B/tokenizer.pth
[05-13 16:34:51|INFO|cosmos_predict2/_src/predict2/tokenizers/wan2pt1.py:669:_video_vae] loading /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/checkpoints/nvidia/Cosmos-Predict2.5-2B/tokenizer.pth
[05-13 16:34:51|INFO|cosmos_predict2/_src/imaginaire/utils/timer.py:150:_log] Time spent on DiffusionModel: set_up_tokenizer: 0.32 s
[05-13 16:34:51|INFO|cosmos_predict2/_src/predict2/conditioner.py:434:__init__] Initialized embedder #0-fps: 
 ReMapkey 
	input key: fps
	Param count: 0 
	Trainable: None
	Dropout rate: 0.0
	Output key: fps 
	Dtype: None
[05-13 16:34:51|INFO|cosmos_predict2/_src/predict2/conditioner.py:434:__init__] Initialized embedder #1-padding_mask: 
 ReMapkey 
	input key: padding_mask
	Param count: 0 
	Trainable: None
	Dropout rate: 0.0
	Output key: padding_mask 
	Dtype: None
[05-13 16:34:51|INFO|cosmos_predict2/_src/predict2/conditioner.py:434:__init__] Initialized embedder #2-text: 
 TextAttr 
	input key: ['t5_text_embeddings']
	Param count: 0 
	Trainable: None
	Dropout rate: 0.2
	Output key: [crossattn_emb]
[05-13 16:34:51|INFO|cosmos_predict2/_src/predict2/conditioner.py:434:__init__] Initialized embedder #3-use_video_condition: 
 BooleanFlag 
	input key: fps
	Param count: 0 
	Trainable: None
	Dropout rate: 0.0
	Output key: use_video_condition 
	 This is a boolean flag
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1624:build_pos_embed] Building positional embedding with rope3d class, impl <class 'cosmos_predict2._src.predict2.networks.minimal_v4_dit.VideoRopePosition3DEmb'>
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:886:__init__] Using AdaLN LoRA Flag:  True. We enable bias if no AdaLN LoRA for backward compatibility.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is None and using 16 heads with a dimension of 128.
[05-13 16:34:51|DEBUG|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:432:__init__] Setting up Attention. Query dim is 2048, context_dim is 1024 and using 16 heads with a dimension of 128.
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1804:enable_selective_checkpoint] Enable selective checkpoint with predict2_2b_720_aggressive, for every 1 blocks. Total blocks: 28
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 0
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 1
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 2
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 3
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 4
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 5
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 6
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 7
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 8
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 9
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 10
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 11
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 12
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 13
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 14
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 15
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 16
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 17
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 18
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 19
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 20
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 21
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 22
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 23
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 24
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 25
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 26
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/networks/minimal_v4_dit.py:1810:enable_selective_checkpoint] Enable selective checkpoint for block 27
[05-13 16:34:52|INFO|cosmos_predict2/_src/imaginaire/utils/timer.py:150:_log] Time spent on meta to cuda and broadcast model states: 0.15 s
[05-13 16:34:52|INFO|cosmos_predict2/_src/imaginaire/utils/timer.py:150:_log] Time spent on Creating PyTorch model: 0.65 s
[05-13 16:34:52|INFO|cosmos_predict2/_src/imaginaire/utils/timer.py:150:_log] Time spent on Creating PyTorch model and ema if enabled: 0.66 s
[05-13 16:34:52|INFO|cosmos_predict2/_src/imaginaire/utils/timer.py:150:_log] Time spent on DiffusionModel: set_up_model: 0.66 s
[05-13 16:34:52|INFO|cosmos_predict2/_src/predict2/text_encoders/text_encoder.py:74:__init__] Instantiating text encoder model...
[05-13 16:34:52|INFO|cosmos_predict2/_src/imaginaire/utils/checkpoint_db.py:183:path] Using local checkpoint Qwen/Qwen2.5-VL-7B-Instruct(7219c6c7-f878-4137-bbdb-76842ea85e70) from /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/checkpoints/nvidia/Cosmos-Reason1-7B
[05-13 16:34:52|INFO|cosmos_predict2/_src/reason1/tokenizer/processor.py:75:__init__] Successfully loaded processor from local cache
[05-13 16:34:52|INFO|cosmos_predict2/_src/reason1/models/vlm_base.py:116:__init__] Setting torch default dtype from torch.float32 to torch.bfloat16
[05-13 16:34:52|INFO|cosmos_predict2/_src/reason1/models/vlm_base.py:123:__init__] Reset torch default dtype to torch.float32
[05-13 16:34:53|INFO|cosmos_predict2/_src/predict2/text_encoders/text_encoder.py:84:__init__] Loading checkpoint from s3://bucket/cosmos_reasoning1/sft_exp700/sft_exp721-1_qwen7b_tl_721_5vs5_s3_balanced_n32_resume_16k/checkpoints/iter_000016000/model/.
[05-13 16:34:53|INFO|cosmos_predict2/_src/imaginaire/utils/checkpoint_db.py:183:path] Using local checkpoint nvidia/Cosmos-Reason1.1-7B(cb3e3ffa-7b08-4c34-822d-61c7aa31a14f) from /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/checkpoints/nvidia/Cosmos-Reason1-7B
[05-13 16:35:01|INFO|cosmos_predict2/_src/predict2/text_encoders/text_encoder.py:113:__init__] Finished loading checkpoint from /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/checkpoints/nvidia/Cosmos-Reason1-7B.
[05-13 16:35:01|INFO|cosmos_predict2/_src/predict2/text_encoders/text_encoder.py:116:__init__] Text encoder model instantiated
[05-13 16:35:01|INFO|cosmos_predict2/_src/imaginaire/utils/timer.py:150:_log] Time spent on instantiate model: 10.51 s
[05-13 16:35:01|INFO|cosmos_predict2/_src/predict2/utils/model_loader.py:227:load_model_state_dict_from_checkpoint] Loading model cached locally from /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/checkpoints/checkpoint/iter_3600/model_ema_bf16.pt
[05-13 16:35:03|CRITICAL|cosmos_predict2/_src/predict2/models/text2world_model_rectified_flow.py:919:load_state_dict] load model in non-strict mode
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.0.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.0.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.0.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.0.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.1.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.1.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.1.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.1.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.2.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.2.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.2.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.2.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.3.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.3.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.3.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.3.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.4.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.4.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.4.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.4.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.5.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.5.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.5.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.5.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.6.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.6.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.6.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.6.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.7.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.7.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.7.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.7.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.8.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.8.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.8.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.8.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.9.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.9.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.9.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.9.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.10.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.10.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.10.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.10.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.11.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.11.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.11.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.11.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.12.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.12.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.12.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.12.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.13.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.13.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.13.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.13.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.14.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.14.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.14.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.14.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.15.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.15.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.15.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.15.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.16.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.16.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.16.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.16.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.17.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.17.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.17.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.17.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.18.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.18.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.18.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.18.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.19.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.19.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.19.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.19.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.20.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.20.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.20.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.20.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.21.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.21.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.21.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.21.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.22.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.22.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.22.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.22.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.23.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.23.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.23.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.23.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.24.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.24.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.24.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.24.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.25.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.25.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.25.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.25.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.26.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.26.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.26.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.26.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.27.self_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.27.self_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.27.cross_attn.q_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key blocks.27.cross_attn.k_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:03|WARNING|cosmos_predict2/_src/imaginaire/utils/checkpointer.py:451:non_strict_load_model] Skipping key t_embedding_norm._extra_state introduced by TransformerEngine for FP8 in the checkpoint.
[05-13 16:35:04|CRITICAL|cosmos_predict2/_src/predict2/models/text2world_model_rectified_flow.py:920:load_state_dict] [RANK 0] _IncompatibleKeys(missing_keys=[], unexpected_keys=[], incorrect_shapes=[])
[05-13 16:35:04|INFO|cosmos_predict2/inference.py:59:__init__] Saved config to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/config.yaml
[05-13 16:35:04|INFO|cosmos_predict2/inference.py:81:generate] Generating 125 samples: ['gr1_behavior_0', 'gr1_behavior_1', 'gr1_behavior_10', 'gr1_behavior_11', 'gr1_behavior_12', 'gr1_behavior_13', 'gr1_behavior_14', 'gr1_behavior_15', 'gr1_behavior_16', 'gr1_behavior_17', 'gr1_behavior_18', 'gr1_behavior_19', 'gr1_behavior_2', 'gr1_behavior_20', 'gr1_behavior_21', 'gr1_behavior_22', 'gr1_behavior_23', 'gr1_behavior_24', 'gr1_behavior_25', 'gr1_behavior_26', 'gr1_behavior_27', 'gr1_behavior_28', 'gr1_behavior_29', 'gr1_behavior_3', 'gr1_behavior_30', 'gr1_behavior_31', 'gr1_behavior_32', 'gr1_behavior_33', 'gr1_behavior_34', 'gr1_behavior_35', 'gr1_behavior_36', 'gr1_behavior_37', 'gr1_behavior_38', 'gr1_behavior_39', 'gr1_behavior_4', 'gr1_behavior_40', 'gr1_behavior_41', 'gr1_behavior_42', 'gr1_behavior_43', 'gr1_behavior_44', 'gr1_behavior_45', 'gr1_behavior_46', 'gr1_behavior_5', 'gr1_behavior_6', 'gr1_behavior_7', 'gr1_behavior_8', 'gr1_behavior_9', 'gr1_env_0', 'gr1_env_1', 'gr1_env_10', 'gr1_env_11', 'gr1_env_12', 'gr1_env_13', 'gr1_env_14', 'gr1_env_15', 'gr1_env_16', 'gr1_env_17', 'gr1_env_18', 'gr1_env_19', 'gr1_env_2', 'gr1_env_20', 'gr1_env_21', 'gr1_env_22', 'gr1_env_23', 'gr1_env_24', 'gr1_env_25', 'gr1_env_26', 'gr1_env_27', 'gr1_env_28', 'gr1_env_3', 'gr1_env_4', 'gr1_env_5', 'gr1_env_6', 'gr1_env_7', 'gr1_env_8', 'gr1_env_9', 'gr1_object_0', 'gr1_object_1', 'gr1_object_10', 'gr1_object_11', 'gr1_object_13', 'gr1_object_14', 'gr1_object_15', 'gr1_object_16', 'gr1_object_17', 'gr1_object_18', 'gr1_object_19', 'gr1_object_2', 'gr1_object_20', 'gr1_object_21', 'gr1_object_22', 'gr1_object_23', 'gr1_object_24', 'gr1_object_25', 'gr1_object_26', 'gr1_object_27', 'gr1_object_28', 'gr1_object_29', 'gr1_object_3', 'gr1_object_30', 'gr1_object_31', 'gr1_object_32', 'gr1_object_33', 'gr1_object_34', 'gr1_object_35', 'gr1_object_36', 'gr1_object_37', 'gr1_object_38', 'gr1_object_39', 'gr1_object_4', 'gr1_object_40', 'gr1_object_41', 'gr1_object_42', 'gr1_object_43', 'gr1_object_44', 'gr1_object_45', 'gr1_object_46', 'gr1_object_47', 'gr1_object_48', 'gr1_object_49', 'gr1_object_5', 'gr1_object_6', 'gr1_object_7', 'gr1_object_8', 'gr1_object_9']
[05-13 16:35:04|INFO|cosmos_predict2/inference.py:85:generate] [1/125] Processing sample gr1_behavior_0
[05-13 16:35:04|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_0' prompt_path=None prompt='Open the box' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/0_Open the box.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 16:35:04|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_0.json
[05-13 16:35:04|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 16:35:04|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 16:35:04|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/0_Open the box.png
[05-13 16:35:05|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.80 GB
[05-13 16:35:05|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 16:50:31|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 16:50:32|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 16:50:34|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_0.mp4
[05-13 16:50:34|INFO|cosmos_predict2/inference.py:85:generate] [2/125] Processing sample gr1_behavior_1
[05-13 16:50:34|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_1' prompt_path=None prompt='Use both hands to pick up pot' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/1_Use both hands to pick up pot.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 16:50:34|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_1.json
[05-13 16:50:34|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 16:50:34|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 16:50:34|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/1_Use both hands to pick up pot.png
[05-13 16:50:34|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 16:50:34|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 17:05:36|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 17:05:37|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 17:05:40|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_1.mp4
[05-13 17:05:40|INFO|cosmos_predict2/inference.py:85:generate] [3/125] Processing sample gr1_behavior_10
[05-13 17:05:40|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_10' prompt_path=None prompt='Use right hand to strum ukelele' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/10_Use right hand to strum ukelele.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 17:05:40|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_10.json
[05-13 17:05:40|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 17:05:40|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 17:05:40|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/10_Use right hand to strum ukelele.png
[05-13 17:05:40|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 17:05:40|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 17:20:43|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 17:20:44|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 17:20:46|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_10.mp4
[05-13 17:20:46|INFO|cosmos_predict2/inference.py:85:generate] [4/125] Processing sample gr1_behavior_11
[05-13 17:20:46|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_11' prompt_path=None prompt='Use the left hand to hold mouse and move the mouse around' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/11_Use the left hand to hold mouse and move the mouse around.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 17:20:46|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_11.json
[05-13 17:20:46|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 17:20:46|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 17:20:46|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/11_Use the left hand to hold mouse and move the mouse around.png
[05-13 17:20:47|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 17:20:47|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 17:35:50|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 17:35:51|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 17:35:53|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_11.mp4
[05-13 17:35:53|INFO|cosmos_predict2/inference.py:85:generate] [5/125] Processing sample gr1_behavior_12
[05-13 17:35:53|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_12' prompt_path=None prompt='Use the left hand to pick up shaker and shake it' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/12_Use the left hand to pick up shaker and shake it.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 17:35:53|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_12.json
[05-13 17:35:53|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 17:35:53|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 17:35:53|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/12_Use the left hand to pick up shaker and shake it.png
[05-13 17:35:53|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 17:35:53|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 17:50:54|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 17:50:55|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 17:50:57|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_12.mp4
[05-13 17:50:57|INFO|cosmos_predict2/inference.py:85:generate] [6/125] Processing sample gr1_behavior_13
[05-13 17:50:57|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_13' prompt_path=None prompt='Use the left hand to pick up the dustpan, use the right hand to pick up the tape dispenser, then sweep the dust on the table into the dustpan' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/13_Use the left hand to pick up the dustpan, use the right hand to pick up the tape dispenser, then sweep the dust on the table into the dustpan.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 17:50:57|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_13.json
[05-13 17:50:57|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 17:50:57|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 17:50:57|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/13_Use the left hand to pick up the dustpan, use the right hand to pick up the tape dispenser, then sweep the dust on the table into the dustpan.png
[05-13 17:50:58|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 17:50:58|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 18:06:00|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 18:06:01|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 18:06:03|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_13.mp4
[05-13 18:06:03|INFO|cosmos_predict2/inference.py:85:generate] [7/125] Processing sample gr1_behavior_14
[05-13 18:06:03|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_14' prompt_path=None prompt='Use the right hand to close lunch box' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/14_Use the right hand to close lunch box.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 18:06:03|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_14.json
[05-13 18:06:03|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 18:06:03|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 18:06:03|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/14_Use the right hand to close lunch box.png
[05-13 18:06:03|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 18:06:03|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 18:21:06|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 18:21:07|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 18:21:09|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_14.mp4
[05-13 18:21:09|INFO|cosmos_predict2/inference.py:85:generate] [8/125] Processing sample gr1_behavior_15
[05-13 18:21:09|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_15' prompt_path=None prompt='Use the right hand to close microwave' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/15_Use the right hand to close microwave.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 18:21:09|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_15.json
[05-13 18:21:09|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 18:21:09|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 18:21:09|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/15_Use the right hand to close microwave.png
[05-13 18:21:10|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 18:21:10|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 18:36:13|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 18:36:14|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 18:36:16|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_15.mp4
[05-13 18:36:16|INFO|cosmos_predict2/inference.py:85:generate] [9/125] Processing sample gr1_behavior_16
[05-13 18:36:16|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_16' prompt_path=None prompt='Use the right hand to grab the pan handle and toss the pan' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/16_Use the right hand to grab the pan handle and toss the pan.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 18:36:16|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_16.json
[05-13 18:36:16|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 18:36:16|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 18:36:16|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/16_Use the right hand to grab the pan handle and toss the pan.png
[05-13 18:36:16|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 18:36:16|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 18:51:19|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 18:51:21|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 18:51:22|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_16.mp4
[05-13 18:51:22|INFO|cosmos_predict2/inference.py:85:generate] [10/125] Processing sample gr1_behavior_17
[05-13 18:51:22|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_17' prompt_path=None prompt='Use the right hand to open macbook' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/17_Use the right hand to open macbook.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 18:51:22|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_17.json
[05-13 18:51:22|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 18:51:22|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 18:51:22|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/17_Use the right hand to open macbook.png
[05-13 18:51:23|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 18:51:23|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 19:06:25|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 19:06:27|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 19:06:28|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_17.mp4
[05-13 19:06:28|INFO|cosmos_predict2/inference.py:85:generate] [11/125] Processing sample gr1_behavior_18
[05-13 19:06:28|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_18' prompt_path=None prompt='Use the right hand to pick up blue scoop and scoop powder from container' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/18_Use the right hand to pick up blue scoop and scoop powder from container.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 19:06:28|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_18.json
[05-13 19:06:28|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 19:06:28|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 19:06:28|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/18_Use the right hand to pick up blue scoop and scoop powder from container.png
[05-13 19:06:28|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 19:06:28|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 19:21:31|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 19:21:32|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 19:21:34|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_18.mp4
[05-13 19:21:34|INFO|cosmos_predict2/inference.py:85:generate] [12/125] Processing sample gr1_behavior_19
[05-13 19:21:34|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_19' prompt_path=None prompt='Use the right hand to pick up expo eraser and erase whiteboard' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/19_Use the right hand to pick up expo eraser and erase whiteboard.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 19:21:34|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_19.json
[05-13 19:21:34|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 19:21:34|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 19:21:34|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/19_Use the right hand to pick up expo eraser and erase whiteboard.png
[05-13 19:21:34|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 19:21:34|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 19:36:34|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 19:36:36|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 19:36:37|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_19.mp4
[05-13 19:36:37|INFO|cosmos_predict2/inference.py:85:generate] [13/125] Processing sample gr1_behavior_2
[05-13 19:36:37|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_2' prompt_path=None prompt='Use knife to cut the object on the cutting board' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/2_Use knife to cut the object on the cutting board.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 19:36:37|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_2.json
[05-13 19:36:37|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 19:36:37|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 19:36:37|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/2_Use knife to cut the object on the cutting board.png
[05-13 19:36:38|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 19:36:38|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 19:51:41|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 19:51:42|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 19:51:43|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_2.mp4
[05-13 19:51:43|INFO|cosmos_predict2/inference.py:85:generate] [14/125] Processing sample gr1_behavior_20
[05-13 19:51:43|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_20' prompt_path=None prompt='Use the right hand to pick up glass and bring it close to the camera as if drinking' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/20_Use the right hand to pick up glass and bring it close to the camera as if drinking.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 19:51:43|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_20.json
[05-13 19:51:43|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 19:51:43|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 19:51:43|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/20_Use the right hand to pick up glass and bring it close to the camera as if drinking.png
[05-13 19:51:44|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 19:51:44|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 20:06:45|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 20:06:46|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 20:06:48|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_20.mp4
[05-13 20:06:48|INFO|cosmos_predict2/inference.py:85:generate] [15/125] Processing sample gr1_behavior_21
[05-13 20:06:48|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_21' prompt_path=None prompt='Use the right hand to pick up hat and put it on top of mini tripod' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/21_Use the right hand to pick up hat and put it on top of mini tripod.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 20:06:48|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_21.json
[05-13 20:06:48|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 20:06:48|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 20:06:48|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/21_Use the right hand to pick up hat and put it on top of mini tripod.png
[05-13 20:06:48|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 20:06:48|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 20:21:50|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 20:21:51|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 20:21:53|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_21.mp4
[05-13 20:21:53|INFO|cosmos_predict2/inference.py:85:generate] [16/125] Processing sample gr1_behavior_22
[05-13 20:21:53|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_22' prompt_path=None prompt='Use the right hand to pick up iron and press the T-shirt' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/22_Use the right hand to pick up iron and press the T-shirt.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 20:21:53|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_22.json
[05-13 20:21:53|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 20:21:53|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 20:21:53|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/22_Use the right hand to pick up iron and press the T-shirt.png
[05-13 20:21:53|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 20:21:53|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 20:36:56|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 20:36:58|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 20:36:59|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_22.mp4
[05-13 20:36:59|INFO|cosmos_predict2/inference.py:85:generate] [17/125] Processing sample gr1_behavior_23
[05-13 20:36:59|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_23' prompt_path=None prompt='Use the right hand to pick up long-reach lighter to light the candle' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/23_Use the right hand to pick up long-reach lighter to light the candle.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 20:36:59|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_23.json
[05-13 20:36:59|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 20:36:59|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 20:36:59|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/23_Use the right hand to pick up long-reach lighter to light the candle.png
[05-13 20:37:00|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 20:37:00|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 20:52:02|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 20:52:03|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 20:52:04|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_23.mp4
[05-13 20:52:04|INFO|cosmos_predict2/inference.py:85:generate] [18/125] Processing sample gr1_behavior_24
[05-13 20:52:04|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_24' prompt_path=None prompt='Use the right hand to pick up marker and write on whiteboard' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/24_Use the right hand to pick up marker and write on whiteboard.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 20:52:04|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_24.json
[05-13 20:52:04|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 20:52:04|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 20:52:04|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/24_Use the right hand to pick up marker and write on whiteboard.png
[05-13 20:52:05|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 20:52:05|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 21:07:08|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 21:07:09|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 21:07:10|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_24.mp4
[05-13 21:07:10|INFO|cosmos_predict2/inference.py:85:generate] [19/125] Processing sample gr1_behavior_25
[05-13 21:07:10|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_25' prompt_path=None prompt='Use the right hand to pick up pink bottle and pour water on flower' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/25_Use the right hand to pick up pink bottle and pour water on flower.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 21:07:10|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_25.json
[05-13 21:07:10|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 21:07:10|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 21:07:10|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/25_Use the right hand to pick up pink bottle and pour water on flower.png
[05-13 21:07:11|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 21:07:11|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 21:22:10|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 21:22:11|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 21:22:13|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_25.mp4
[05-13 21:22:13|INFO|cosmos_predict2/inference.py:85:generate] [20/125] Processing sample gr1_behavior_26
[05-13 21:22:13|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_26' prompt_path=None prompt='Use the right hand to pick up plate and place it onto the dish rack' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/26_Use the right hand to pick up plate and place it onto the dish rack.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 21:22:13|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_26.json
[05-13 21:22:13|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 21:22:13|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 21:22:13|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/26_Use the right hand to pick up plate and place it onto the dish rack.png
[05-13 21:22:14|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 21:22:14|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 21:37:15|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 21:37:16|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 21:37:19|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_26.mp4
[05-13 21:37:19|INFO|cosmos_predict2/inference.py:85:generate] [21/125] Processing sample gr1_behavior_27
[05-13 21:37:19|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_27' prompt_path=None prompt='Use the right hand to pick up rag and erase whiteboard' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/27_Use the right hand to pick up rag and erase whiteboard.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 21:37:19|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_27.json
[05-13 21:37:19|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 21:37:19|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 21:37:19|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/27_Use the right hand to pick up rag and erase whiteboard.png
[05-13 21:37:19|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 21:37:19|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 21:52:21|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 21:52:23|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 21:52:24|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_27.mp4
[05-13 21:52:24|INFO|cosmos_predict2/inference.py:85:generate] [22/125] Processing sample gr1_behavior_28
[05-13 21:52:24|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_28' prompt_path=None prompt='Use the right hand to pick up sauce bottle and shake it' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/28_Use the right hand to pick up sauce bottle and shake it.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 21:52:24|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_28.json
[05-13 21:52:24|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 21:52:24|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 21:52:24|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/28_Use the right hand to pick up sauce bottle and shake it.png
[05-13 21:52:24|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 21:52:24|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 22:07:27|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 22:07:28|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 22:07:30|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_28.mp4
[05-13 22:07:30|INFO|cosmos_predict2/inference.py:85:generate] [23/125] Processing sample gr1_behavior_29
[05-13 22:07:30|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_29' prompt_path=None prompt='Use the right hand to pick up scraper and scrape cutting board' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/29_Use the right hand to pick up scraper and scrape cutting board.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 22:07:30|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_29.json
[05-13 22:07:30|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 22:07:30|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 22:07:30|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/29_Use the right hand to pick up scraper and scrape cutting board.png
[05-13 22:07:31|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 22:07:31|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 22:22:32|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 22:22:34|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 22:22:36|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_29.mp4
[05-13 22:22:36|INFO|cosmos_predict2/inference.py:85:generate] [24/125] Processing sample gr1_behavior_3
[05-13 22:22:36|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_3' prompt_path=None prompt='Use left hand to pick up glass on the table and clink against the glass held in the right hand' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/3_Use left hand to pick up glass on the table and clink against the glass held in the right hand.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 22:22:36|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_3.json
[05-13 22:22:36|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 22:22:36|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 22:22:36|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/3_Use left hand to pick up glass on the table and clink against the glass held in the right hand.png
[05-13 22:22:36|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 22:22:36|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 22:37:38|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 22:37:39|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 22:37:41|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_3.mp4
[05-13 22:37:41|INFO|cosmos_predict2/inference.py:85:generate] [25/125] Processing sample gr1_behavior_30
[05-13 22:37:41|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_30' prompt_path=None prompt='Use the right hand to pick up shaker and crack it on side of bowl' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/30_Use the right hand to pick up shaker and crack it on side of bowl.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 22:37:41|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_30.json
[05-13 22:37:41|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 22:37:41|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 22:37:41|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/30_Use the right hand to pick up shaker and crack it on side of bowl.png
[05-13 22:37:41|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 22:37:41|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 22:52:43|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 22:52:45|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 22:52:47|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_30.mp4
[05-13 22:52:47|INFO|cosmos_predict2/inference.py:85:generate] [26/125] Processing sample gr1_behavior_31
[05-13 22:52:47|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_31' prompt_path=None prompt='Use the right hand to pick up spatula and spread butter on bread' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/31_Use the right hand to pick up spatula and spread butter on bread.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 22:52:47|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_31.json
[05-13 22:52:47|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 22:52:47|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 22:52:47|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/31_Use the right hand to pick up spatula and spread butter on bread.png
[05-13 22:52:47|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 22:52:47|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 23:07:49|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 23:07:51|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 23:07:53|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_31.mp4
[05-13 23:07:53|INFO|cosmos_predict2/inference.py:85:generate] [27/125] Processing sample gr1_behavior_32
[05-13 23:07:53|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_32' prompt_path=None prompt='Use the right hand to pick up tambourine and hit it with left hand' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/32_Use the right hand to pick up tambourine and hit it with left hand.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 23:07:53|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_32.json
[05-13 23:07:53|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 23:07:53|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 23:07:53|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/32_Use the right hand to pick up tambourine and hit it with left hand.png
[05-13 23:07:53|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 23:07:53|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 23:22:56|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 23:22:57|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 23:22:58|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_32.mp4
[05-13 23:22:58|INFO|cosmos_predict2/inference.py:85:generate] [28/125] Processing sample gr1_behavior_33
[05-13 23:22:58|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_33' prompt_path=None prompt='Use the right hand to pick up the bottle and pour water into the glass' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/33_Use the right hand to pick up the bottle and pour water into the glass.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 23:22:58|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_33.json
[05-13 23:22:58|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 23:22:58|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 23:22:58|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/33_Use the right hand to pick up the bottle and pour water into the glass.png
[05-13 23:22:59|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 23:22:59|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 23:38:03|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 23:38:04|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 23:38:05|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_33.mp4
[05-13 23:38:05|INFO|cosmos_predict2/inference.py:85:generate] [29/125] Processing sample gr1_behavior_34
[05-13 23:38:05|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_34' prompt_path=None prompt='Use the right hand to pick up tongs and squeeze it' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/34_Use the right hand to pick up tongs and squeeze it.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 23:38:05|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_34.json
[05-13 23:38:05|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 23:38:05|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 23:38:05|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/34_Use the right hand to pick up tongs and squeeze it.png
[05-13 23:38:06|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 23:38:06|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-13 23:53:05|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-13 23:53:06|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-13 23:53:08|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_34.mp4
[05-13 23:53:08|INFO|cosmos_predict2/inference.py:85:generate] [30/125] Processing sample gr1_behavior_35
[05-13 23:53:08|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_35' prompt_path=None prompt='Use the right hand to pick up toy and wave it' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/35_Use the right hand to pick up toy and wave it.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-13 23:53:08|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_35.json
[05-13 23:53:08|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-13 23:53:08|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-13 23:53:08|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/35_Use the right hand to pick up toy and wave it.png
[05-13 23:53:08|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-13 23:53:08|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 00:08:10|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 00:08:11|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 00:08:13|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_35.mp4
[05-14 00:08:13|INFO|cosmos_predict2/inference.py:85:generate] [31/125] Processing sample gr1_behavior_36
[05-14 00:08:13|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_36' prompt_path=None prompt='Use the right hand to pick up vacuum and sweep vacuum from left to right across the table' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/36_Use the right hand to pick up vacuum and sweep vacuum from left to right across the table.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 00:08:13|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_36.json
[05-14 00:08:13|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 00:08:13|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 00:08:13|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/36_Use the right hand to pick up vacuum and sweep vacuum from left to right across the table.png
[05-14 00:08:13|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 00:08:13|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 00:23:16|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 00:23:17|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 00:23:19|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_36.mp4
[05-14 00:23:19|INFO|cosmos_predict2/inference.py:85:generate] [32/125] Processing sample gr1_behavior_37
[05-14 00:23:19|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_37' prompt_path=None prompt='Use the right hand to pick up wooden spoon to the table' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/37_Use the right hand to pick up wooden spoon to the table.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 00:23:19|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_37.json
[05-14 00:23:19|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 00:23:19|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 00:23:19|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/37_Use the right hand to pick up wooden spoon to the table.png
[05-14 00:23:19|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 00:23:19|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 00:38:21|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 00:38:22|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 00:38:24|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_37.mp4
[05-14 00:38:24|INFO|cosmos_predict2/inference.py:85:generate] [33/125] Processing sample gr1_behavior_38
[05-14 00:38:24|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_38' prompt_path=None prompt='Use the right hand to press calculator' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/38_Use the right hand to press calculator.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 00:38:24|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_38.json
[05-14 00:38:24|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 00:38:24|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 00:38:24|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/38_Use the right hand to press calculator.png
[05-14 00:38:24|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 00:38:24|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 00:53:26|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 00:53:28|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 00:53:29|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_38.mp4
[05-14 00:53:29|INFO|cosmos_predict2/inference.py:85:generate] [34/125] Processing sample gr1_behavior_39
[05-14 00:53:29|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_39' prompt_path=None prompt='Use the right hand to press down the toaster lever' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/39_Use the right hand to press down the toaster lever.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 00:53:29|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_39.json
[05-14 00:53:29|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 00:53:29|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 00:53:29|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/39_Use the right hand to press down the toaster lever.png
[05-14 00:53:29|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 00:53:29|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 01:08:31|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 01:08:32|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 01:08:34|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_39.mp4
[05-14 01:08:34|INFO|cosmos_predict2/inference.py:85:generate] [35/125] Processing sample gr1_behavior_4
[05-14 01:08:34|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_4' prompt_path=None prompt='Use right hand to pick up ball and throw it away' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/4_Use right hand to pick up ball and throw it away.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 01:08:34|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_4.json
[05-14 01:08:34|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 01:08:34|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 01:08:34|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/4_Use right hand to pick up ball and throw it away.png
[05-14 01:08:34|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 01:08:34|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 01:23:36|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 01:23:37|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 01:23:39|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_4.mp4
[05-14 01:23:39|INFO|cosmos_predict2/inference.py:85:generate] [36/125] Processing sample gr1_behavior_40
[05-14 01:23:39|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_40' prompt_path=None prompt='Use the right hand to press keyboard' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/40_Use the right hand to press keyboard.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 01:23:39|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_40.json
[05-14 01:23:39|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 01:23:39|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 01:23:39|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/40_Use the right hand to press keyboard.png
[05-14 01:23:39|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 01:23:39|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 01:38:43|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 01:38:44|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 01:38:45|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_40.mp4
[05-14 01:38:45|INFO|cosmos_predict2/inference.py:85:generate] [37/125] Processing sample gr1_behavior_41
[05-14 01:38:45|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_41' prompt_path=None prompt='Use the right hand to press red button' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/41_Use the right hand to press red button.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 01:38:45|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_41.json
[05-14 01:38:45|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 01:38:45|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 01:38:45|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/41_Use the right hand to press red button.png
[05-14 01:38:46|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 01:38:46|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 01:53:49|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 01:53:50|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 01:53:51|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_41.mp4
[05-14 01:53:51|INFO|cosmos_predict2/inference.py:85:generate] [38/125] Processing sample gr1_behavior_42
[05-14 01:53:51|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_42' prompt_path=None prompt='Use the right hand to pull out a wipe' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/42_Use the right hand to pull out a wipe.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 01:53:51|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_42.json
[05-14 01:53:51|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 01:53:51|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 01:53:51|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/42_Use the right hand to pull out a wipe.png
[05-14 01:53:52|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 01:53:52|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 02:08:55|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 02:08:56|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 02:08:58|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_42.mp4
[05-14 02:08:58|INFO|cosmos_predict2/inference.py:85:generate] [39/125] Processing sample gr1_behavior_43
[05-14 02:08:58|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_43' prompt_path=None prompt='Use the right hand to swing ping pong paddle' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/43_Use the right hand to swing ping pong paddle.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 02:08:58|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_43.json
[05-14 02:08:58|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 02:08:58|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 02:08:58|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/43_Use the right hand to swing ping pong paddle.png
[05-14 02:08:58|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 02:08:58|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 02:24:02|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 02:24:03|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 02:24:05|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_43.mp4
[05-14 02:24:05|INFO|cosmos_predict2/inference.py:85:generate] [40/125] Processing sample gr1_behavior_44
[05-14 02:24:05|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_44' prompt_path=None prompt='Use the right hand to tap on fingerboard to make it jump' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/44_Use the right hand to tap on fingerboard to make it jump.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 02:24:05|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_44.json
[05-14 02:24:05|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 02:24:05|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 02:24:05|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/44_Use the right hand to tap on fingerboard to make it jump.png
[05-14 02:24:05|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 02:24:05|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 02:39:07|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 02:39:08|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 02:39:10|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_44.mp4
[05-14 02:39:10|INFO|cosmos_predict2/inference.py:85:generate] [41/125] Processing sample gr1_behavior_45
[05-14 02:39:10|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_45' prompt_path=None prompt='Use the right hand to turn stove knob' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/45_Use the right hand to turn stove knob.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 02:39:10|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_45.json
[05-14 02:39:10|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 02:39:10|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 02:39:10|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/45_Use the right hand to turn stove knob.png
[05-14 02:39:11|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 02:39:11|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 02:54:14|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 02:54:16|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 02:54:18|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_45.mp4
[05-14 02:54:18|INFO|cosmos_predict2/inference.py:85:generate] [42/125] Processing sample gr1_behavior_46
[05-14 02:54:18|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_46' prompt_path=None prompt='Use the right hand to unroll black mat' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/46_Use the right hand to unroll black mat.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 02:54:18|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_46.json
[05-14 02:54:18|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 02:54:18|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 02:54:18|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/46_Use the right hand to unroll black mat.png
[05-14 02:54:19|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 02:54:19|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 03:09:25|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 03:09:26|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 03:09:28|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_46.mp4
[05-14 03:09:28|INFO|cosmos_predict2/inference.py:85:generate] [43/125] Processing sample gr1_behavior_5
[05-14 03:09:28|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_5' prompt_path=None prompt='Use right hand to pick up bottle and spray water onto plate' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/5_Use right hand to pick up bottle and spray water onto plate.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 03:09:28|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_5.json
[05-14 03:09:28|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 03:09:28|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 03:09:28|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/5_Use right hand to pick up bottle and spray water onto plate.png
[05-14 03:09:28|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 03:09:28|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 03:24:35|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 03:24:37|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 03:24:39|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_5.mp4
[05-14 03:24:39|INFO|cosmos_predict2/inference.py:85:generate] [44/125] Processing sample gr1_behavior_6
[05-14 03:24:39|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_6' prompt_path=None prompt='Use right hand to pick up hammer and strike the nail into wood' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/6_Use right hand to pick up hammer and strike the nail into wood.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 03:24:39|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_6.json
[05-14 03:24:39|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 03:24:39|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 03:24:39|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/6_Use right hand to pick up hammer and strike the nail into wood.png
[05-14 03:24:39|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 03:24:39|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 03:39:48|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 03:39:49|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 03:39:51|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_6.mp4
[05-14 03:39:51|INFO|cosmos_predict2/inference.py:85:generate] [45/125] Processing sample gr1_behavior_7
[05-14 03:39:51|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_7' prompt_path=None prompt='Use right hand to pick up sponge and wipe the plate held in the right hand' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/7_Use right hand to pick up sponge and wipe the plate held in the right hand.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 03:39:51|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_7.json
[05-14 03:39:51|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 03:39:51|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 03:39:51|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/7_Use right hand to pick up sponge and wipe the plate held in the right hand.png
[05-14 03:39:52|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 03:39:52|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 03:55:00|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 03:55:01|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 03:55:03|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_7.mp4
[05-14 03:55:03|INFO|cosmos_predict2/inference.py:85:generate] [46/125] Processing sample gr1_behavior_8
[05-14 03:55:03|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_8' prompt_path=None prompt='Use right hand to pick up toy and wave it' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/8_Use right hand to pick up toy and wave it.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 03:55:03|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_8.json
[05-14 03:55:03|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 03:55:03|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 03:55:03|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/8_Use right hand to pick up toy and wave it.png
[05-14 03:55:03|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 03:55:03|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 04:10:12|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 04:10:13|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 04:10:14|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_8.mp4
[05-14 04:10:14|INFO|cosmos_predict2/inference.py:85:generate] [47/125] Processing sample gr1_behavior_9
[05-14 04:10:14|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_behavior_9' prompt_path=None prompt='Use right hand to stir with spoon' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/9_Use right hand to stir with spoon.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 04:10:14|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_9.json
[05-14 04:10:14|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 04:10:14|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 04:10:14|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_behavior/9_Use right hand to stir with spoon.png
[05-14 04:10:15|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 04:10:15|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 04:25:23|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 04:25:24|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 04:25:26|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_behavior_9.mp4
[05-14 04:25:26|INFO|cosmos_predict2/inference.py:85:generate] [48/125] Processing sample gr1_env_0
[05-14 04:25:26|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_0' prompt_path=None prompt='Use the right hand to pick up the spatula and perform a serving motion from the bowl onto the metal plate' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/0_Use the right hand to pick up the spatula and perform a serving motion from the bowl onto the metal plate.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 04:25:26|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_0.json
[05-14 04:25:26|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 04:25:26|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 04:25:26|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/0_Use the right hand to pick up the spatula and perform a serving motion from the bowl onto the metal plate.png
[05-14 04:25:27|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 04:25:27|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 04:40:35|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 04:40:36|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 04:40:37|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_0.mp4
[05-14 04:40:37|INFO|cosmos_predict2/inference.py:85:generate] [49/125] Processing sample gr1_env_1
[05-14 04:40:37|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_1' prompt_path=None prompt='Use the right hand to close lunch box' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/1_Use the right hand to close lunch box.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 04:40:37|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_1.json
[05-14 04:40:37|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 04:40:37|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 04:40:37|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/1_Use the right hand to close lunch box.png
[05-14 04:40:38|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 04:40:38|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 04:55:46|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 04:55:47|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 04:55:49|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_1.mp4
[05-14 04:55:49|INFO|cosmos_predict2/inference.py:85:generate] [50/125] Processing sample gr1_env_10
[05-14 04:55:49|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_10' prompt_path=None prompt='Use the right hand to pick up chip with green packaging to white bag' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/10_Use the right hand to pick up chip with green packaging to white bag.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 04:55:49|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_10.json
[05-14 04:55:49|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 04:55:49|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 04:55:49|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/10_Use the right hand to pick up chip with green packaging to white bag.png
[05-14 04:55:50|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 04:55:50|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 05:10:59|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 05:11:00|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 05:11:01|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_10.mp4
[05-14 05:11:01|INFO|cosmos_predict2/inference.py:85:generate] [51/125] Processing sample gr1_env_11
[05-14 05:11:01|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_11' prompt_path=None prompt='Use the right hand to pick up cup to green trash can' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/11_Use the right hand to pick up cup to green trash can.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 05:11:01|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_11.json
[05-14 05:11:01|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 05:11:01|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 05:11:01|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/11_Use the right hand to pick up cup to green trash can.png
[05-14 05:11:02|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 05:11:02|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 05:26:10|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 05:26:11|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 05:26:13|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_11.mp4
[05-14 05:26:13|INFO|cosmos_predict2/inference.py:85:generate] [52/125] Processing sample gr1_env_12
[05-14 05:26:13|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_12' prompt_path=None prompt='Use the right hand to pick up orange to metal platform' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/12_Use the right hand to pick up orange to metal platform.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 05:26:13|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_12.json
[05-14 05:26:13|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 05:26:13|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 05:26:13|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/12_Use the right hand to pick up orange to metal platform.png
[05-14 05:26:14|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 05:26:14|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 05:41:22|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 05:41:24|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 05:41:26|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_12.mp4
[05-14 05:41:26|INFO|cosmos_predict2/inference.py:85:generate] [53/125] Processing sample gr1_env_13
[05-14 05:41:26|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_13' prompt_path=None prompt='Use the right hand to pick up orange to the serving dish' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/13_Use the right hand to pick up orange to the serving dish.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 05:41:26|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_13.json
[05-14 05:41:26|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 05:41:26|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 05:41:26|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/13_Use the right hand to pick up orange to the serving dish.png
[05-14 05:41:27|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 05:41:27|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 05:56:34|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 05:56:36|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 05:56:37|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_13.mp4
[05-14 05:56:37|INFO|cosmos_predict2/inference.py:85:generate] [54/125] Processing sample gr1_env_14
[05-14 05:56:37|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_14' prompt_path=None prompt='Use the right hand to pick up paper cup to metal container' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/14_Use the right hand to pick up paper cup to metal container.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 05:56:37|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_14.json
[05-14 05:56:37|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 05:56:37|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 05:56:37|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/14_Use the right hand to pick up paper cup to metal container.png
[05-14 05:56:38|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 05:56:38|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 06:11:45|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 06:11:46|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 06:11:48|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_14.mp4
[05-14 06:11:48|INFO|cosmos_predict2/inference.py:85:generate] [55/125] Processing sample gr1_env_15
[05-14 06:11:48|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_15' prompt_path=None prompt='Use the right hand to pick up paper roll to metal container' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/15_Use the right hand to pick up paper roll to metal container.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 06:11:48|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_15.json
[05-14 06:11:48|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 06:11:48|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 06:11:48|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/15_Use the right hand to pick up paper roll to metal container.png
[05-14 06:11:49|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 06:11:49|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 06:26:57|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 06:26:58|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 06:26:59|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_15.mp4
[05-14 06:26:59|INFO|cosmos_predict2/inference.py:85:generate] [56/125] Processing sample gr1_env_16
[05-14 06:26:59|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_16' prompt_path=None prompt='Use the right hand to pick up pear to bucket' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/16_Use the right hand to pick up pear to bucket.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 06:26:59|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_16.json
[05-14 06:26:59|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 06:26:59|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 06:26:59|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/16_Use the right hand to pick up pear to bucket.png
[05-14 06:27:00|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 06:27:00|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 06:42:08|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 06:42:09|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 06:42:10|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_16.mp4
[05-14 06:42:10|INFO|cosmos_predict2/inference.py:85:generate] [57/125] Processing sample gr1_env_17
[05-14 06:42:10|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_17' prompt_path=None prompt='Use the right hand to pick up plastic pitcher and pour water onto green plant' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/17_Use the right hand to pick up plastic pitcher and pour water onto green plant.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 06:42:10|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_17.json
[05-14 06:42:10|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 06:42:10|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 06:42:10|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/17_Use the right hand to pick up plastic pitcher and pour water onto green plant.png
[05-14 06:42:11|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 06:42:11|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 06:57:19|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 06:57:20|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 06:57:22|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_17.mp4
[05-14 06:57:22|INFO|cosmos_predict2/inference.py:85:generate] [58/125] Processing sample gr1_env_18
[05-14 06:57:22|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_18' prompt_path=None prompt='Use the right hand to pick up red apple on the wooden box to the bucket on the left' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/18_Use the right hand to pick up red apple on the wooden box to the bucket on the left.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 06:57:22|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_18.json
[05-14 06:57:22|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 06:57:22|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 06:57:22|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/18_Use the right hand to pick up red apple on the wooden box to the bucket on the left.png
[05-14 06:57:22|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 06:57:22|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 07:12:30|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 07:12:32|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 07:12:34|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_18.mp4
[05-14 07:12:34|INFO|cosmos_predict2/inference.py:85:generate] [59/125] Processing sample gr1_env_19
[05-14 07:12:34|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_19' prompt_path=None prompt='Use the right hand to pick up sandwich to pizza box' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/19_Use the right hand to pick up sandwich to pizza box.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 07:12:34|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_19.json
[05-14 07:12:34|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 07:12:34|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 07:12:34|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/19_Use the right hand to pick up sandwich to pizza box.png
[05-14 07:12:35|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 07:12:35|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 07:27:44|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 07:27:45|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 07:27:47|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_19.mp4
[05-14 07:27:47|INFO|cosmos_predict2/inference.py:85:generate] [60/125] Processing sample gr1_env_2
[05-14 07:27:47|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_2' prompt_path=None prompt='Use the right hand to close the black drawer' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/2_Use the right hand to close the black drawer.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 07:27:47|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_2.json
[05-14 07:27:47|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 07:27:47|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 07:27:47|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/2_Use the right hand to close the black drawer.png
[05-14 07:27:47|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 07:27:47|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 07:42:55|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 07:42:56|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 07:42:57|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_2.mp4
[05-14 07:42:57|INFO|cosmos_predict2/inference.py:85:generate] [61/125] Processing sample gr1_env_20
[05-14 07:42:57|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_20' prompt_path=None prompt='Use the right hand to pick up silver lid and cover the black pot' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/20_Use the right hand to pick up silver lid and cover the black pot.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 07:42:57|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_20.json
[05-14 07:42:57|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 07:42:57|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 07:42:57|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/20_Use the right hand to pick up silver lid and cover the black pot.png
[05-14 07:42:58|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 07:42:58|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 07:58:06|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 07:58:08|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 07:58:09|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_20.mp4
[05-14 07:58:09|INFO|cosmos_predict2/inference.py:85:generate] [62/125] Processing sample gr1_env_21
[05-14 07:58:09|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_21' prompt_path=None prompt='Use the right hand to pick up spatula to the table' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/21_Use the right hand to pick up spatula to the table.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 07:58:09|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_21.json
[05-14 07:58:09|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 07:58:09|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 07:58:09|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/21_Use the right hand to pick up spatula to the table.png
[05-14 07:58:10|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 07:58:10|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 08:13:19|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 08:13:20|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 08:13:21|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_21.mp4
[05-14 08:13:21|INFO|cosmos_predict2/inference.py:85:generate] [63/125] Processing sample gr1_env_22
[05-14 08:13:21|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_22' prompt_path=None prompt='Use the right hand to pick up the bottle and move it over the barcode scanner to check ou' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/22_Use the right hand to pick up the bottle and move it over the barcode scanner to check ou.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 08:13:21|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_22.json
[05-14 08:13:21|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 08:13:21|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 08:13:21|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/22_Use the right hand to pick up the bottle and move it over the barcode scanner to check ou.png
[05-14 08:13:21|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 08:13:21|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 08:28:30|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 08:28:31|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 08:28:32|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_22.mp4
[05-14 08:28:32|INFO|cosmos_predict2/inference.py:85:generate] [64/125] Processing sample gr1_env_23
[05-14 08:28:32|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_23' prompt_path=None prompt='Use the right hand to pick up the glass, place it under the Coca-Cola dispenser, and press the lever to fill the glass with Coca-Cola' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/23_Use the right hand to pick up the glass, place it under the Coca-Cola dispenser, and press the lever to fill the glass with Coca-Cola.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 08:28:32|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_23.json
[05-14 08:28:32|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 08:28:32|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 08:28:32|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/23_Use the right hand to pick up the glass, place it under the Coca-Cola dispenser, and press the lever to fill the glass with Coca-Cola.png
[05-14 08:28:32|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 08:28:32|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 08:43:41|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 08:43:42|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 08:43:43|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_23.mp4
[05-14 08:43:43|INFO|cosmos_predict2/inference.py:85:generate] [65/125] Processing sample gr1_env_24
[05-14 08:43:43|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_24' prompt_path=None prompt='Use the right hand to pick up the glass, place it under the Sprite dispenser, and press the lever to fill the glass with Sprite' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/24_Use the right hand to pick up the glass, place it under the Sprite dispenser, and press the lever to fill the glass with Sprite.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 08:43:43|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_24.json
[05-14 08:43:43|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 08:43:43|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 08:43:43|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/24_Use the right hand to pick up the glass, place it under the Sprite dispenser, and press the lever to fill the glass with Sprite.png
[05-14 08:43:43|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 08:43:43|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 08:58:51|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 08:58:52|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 08:58:53|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_24.mp4
[05-14 08:58:53|INFO|cosmos_predict2/inference.py:85:generate] [66/125] Processing sample gr1_env_25
[05-14 08:58:53|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_25' prompt_path=None prompt='Use the right hand to pick up the sauce bottle to white tray' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/25_Use the right hand to pick up the sauce bottle to white tray.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 08:58:53|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_25.json
[05-14 08:58:53|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 08:58:53|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 08:58:53|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/25_Use the right hand to pick up the sauce bottle to white tray.png
[05-14 08:58:54|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 08:58:54|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 09:13:59|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 09:14:00|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 09:14:01|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_25.mp4
[05-14 09:14:01|INFO|cosmos_predict2/inference.py:85:generate] [67/125] Processing sample gr1_env_26
[05-14 09:14:01|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_26' prompt_path=None prompt='Use the right hand to pick up the spoon and perform a mashing motion on the paper' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/26_Use the right hand to pick up the spoon and perform a mashing motion on the paper.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 09:14:01|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_26.json
[05-14 09:14:01|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 09:14:01|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 09:14:01|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/26_Use the right hand to pick up the spoon and perform a mashing motion on the paper.png
[05-14 09:14:01|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 09:14:01|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 09:29:05|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 09:29:06|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 09:29:09|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_26.mp4
[05-14 09:29:09|INFO|cosmos_predict2/inference.py:85:generate] [68/125] Processing sample gr1_env_27
[05-14 09:29:09|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_27' prompt_path=None prompt='Use the right hand to pick up the spoon from the bowl and serve the contents onto the plate' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/27_Use the right hand to pick up the spoon from the bowl and serve the contents onto the plate.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 09:29:09|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_27.json
[05-14 09:29:09|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 09:29:09|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 09:29:09|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/27_Use the right hand to pick up the spoon from the bowl and serve the contents onto the plate.png
[05-14 09:29:10|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 09:29:10|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 09:44:14|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 09:44:15|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 09:44:17|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_27.mp4
[05-14 09:44:17|INFO|cosmos_predict2/inference.py:85:generate] [69/125] Processing sample gr1_env_28
[05-14 09:44:17|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_28' prompt_path=None prompt='Use the right hand to pick up the wooden spoon and perform a scooping motion between the two bowls' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/28_Use the right hand to pick up the wooden spoon and perform a scooping motion between the two bowls.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 09:44:17|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_28.json
[05-14 09:44:17|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 09:44:17|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 09:44:17|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/28_Use the right hand to pick up the wooden spoon and perform a scooping motion between the two bowls.png
[05-14 09:44:18|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 09:44:18|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 09:59:26|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 09:59:27|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 09:59:28|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_28.mp4
[05-14 09:59:28|INFO|cosmos_predict2/inference.py:85:generate] [70/125] Processing sample gr1_env_3
[05-14 09:59:28|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_3' prompt_path=None prompt='Use the right hand to close the lid of the soup tureen' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/3_Use the right hand to close the lid of the soup tureen.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 09:59:28|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_3.json
[05-14 09:59:28|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 09:59:28|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 09:59:28|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/3_Use the right hand to close the lid of the soup tureen.png
[05-14 09:59:29|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 09:59:29|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 10:14:37|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 10:14:38|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 10:14:40|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_3.mp4
[05-14 10:14:40|INFO|cosmos_predict2/inference.py:85:generate] [71/125] Processing sample gr1_env_4
[05-14 10:14:40|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_4' prompt_path=None prompt='Use the right hand to close waffle maker' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/4_Use the right hand to close waffle maker.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 10:14:40|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_4.json
[05-14 10:14:40|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 10:14:40|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 10:14:40|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/4_Use the right hand to close waffle maker.png
[05-14 10:14:40|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 10:14:40|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 10:29:48|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 10:29:49|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 10:29:51|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_4.mp4
[05-14 10:29:51|INFO|cosmos_predict2/inference.py:85:generate] [72/125] Processing sample gr1_env_5
[05-14 10:29:51|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_5' prompt_path=None prompt='Use the right hand to grab the handle of lid and lift up to uncover the pot' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/5_Use the right hand to grab the handle of lid and lift up to uncover the pot.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 10:29:51|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_5.json
[05-14 10:29:51|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 10:29:51|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 10:29:51|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/5_Use the right hand to grab the handle of lid and lift up to uncover the pot.png
[05-14 10:29:52|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 10:29:52|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 10:45:00|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 10:45:02|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 10:45:03|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_5.mp4
[05-14 10:45:03|INFO|cosmos_predict2/inference.py:85:generate] [73/125] Processing sample gr1_env_6
[05-14 10:45:03|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_6' prompt_path=None prompt='Use the right hand to grab the handle of the wooden bucket and lift it up' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/6_Use the right hand to grab the handle of the wooden bucket and lift it up.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 10:45:03|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_6.json
[05-14 10:45:03|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 10:45:03|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 10:45:03|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/6_Use the right hand to grab the handle of the wooden bucket and lift it up.png
[05-14 10:45:03|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 10:45:03|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 11:00:10|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 11:00:11|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 11:00:13|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_6.mp4
[05-14 11:00:13|INFO|cosmos_predict2/inference.py:85:generate] [74/125] Processing sample gr1_env_7
[05-14 11:00:13|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_7' prompt_path=None prompt='Use the right hand to grab the pole of the sign and lift it up' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/7_Use the right hand to grab the pole of the sign and lift it up.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 11:00:13|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_7.json
[05-14 11:00:13|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 11:00:13|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 11:00:13|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/7_Use the right hand to grab the pole of the sign and lift it up.png
[05-14 11:00:13|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 11:00:13|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 11:15:20|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 11:15:21|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 11:15:23|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_7.mp4
[05-14 11:15:23|INFO|cosmos_predict2/inference.py:85:generate] [75/125] Processing sample gr1_env_8
[05-14 11:15:23|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_8' prompt_path=None prompt='Use the right hand to knock over the allergy information sign on the counter' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/8_Use the right hand to knock over the allergy information sign on the counter.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 11:15:23|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_8.json
[05-14 11:15:23|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 11:15:23|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 11:15:23|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/8_Use the right hand to knock over the allergy information sign on the counter.png
[05-14 11:15:24|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 11:15:24|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 11:30:31|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 11:30:32|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 11:30:33|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_8.mp4
[05-14 11:30:33|INFO|cosmos_predict2/inference.py:85:generate] [76/125] Processing sample gr1_env_9
[05-14 11:30:33|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_env_9' prompt_path=None prompt='Use the right hand to knock over the green Pocky box' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/9_Use the right hand to knock over the green Pocky box.png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 11:30:33|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_9.json
[05-14 11:30:33|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 11:30:33|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 11:30:33|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_env/9_Use the right hand to knock over the green Pocky box.png
[05-14 11:30:34|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 11:30:34|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 11:45:42|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 11:45:44|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 11:45:46|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_env_9.mp4
[05-14 11:45:46|INFO|cosmos_predict2/inference.py:85:generate] [77/125] Processing sample gr1_object_0
[05-14 11:45:46|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_0' prompt_path=None prompt='Use the left hand to pick up dark green cucumber from on circular gray mat to above beige bowl.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/0_Use the left hand to pick up dark green cucumber from on circular gray mat to above beige bowl..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 11:45:46|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_0.json
[05-14 11:45:46|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 11:45:46|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 11:45:46|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/0_Use the left hand to pick up dark green cucumber from on circular gray mat to above beige bowl..png
[05-14 11:45:47|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 11:45:47|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 12:00:53|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 12:00:54|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 12:00:56|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_0.mp4
[05-14 12:00:56|INFO|cosmos_predict2/inference.py:85:generate] [78/125] Processing sample gr1_object_1
[05-14 12:00:56|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_1' prompt_path=None prompt='Use the right hand to pick up orange juice carton from center of pink plate to center of green bowl.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/1_Use the right hand to pick up orange juice carton from center of pink plate to center of green bowl..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 12:00:56|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_1.json
[05-14 12:00:56|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 12:00:56|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 12:00:56|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/1_Use the right hand to pick up orange juice carton from center of pink plate to center of green bowl..png
[05-14 12:00:56|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 12:00:56|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 12:16:02|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 12:16:04|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 12:16:06|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_1.mp4
[05-14 12:16:06|INFO|cosmos_predict2/inference.py:85:generate] [79/125] Processing sample gr1_object_10
[05-14 12:16:06|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_10' prompt_path=None prompt='Use the right hand to pick up yellow star fruit from right side of table to white shelf.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/10_Use the right hand to pick up yellow star fruit from right side of table to white shelf..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 12:16:06|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_10.json
[05-14 12:16:06|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 12:16:06|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 12:16:06|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/10_Use the right hand to pick up yellow star fruit from right side of table to white shelf..png
[05-14 12:16:07|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 12:16:07|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 12:31:15|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 12:31:16|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 12:31:18|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_10.mp4
[05-14 12:31:18|INFO|cosmos_predict2/inference.py:85:generate] [80/125] Processing sample gr1_object_11
[05-14 12:31:18|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_11' prompt_path=None prompt='Use the right hand to pick up green pepper from black shelf to inside brown paper bag.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/11_Use the right hand to pick up green pepper from black shelf to inside brown paper bag..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 12:31:18|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_11.json
[05-14 12:31:18|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 12:31:18|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 12:31:18|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/11_Use the right hand to pick up green pepper from black shelf to inside brown paper bag..png
[05-14 12:31:18|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 12:31:18|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 12:46:24|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 12:46:25|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 12:46:26|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_11.mp4
[05-14 12:46:26|INFO|cosmos_predict2/inference.py:85:generate] [81/125] Processing sample gr1_object_13
[05-14 12:46:26|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_13' prompt_path=None prompt='Use the right hand to pick up red bell pepper from center of tan countertop to brown bottom rack.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/13_Use the right hand to pick up red bell pepper from center of tan countertop to brown bottom rack..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 12:46:26|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_13.json
[05-14 12:46:26|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 12:46:26|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 12:46:26|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/13_Use the right hand to pick up red bell pepper from center of tan countertop to brown bottom rack..png
[05-14 12:46:27|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 12:46:27|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 13:01:30|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 13:01:31|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 13:01:33|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_13.mp4
[05-14 13:01:33|INFO|cosmos_predict2/inference.py:85:generate] [82/125] Processing sample gr1_object_14
[05-14 13:01:33|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_14' prompt_path=None prompt="Use the right hand to pick up rubik's cube from from the top of the three-tiered wooden shelf to to the bottom of the three-tiered wooden shelf." negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath("/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/14_Use the right hand to pick up rubik's cube from from the top of the three-tiered wooden shelf to to the bottom of the three-tiered wooden shelf..png") resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 13:01:33|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_14.json
[05-14 13:01:33|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 13:01:33|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 13:01:33|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/14_Use the right hand to pick up rubik's cube from from the top of the three-tiered wooden shelf to to the bottom of the three-tiered wooden shelf..png
[05-14 13:01:33|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 13:01:33|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 13:16:37|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 13:16:38|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 13:16:40|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_14.mp4
[05-14 13:16:40|INFO|cosmos_predict2/inference.py:85:generate] [83/125] Processing sample gr1_object_15
[05-14 13:16:40|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_15' prompt_path=None prompt='Use the left hand to pick up red glass from light red plate to turquoise plate.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/15_Use the left hand to pick up red glass from light red plate to turquoise plate..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 13:16:40|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_15.json
[05-14 13:16:40|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 13:16:40|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 13:16:40|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/15_Use the left hand to pick up red glass from light red plate to turquoise plate..png
[05-14 13:16:41|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 13:16:41|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 13:31:46|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 13:31:47|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 13:31:48|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_15.mp4
[05-14 13:31:48|INFO|cosmos_predict2/inference.py:85:generate] [84/125] Processing sample gr1_object_16
[05-14 13:31:48|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_16' prompt_path=None prompt='Use the right hand to pick up milk carton from center of pink plate to pale turquoise plate.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/16_Use the right hand to pick up milk carton from center of pink plate to pale turquoise plate..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 13:31:48|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_16.json
[05-14 13:31:48|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 13:31:48|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 13:31:48|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/16_Use the right hand to pick up milk carton from center of pink plate to pale turquoise plate..png
[05-14 13:31:49|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 13:31:49|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 13:46:53|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 13:46:54|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 13:46:56|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_16.mp4
[05-14 13:46:56|INFO|cosmos_predict2/inference.py:85:generate] [85/125] Processing sample gr1_object_17
[05-14 13:46:56|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_17' prompt_path=None prompt='Use the right hand to pick up orange from pink plate to pale turquoise plate.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/17_Use the right hand to pick up orange from pink plate to pale turquoise plate..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 13:46:56|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_17.json
[05-14 13:46:56|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 13:46:56|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 13:46:56|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/17_Use the right hand to pick up orange from pink plate to pale turquoise plate..png
[05-14 13:46:57|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 13:46:57|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 14:02:03|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 14:02:04|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 14:02:05|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_17.mp4
[05-14 14:02:05|INFO|cosmos_predict2/inference.py:85:generate] [86/125] Processing sample gr1_object_18
[05-14 14:02:05|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_18' prompt_path=None prompt='Use the right hand to pick up corn on the cob from black shelf to inside brown paper bag.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/18_Use the right hand to pick up corn on the cob from black shelf to inside brown paper bag..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 14:02:05|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_18.json
[05-14 14:02:05|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 14:02:05|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 14:02:05|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/18_Use the right hand to pick up corn on the cob from black shelf to inside brown paper bag..png
[05-14 14:02:06|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 14:02:06|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 14:17:12|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 14:17:13|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 14:17:16|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_18.mp4
[05-14 14:17:16|INFO|cosmos_predict2/inference.py:85:generate] [87/125] Processing sample gr1_object_19
[05-14 14:17:16|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_19' prompt_path=None prompt='Use the left hand to pick up yellow mustard bottle from tan table to middle level of white shelf.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/19_Use the left hand to pick up yellow mustard bottle from tan table to middle level of white shelf..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 14:17:16|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_19.json
[05-14 14:17:16|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 14:17:16|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 14:17:16|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/19_Use the left hand to pick up yellow mustard bottle from tan table to middle level of white shelf..png
[05-14 14:17:17|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 14:17:17|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 14:32:22|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 14:32:23|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 14:32:25|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_19.mp4
[05-14 14:32:25|INFO|cosmos_predict2/inference.py:85:generate] [88/125] Processing sample gr1_object_2
[05-14 14:32:25|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_2' prompt_path=None prompt='Use the right hand to pick up tall red glass from center of table to center of bright blue plate.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/2_Use the right hand to pick up tall red glass from center of table to center of bright blue plate..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 14:32:25|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_2.json
[05-14 14:32:25|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 14:32:25|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 14:32:25|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/2_Use the right hand to pick up tall red glass from center of table to center of bright blue plate..png
[05-14 14:32:26|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 14:32:26|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 14:47:30|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 14:47:32|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 14:47:33|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_2.mp4
[05-14 14:47:33|INFO|cosmos_predict2/inference.py:85:generate] [89/125] Processing sample gr1_object_20
[05-14 14:47:33|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_20' prompt_path=None prompt='Use the right hand to pick up pink peach from center of white table to bottom level of the shelf.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/20_Use the right hand to pick up pink peach from center of white table to bottom level of the shelf..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 14:47:33|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_20.json
[05-14 14:47:33|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 14:47:33|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 14:47:33|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/20_Use the right hand to pick up pink peach from center of white table to bottom level of the shelf..png
[05-14 14:47:34|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 14:47:34|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 15:02:39|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 15:02:40|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 15:02:42|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_20.mp4
[05-14 15:02:42|INFO|cosmos_predict2/inference.py:85:generate] [90/125] Processing sample gr1_object_21
[05-14 15:02:42|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_21' prompt_path=None prompt='Use the right hand to pick up brown glass cup from tan table to blue plate.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/21_Use the right hand to pick up brown glass cup from tan table to blue plate..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 15:02:42|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_21.json
[05-14 15:02:42|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 15:02:42|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 15:02:42|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/21_Use the right hand to pick up brown glass cup from tan table to blue plate..png
[05-14 15:02:42|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 15:02:42|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 15:17:48|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 15:17:49|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 15:17:50|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_21.mp4
[05-14 15:17:50|INFO|cosmos_predict2/inference.py:85:generate] [91/125] Processing sample gr1_object_22
[05-14 15:17:50|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_22' prompt_path=None prompt='Use the left hand to pick up white plate from black rack to on top of teal plate.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/22_Use the left hand to pick up white plate from black rack to on top of teal plate..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 15:17:50|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_22.json
[05-14 15:17:50|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 15:17:50|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 15:17:50|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/22_Use the left hand to pick up white plate from black rack to on top of teal plate..png
[05-14 15:17:51|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 15:17:51|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 15:32:55|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 15:32:56|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 15:32:57|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_22.mp4
[05-14 15:32:57|INFO|cosmos_predict2/inference.py:85:generate] [92/125] Processing sample gr1_object_23
[05-14 15:32:57|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_23' prompt_path=None prompt='Use the right hand to pick up white mug from table mat to dark wooden tray.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/23_Use the right hand to pick up white mug from table mat to dark wooden tray..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 15:32:57|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_23.json
[05-14 15:32:57|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 15:32:57|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 15:32:57|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/23_Use the right hand to pick up white mug from table mat to dark wooden tray..png
[05-14 15:32:58|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 15:32:58|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 15:48:01|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 15:48:02|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 15:48:05|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_23.mp4
[05-14 15:48:05|INFO|cosmos_predict2/inference.py:85:generate] [93/125] Processing sample gr1_object_24
[05-14 15:48:05|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_24' prompt_path=None prompt='Use the right hand to pick up orange tangerine from pink plate to light blue plate.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/24_Use the right hand to pick up orange tangerine from pink plate to light blue plate..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 15:48:05|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_24.json
[05-14 15:48:05|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 15:48:05|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 15:48:05|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/24_Use the right hand to pick up orange tangerine from pink plate to light blue plate..png
[05-14 15:48:06|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 15:48:06|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 16:03:12|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 16:03:14|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 16:03:16|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_24.mp4
[05-14 16:03:16|INFO|cosmos_predict2/inference.py:85:generate] [94/125] Processing sample gr1_object_25
[05-14 16:03:16|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_25' prompt_path=None prompt='Use the left hand to pick up green pepper from tan table, below the bright blue plate to pale turquoise plate.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/25_Use the left hand to pick up green pepper from tan table, below the bright blue plate to pale turquoise plate..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 16:03:16|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_25.json
[05-14 16:03:16|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 16:03:16|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 16:03:16|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/25_Use the left hand to pick up green pepper from tan table, below the bright blue plate to pale turquoise plate..png
[05-14 16:03:16|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 16:03:16|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 16:18:20|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 16:18:21|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 16:18:23|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_25.mp4
[05-14 16:18:23|INFO|cosmos_predict2/inference.py:85:generate] [95/125] Processing sample gr1_object_26
[05-14 16:18:23|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_26' prompt_path=None prompt='Use the left hand to pick up turquoise plate from plate rack to cornflower blue plate.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/26_Use the left hand to pick up turquoise plate from plate rack to cornflower blue plate..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 16:18:23|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_26.json
[05-14 16:18:23|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 16:18:23|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 16:18:23|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/26_Use the left hand to pick up turquoise plate from plate rack to cornflower blue plate..png
[05-14 16:18:24|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 16:18:24|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 16:33:29|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 16:33:30|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 16:33:32|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_26.mp4
[05-14 16:33:32|INFO|cosmos_predict2/inference.py:85:generate] [96/125] Processing sample gr1_object_27
[05-14 16:33:32|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_27' prompt_path=None prompt='Use the left hand to pick up dragonfruit from center of tan table to top level of wooden shelf.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/27_Use the left hand to pick up dragonfruit from center of tan table to top level of wooden shelf..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 16:33:32|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_27.json
[05-14 16:33:32|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 16:33:32|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 16:33:32|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/27_Use the left hand to pick up dragonfruit from center of tan table to top level of wooden shelf..png
[05-14 16:33:33|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 16:33:33|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 16:48:38|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 16:48:39|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 16:48:41|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_27.mp4
[05-14 16:48:41|INFO|cosmos_predict2/inference.py:85:generate] [97/125] Processing sample gr1_object_28
[05-14 16:48:41|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_28' prompt_path=None prompt='Use the right hand to pick up tall red glass from center of tan table to bright blue plate.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/28_Use the right hand to pick up tall red glass from center of tan table to bright blue plate..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 16:48:41|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_28.json
[05-14 16:48:41|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 16:48:41|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 16:48:41|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/28_Use the right hand to pick up tall red glass from center of tan table to bright blue plate..png
[05-14 16:48:42|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 16:48:42|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 17:03:46|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 17:03:47|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 17:03:49|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_28.mp4
[05-14 17:03:49|INFO|cosmos_predict2/inference.py:85:generate] [98/125] Processing sample gr1_object_29
[05-14 17:03:49|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_29' prompt_path=None prompt='Use the right hand to pick up red tomato from upper black tray of plastic shelf to inside of brown paper bag.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/29_Use the right hand to pick up red tomato from upper black tray of plastic shelf to inside of brown paper bag..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 17:03:49|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_29.json
[05-14 17:03:49|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 17:03:49|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 17:03:49|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/29_Use the right hand to pick up red tomato from upper black tray of plastic shelf to inside of brown paper bag..png
[05-14 17:03:49|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 17:03:49|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 17:18:54|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 17:18:55|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 17:18:57|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_29.mp4
[05-14 17:18:57|INFO|cosmos_predict2/inference.py:85:generate] [99/125] Processing sample gr1_object_3
[05-14 17:18:57|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_3' prompt_path=None prompt='Use the right hand to pick up green bok choy from from the center of the table to to the bottom rack of the black two-tiered wire basket.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/3_Use the right hand to pick up green bok choy from from the center of the table to to the bottom rack of the black two-tiered wire basket..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 17:18:57|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_3.json
[05-14 17:18:57|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 17:18:57|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 17:18:57|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/3_Use the right hand to pick up green bok choy from from the center of the table to to the bottom rack of the black two-tiered wire basket..png
[05-14 17:18:58|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 17:18:58|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 17:34:01|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 17:34:02|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 17:34:04|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_3.mp4
[05-14 17:34:04|INFO|cosmos_predict2/inference.py:85:generate] [100/125] Processing sample gr1_object_30
[05-14 17:34:04|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_30' prompt_path=None prompt='Use the right hand to pick up purple grapes from center of table to white tray.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/30_Use the right hand to pick up purple grapes from center of table to white tray..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 17:34:04|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_30.json
[05-14 17:34:04|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 17:34:04|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 17:34:04|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/30_Use the right hand to pick up purple grapes from center of table to white tray..png
[05-14 17:34:05|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 17:34:05|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 17:49:09|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 17:49:11|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 17:49:12|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_30.mp4
[05-14 17:49:12|INFO|cosmos_predict2/inference.py:85:generate] [101/125] Processing sample gr1_object_31
[05-14 17:49:12|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_31' prompt_path=None prompt='Use the right hand to pick up tomato from middle of table to bottom white shelf.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/31_Use the right hand to pick up tomato from middle of table to bottom white shelf..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 17:49:12|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_31.json
[05-14 17:49:12|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 17:49:12|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 17:49:12|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/31_Use the right hand to pick up tomato from middle of table to bottom white shelf..png
[05-14 17:49:13|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 17:49:13|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 18:04:20|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 18:04:21|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 18:04:23|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_31.mp4
[05-14 18:04:23|INFO|cosmos_predict2/inference.py:85:generate] [102/125] Processing sample gr1_object_32
[05-14 18:04:23|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_32' prompt_path=None prompt='Use the right hand to pick up milk bottle from light blue bowl to light blue plate.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/32_Use the right hand to pick up milk bottle from light blue bowl to light blue plate..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 18:04:23|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_32.json
[05-14 18:04:23|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 18:04:23|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 18:04:23|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/32_Use the right hand to pick up milk bottle from light blue bowl to light blue plate..png
[05-14 18:04:24|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 18:04:24|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 18:19:29|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 18:19:31|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 18:19:33|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_32.mp4
[05-14 18:19:33|INFO|cosmos_predict2/inference.py:85:generate] [103/125] Processing sample gr1_object_33
[05-14 18:19:33|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_33' prompt_path=None prompt='Use the right hand to pick up red cup from blue plate to blue plate.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/33_Use the right hand to pick up red cup from blue plate to blue plate..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 18:19:33|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_33.json
[05-14 18:19:33|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 18:19:33|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 18:19:33|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/33_Use the right hand to pick up red cup from blue plate to blue plate..png
[05-14 18:19:33|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 18:19:33|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 18:34:37|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 18:34:38|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 18:34:40|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_33.mp4
[05-14 18:34:40|INFO|cosmos_predict2/inference.py:85:generate] [104/125] Processing sample gr1_object_34
[05-14 18:34:40|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_34' prompt_path=None prompt='Use the right hand to pick up milk carton from pale turquoise plate in the center of the table to second level of wooden shelf.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/34_Use the right hand to pick up milk carton from pale turquoise plate in the center of the table to second level of wooden shelf..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 18:34:40|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_34.json
[05-14 18:34:40|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 18:34:40|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 18:34:40|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/34_Use the right hand to pick up milk carton from pale turquoise plate in the center of the table to second level of wooden shelf..png
[05-14 18:34:40|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 18:34:40|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 18:49:45|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 18:49:46|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 18:49:48|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_34.mp4
[05-14 18:49:48|INFO|cosmos_predict2/inference.py:85:generate] [105/125] Processing sample gr1_object_35
[05-14 18:49:48|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_35' prompt_path=None prompt='Use the right hand to pick up peach from center of table to bottom level white tray.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/35_Use the right hand to pick up peach from center of table to bottom level white tray..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 18:49:48|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_35.json
[05-14 18:49:48|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 18:49:48|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 18:49:48|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/35_Use the right hand to pick up peach from center of table to bottom level white tray..png
[05-14 18:49:49|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 18:49:49|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 19:04:54|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 19:04:55|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 19:04:57|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_35.mp4
[05-14 19:04:57|INFO|cosmos_predict2/inference.py:85:generate] [106/125] Processing sample gr1_object_36
[05-14 19:04:57|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_36' prompt_path=None prompt='Use the left hand to pick up red glass from dark blue plate to light green plate.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/36_Use the left hand to pick up red glass from dark blue plate to light green plate..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 19:04:57|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_36.json
[05-14 19:04:57|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 19:04:57|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 19:04:57|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/36_Use the left hand to pick up red glass from dark blue plate to light green plate..png
[05-14 19:04:57|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 19:04:57|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 19:20:02|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 19:20:03|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 19:20:06|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_36.mp4
[05-14 19:20:06|INFO|cosmos_predict2/inference.py:85:generate] [107/125] Processing sample gr1_object_37
[05-14 19:20:06|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_37' prompt_path=None prompt='Use the right hand to pick up pink peach from right side of table to lower white tray of plastic shelf.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/37_Use the right hand to pick up pink peach from right side of table to lower white tray of plastic shelf..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 19:20:06|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_37.json
[05-14 19:20:06|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 19:20:06|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 19:20:06|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/37_Use the right hand to pick up pink peach from right side of table to lower white tray of plastic shelf..png
[05-14 19:20:07|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 19:20:07|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 19:35:12|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 19:35:13|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 19:35:14|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_37.mp4
[05-14 19:35:14|INFO|cosmos_predict2/inference.py:85:generate] [108/125] Processing sample gr1_object_38
[05-14 19:35:14|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_38' prompt_path=None prompt='Use the left hand to pick up green apple from tan countertop to brown box.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/38_Use the left hand to pick up green apple from tan countertop to brown box..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 19:35:14|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_38.json
[05-14 19:35:14|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 19:35:14|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 19:35:14|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/38_Use the left hand to pick up green apple from tan countertop to brown box..png
[05-14 19:35:14|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 19:35:14|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 19:50:19|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 19:50:20|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 19:50:22|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_38.mp4
[05-14 19:50:22|INFO|cosmos_predict2/inference.py:85:generate] [109/125] Processing sample gr1_object_39
[05-14 19:50:22|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_39' prompt_path=None prompt='Use the left hand to pick up the tall orange cup from from the large blue plate to to the small teal plate.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/39_Use the left hand to pick up the tall orange cup from from the large blue plate to to the small teal plate..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 19:50:22|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_39.json
[05-14 19:50:22|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 19:50:22|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 19:50:22|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/39_Use the left hand to pick up the tall orange cup from from the large blue plate to to the small teal plate..png
[05-14 19:50:22|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 19:50:22|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 20:05:27|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 20:05:28|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 20:05:29|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_39.mp4
[05-14 20:05:29|INFO|cosmos_predict2/inference.py:85:generate] [110/125] Processing sample gr1_object_4
[05-14 20:05:29|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_4' prompt_path=None prompt='Use the right hand to pick up orange carrot from center of table to lower white shelf.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/4_Use the right hand to pick up orange carrot from center of table to lower white shelf..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 20:05:29|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_4.json
[05-14 20:05:29|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 20:05:29|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 20:05:29|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/4_Use the right hand to pick up orange carrot from center of table to lower white shelf..png
[05-14 20:05:29|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 20:05:29|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 20:20:34|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 20:20:35|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 20:20:37|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_4.mp4
[05-14 20:20:37|INFO|cosmos_predict2/inference.py:85:generate] [111/125] Processing sample gr1_object_40
[05-14 20:20:37|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_40' prompt_path=None prompt='Use the right hand to pick up red apple from brown tray to beige placemat.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/40_Use the right hand to pick up red apple from brown tray to beige placemat..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 20:20:37|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_40.json
[05-14 20:20:37|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 20:20:37|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 20:20:37|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/40_Use the right hand to pick up red apple from brown tray to beige placemat..png
[05-14 20:20:37|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 20:20:37|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 20:35:43|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 20:35:44|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 20:35:46|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_40.mp4
[05-14 20:35:46|INFO|cosmos_predict2/inference.py:85:generate] [112/125] Processing sample gr1_object_41
[05-14 20:35:46|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_41' prompt_path=None prompt='Use the right hand to pick up orange juice box from brown box to white placemat.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/41_Use the right hand to pick up orange juice box from brown box to white placemat..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 20:35:46|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_41.json
[05-14 20:35:46|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 20:35:46|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 20:35:46|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/41_Use the right hand to pick up orange juice box from brown box to white placemat..png
[05-14 20:35:46|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 20:35:46|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 20:50:50|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 20:50:51|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 20:50:52|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_41.mp4
[05-14 20:50:52|INFO|cosmos_predict2/inference.py:85:generate] [113/125] Processing sample gr1_object_42
[05-14 20:50:52|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_42' prompt_path=None prompt="Use the right hand to pick up rubik's cube from bottom brown shelf to top brown shelf." negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath("/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/42_Use the right hand to pick up rubik's cube from bottom brown shelf to top brown shelf..png") resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 20:50:52|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_42.json
[05-14 20:50:52|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 20:50:52|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 20:50:52|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/42_Use the right hand to pick up rubik's cube from bottom brown shelf to top brown shelf..png
[05-14 20:50:53|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 20:50:53|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 21:05:57|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 21:05:58|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 21:06:00|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_42.mp4
[05-14 21:06:00|INFO|cosmos_predict2/inference.py:85:generate] [114/125] Processing sample gr1_object_43
[05-14 21:06:00|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_43' prompt_path=None prompt='Use the left hand to pick up pink dragonfruit from blue plate to pale turquoise plate.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/43_Use the left hand to pick up pink dragonfruit from blue plate to pale turquoise plate..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 21:06:00|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_43.json
[05-14 21:06:00|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 21:06:00|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 21:06:00|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/43_Use the left hand to pick up pink dragonfruit from blue plate to pale turquoise plate..png
[05-14 21:06:00|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 21:06:00|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 21:21:06|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 21:21:07|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 21:21:08|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_43.mp4
[05-14 21:21:08|INFO|cosmos_predict2/inference.py:85:generate] [115/125] Processing sample gr1_object_44
[05-14 21:21:08|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_44' prompt_path=None prompt='Use the right hand to pick up eggplant from grey mat on the center of the table to bowl in the center left of the table.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/44_Use the right hand to pick up eggplant from grey mat on the center of the table to bowl in the center left of the table..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 21:21:08|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_44.json
[05-14 21:21:08|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 21:21:08|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 21:21:08|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/44_Use the right hand to pick up eggplant from grey mat on the center of the table to bowl in the center left of the table..png
[05-14 21:21:09|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 21:21:09|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 21:36:14|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 21:36:15|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 21:36:17|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_44.mp4
[05-14 21:36:17|INFO|cosmos_predict2/inference.py:85:generate] [116/125] Processing sample gr1_object_45
[05-14 21:36:17|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_45' prompt_path=None prompt="Use the right hand to pick up rubik's cube from turquoise plate to middle tier of multi level wooden shelf." negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath("/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/45_Use the right hand to pick up rubik's cube from turquoise plate to middle tier of multi level wooden shelf..png") resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 21:36:17|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_45.json
[05-14 21:36:17|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 21:36:17|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 21:36:17|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/45_Use the right hand to pick up rubik's cube from turquoise plate to middle tier of multi level wooden shelf..png
[05-14 21:36:17|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 21:36:17|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 21:51:23|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 21:51:24|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 21:51:27|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_45.mp4
[05-14 21:51:27|INFO|cosmos_predict2/inference.py:85:generate] [117/125] Processing sample gr1_object_46
[05-14 21:51:27|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_46' prompt_path=None prompt='Use the right hand to pick up dragon fruit from gray basket to tan table towards the middle.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/46_Use the right hand to pick up dragon fruit from gray basket to tan table towards the middle..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 21:51:27|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_46.json
[05-14 21:51:27|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 21:51:27|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 21:51:27|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/46_Use the right hand to pick up dragon fruit from gray basket to tan table towards the middle..png
[05-14 21:51:28|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 21:51:28|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 22:06:31|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 22:06:32|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 22:06:35|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_46.mp4
[05-14 22:06:35|INFO|cosmos_predict2/inference.py:85:generate] [118/125] Processing sample gr1_object_47
[05-14 22:06:35|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_47' prompt_path=None prompt='Use the right hand to pick up blue circular clock from brown top shelf to blue plate.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/47_Use the right hand to pick up blue circular clock from brown top shelf to blue plate..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 22:06:35|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_47.json
[05-14 22:06:35|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 22:06:35|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 22:06:35|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/47_Use the right hand to pick up blue circular clock from brown top shelf to blue plate..png
[05-14 22:06:36|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 22:06:36|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 22:21:41|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 22:21:43|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 22:21:44|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_47.mp4
[05-14 22:21:44|INFO|cosmos_predict2/inference.py:85:generate] [119/125] Processing sample gr1_object_48
[05-14 22:21:44|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_48' prompt_path=None prompt='Use the right hand to pick up green bell pepper from middle of table to lower level of wire fruit basket.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/48_Use the right hand to pick up green bell pepper from middle of table to lower level of wire fruit basket..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 22:21:44|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_48.json
[05-14 22:21:44|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 22:21:44|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 22:21:44|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/48_Use the right hand to pick up green bell pepper from middle of table to lower level of wire fruit basket..png
[05-14 22:21:45|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 22:21:45|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 22:36:50|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 22:36:51|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 22:36:53|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_48.mp4
[05-14 22:36:53|INFO|cosmos_predict2/inference.py:85:generate] [120/125] Processing sample gr1_object_49
[05-14 22:36:53|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_49' prompt_path=None prompt="Use the right hand to pick up rubik's cube from center of pink plate to center of turquoise plate." negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath("/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/49_Use the right hand to pick up rubik's cube from center of pink plate to center of turquoise plate..png") resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 22:36:53|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_49.json
[05-14 22:36:53|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 22:36:53|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 22:36:53|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/49_Use the right hand to pick up rubik's cube from center of pink plate to center of turquoise plate..png
[05-14 22:36:54|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 22:36:54|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 22:51:59|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 22:52:00|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 22:52:02|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_49.mp4
[05-14 22:52:02|INFO|cosmos_predict2/inference.py:85:generate] [121/125] Processing sample gr1_object_5
[05-14 22:52:02|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_5' prompt_path=None prompt='Use the right hand to pick up green bok choy from top tier of two tier black and white shelf to brown paper bag.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/5_Use the right hand to pick up green bok choy from top tier of two tier black and white shelf to brown paper bag..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 22:52:02|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_5.json
[05-14 22:52:02|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 22:52:02|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 22:52:02|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/5_Use the right hand to pick up green bok choy from top tier of two tier black and white shelf to brown paper bag..png
[05-14 22:52:02|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 22:52:02|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 23:07:10|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 23:07:12|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 23:07:13|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_5.mp4
[05-14 23:07:13|INFO|cosmos_predict2/inference.py:85:generate] [122/125] Processing sample gr1_object_6
[05-14 23:07:13|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_6' prompt_path=None prompt='Use the right hand to pick up orange from center of the table to lower level white shelf on the right side of the table.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/6_Use the right hand to pick up orange from center of the table to lower level white shelf on the right side of the table..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 23:07:13|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_6.json
[05-14 23:07:13|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 23:07:13|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 23:07:13|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/6_Use the right hand to pick up orange from center of the table to lower level white shelf on the right side of the table..png
[05-14 23:07:13|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 23:07:13|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 23:22:21|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 23:22:22|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 23:22:24|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_6.mp4
[05-14 23:22:24|INFO|cosmos_predict2/inference.py:85:generate] [123/125] Processing sample gr1_object_7
[05-14 23:22:24|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_7' prompt_path=None prompt='Use the right hand to pick up the milk box from center of pink bowl to center of pink plate.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/7_Use the right hand to pick up the milk box from center of pink bowl to center of pink plate..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 23:22:24|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_7.json
[05-14 23:22:24|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 23:22:24|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 23:22:24|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/7_Use the right hand to pick up the milk box from center of pink bowl to center of pink plate..png
[05-14 23:22:24|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 23:22:24|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 23:37:30|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 23:37:31|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 23:37:32|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_7.mp4
[05-14 23:37:32|INFO|cosmos_predict2/inference.py:85:generate] [124/125] Processing sample gr1_object_8
[05-14 23:37:32|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_8' prompt_path=None prompt="Use the right hand to pick up rubik's cube from from the bottom of the three-tiered wooden shelf to to the top of the three-tiered wooden shelf." negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath("/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/8_Use the right hand to pick up rubik's cube from from the bottom of the three-tiered wooden shelf to to the top of the three-tiered wooden shelf..png") resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 23:37:32|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_8.json
[05-14 23:37:32|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 23:37:32|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 23:37:32|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/8_Use the right hand to pick up rubik's cube from from the bottom of the three-tiered wooden shelf to to the top of the three-tiered wooden shelf..png
[05-14 23:37:33|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 23:37:33|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-14 23:52:39|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-14 23:52:40|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-14 23:52:43|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_8.mp4
[05-14 23:52:43|INFO|cosmos_predict2/inference.py:85:generate] [125/125] Processing sample gr1_object_9
[05-14 23:52:43|DEBUG|cosmos_predict2/inference.py:92:_generate_sample] InferenceArguments(name='gr1_object_9' prompt_path=None prompt='Use the right hand to pick up milk carton from from the angled wooden shelf to to the cyan plate.' negative_prompt='The video captures a series of frames showing ugly scenes, static with no motion, motion blur, over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and flickering. Overall, the video is of poor quality.' seed=0 guidance=7 inference_type=<InferenceType.IMAGE2WORLD: 'image2world'> input_path=PosixPath('/pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/9_Use the right hand to pick up milk carton from from the angled wooden shelf to to the cyan plate..png') resolution='800,1280' num_output_frames=77 num_steps=35 enable_autoregressive=False chunk_size=77 chunk_overlap=1)
[05-14 23:52:43|INFO|cosmos_predict2/inference.py:98:_generate_sample] Saved arguments to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_9.json
[05-14 23:52:43|WARNING|cosmos_predict2/inference.py:112:_generate_sample] Guardrail checks on prompt are disabled
[05-14 23:52:43|INFO|cosmos_predict2/inference.py:132:_generate_sample] Generating video with standard mode...
[05-14 23:52:43|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:521:generate_vid2world] Processing image input: /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/video_world/video_gen_physics/tmp/dream_gen_benchmark/gr1_object/9_Use the right hand to pick up milk carton from from the angled wooden shelf to to the cyan plate..png
[05-14 23:52:43|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:559:generate_vid2world] GPU memory usage after getting data_batch: 19.87 GB
[05-14 23:52:43|INFO|cosmos_predict2/_src/predict2/inference/video2world.py:597:generate_vid2world] [Memory Optimization] Starting latent sample generation
[05-15 00:07:51|WARNING|cosmos_predict2/inference.py:167:_generate_sample] Guardrail checks on video are disabled
[05-15 00:07:52|DEBUG|cosmos_predict2/_src/imaginaire/utils/easy_io/handlers/imageio_video_handler.py:163:dump_to_fileobj] mimsave_kwargs: {'fps': 16, 'quality': 5, 'macro_block_size': 1, 'ffmpeg_params': ['-s', '1280x800'], 'output_params': ['-f', 'mp4']}
[05-15 00:07:53|SUCCESS|cosmos_predict2/inference.py:170:_generate_sample] Saved video to /pfss/mlde/workspaces/mlde_wsp_IAS_SAMMerge/VLA/doanh/cosmos-predict2.5/outputs/benchmark_inference_results/gr1_object_9.mp4