idm / analysis /diff_seed_1 /summary.json
zcyqyq's picture
Upload folder using huggingface_hub
56092de verified
Raw History Blame Contribute Delete
10.8 kB
{
"pair": "diff_seed",
"models": [
{
"checkpoint": "/vepfs-mlp2/c20250508/242303041/caiyi/Structured-4D-Model/outputs/recon_image_resnet_idm_sweeps/20260812-163334/1pct/checkpoints/20000.ckpt",
"kind": "image",
"backend": "unet",
"training_seed": 42,
"horizon": 64,
"weights": "model",
"checked_stat_files": [
"/vepfs-mlp2/c20250508/242303041/caiyi/Structured-4D-Model/outputs/recon_image_resnet_idm_sweeps/20260812-163334/1pct/StackCube-v1_image_data_stat.json"
],
"inference_passes": 1
},
{
"checkpoint": "/vepfs-mlp2/c20250508/242303041/caiyi/Structured-4D-Model/outputs/iterative_image_diff_seed_1to16/20260928-022436/bootstrap/seed_43/checkpoints/20000.ckpt",
"kind": "image",
"backend": "unet",
"training_seed": 43,
"horizon": 64,
"weights": "model",
"checked_stat_files": [
"/vepfs-mlp2/c20250508/242303041/caiyi/Structured-4D-Model/outputs/iterative_image_diff_seed_1to16/20260928-022436/bootstrap/seed_43/StackCube-v1_image_data_stat.json"
],
"inference_passes": 1
}
],
"arguments": {
"pair": "diff_seed",
"checkpoint_a": "/vepfs-mlp2/c20250508/242303041/caiyi/Structured-4D-Model/outputs/recon_image_resnet_idm_sweeps/20260812-163334/1pct/checkpoints/20000.ckpt",
"checkpoint_b": "/vepfs-mlp2/c20250508/242303041/caiyi/Structured-4D-Model/outputs/iterative_image_diff_seed_1to16/20260928-022436/bootstrap/seed_43/checkpoints/20000.ckpt",
"image_h5": "/vepfs-mlp2/c20250508/242303041/caiyi/Structured-4D-Model/render/image_idm_four_checkpoints_200/full/rollout_recon_rgb_224.h5",
"pointcloud_h5": "/vepfs-mlp2/c20250508/242303041/caiyi/Structured-4D-Model/render/image_idm_four_checkpoints_200/full/rollout_pointclouds.h5",
"output_dir": "outputs/rollout200_correlation_new_pair/diff_seed_1",
"obs_gap": 64,
"compare_steps": 32,
"step_stride": 32,
"batch_size": 8,
"seeds": [
42
],
"device": "cuda:0"
},
"num_chunks": 739,
"num_trajectories": 200,
"metrics": {
"arm": {
"means": {
"disagreement": 0.0033260114767449164,
"error_a": 0.0035830417412364124,
"error_b": 0.005043686401679642,
"mean_error": 0.004313364071458028
},
"correlations": {
"error_a": {
"chunk": {
"n": 739,
"pearson": 0.1222688717566926,
"spearman": 0.11621484846673798
},
"trajectory": {
"n": 200,
"pearson": 0.11741286754251265,
"spearman": 0.15594839870996777
}
},
"error_b": {
"chunk": {
"n": 739,
"pearson": 0.41969508365582764,
"spearman": 0.38106074324101324
},
"trajectory": {
"n": 200,
"pearson": 0.5187534956198987,
"spearman": 0.49114827870696776
}
},
"mean_error": {
"chunk": {
"n": 739,
"pearson": 0.3428202686110324,
"spearman": 0.39248094591175225
},
"trajectory": {
"n": 200,
"pearson": 0.3949199896198202,
"spearman": 0.4293347333683343
}
}
}
},
"gripper": {
"means": {
"disagreement": 0.769149939667231,
"error_a": 0.906909602216408,
"error_b": 1.1394822955686388,
"mean_error": 1.0231959488925235
},
"correlations": {
"error_a": {
"chunk": {
"n": 739,
"pearson": 0.21330205893235146,
"spearman": 0.43535658402566974
},
"trajectory": {
"n": 200,
"pearson": 0.10414473995437332,
"spearman": 0.10361959048976226
}
},
"error_b": {
"chunk": {
"n": 739,
"pearson": 0.4486463768789744,
"spearman": 0.5830387374788694
},
"trajectory": {
"n": 200,
"pearson": 0.6200008592889438,
"spearman": 0.6070036750918775
}
},
"mean_error": {
"chunk": {
"n": 739,
"pearson": 0.3790466308313798,
"spearman": 0.6039247849384533
},
"trajectory": {
"n": 200,
"pearson": 0.41933898635329725,
"spearman": 0.4685667141678543
}
}
}
},
"overall": {
"means": {
"disagreement": 0.09905400250055567,
"error_a": 0.11649886180063289,
"error_b": 0.14684851254754952,
"mean_error": 0.13167368717409123
},
"correlations": {
"error_a": {
"chunk": {
"n": 739,
"pearson": 0.21070899739306065,
"spearman": 0.3333790934734861
},
"trajectory": {
"n": 200,
"pearson": 0.09531514807835734,
"spearman": 0.09430435760894024
}
},
"error_b": {
"chunk": {
"n": 739,
"pearson": 0.44758837539622703,
"spearman": 0.5208907676612882
},
"trajectory": {
"n": 200,
"pearson": 0.6183365622017337,
"spearman": 0.6050596264906623
}
},
"mean_error": {
"chunk": {
"n": 739,
"pearson": 0.37721046206740283,
"spearman": 0.533760755045286
},
"trajectory": {
"n": 200,
"pearson": 0.41390879551606896,
"spearman": 0.46195454886372167
}
}
}
}
},
"per_seed": [
{
"seed": 42,
"metrics": {
"arm": {
"means": {
"disagreement": 0.0033260114767449164,
"error_a": 0.0035830417412364124,
"error_b": 0.005043686401679642,
"mean_error": 0.004313364071458028
},
"correlations": {
"error_a": {
"chunk": {
"n": 739,
"pearson": 0.1222688717566926,
"spearman": 0.11621484846673798
},
"trajectory": {
"n": 200,
"pearson": 0.11741286754251265,
"spearman": 0.15594839870996777
}
},
"error_b": {
"chunk": {
"n": 739,
"pearson": 0.41969508365582764,
"spearman": 0.38106074324101324
},
"trajectory": {
"n": 200,
"pearson": 0.5187534956198987,
"spearman": 0.49114827870696776
}
},
"mean_error": {
"chunk": {
"n": 739,
"pearson": 0.3428202686110324,
"spearman": 0.39248094591175225
},
"trajectory": {
"n": 200,
"pearson": 0.3949199896198202,
"spearman": 0.4293347333683343
}
}
}
},
"gripper": {
"means": {
"disagreement": 0.769149939667231,
"error_a": 0.906909602216408,
"error_b": 1.1394822955686388,
"mean_error": 1.0231959488925235
},
"correlations": {
"error_a": {
"chunk": {
"n": 739,
"pearson": 0.21330205893235146,
"spearman": 0.43535658402566974
},
"trajectory": {
"n": 200,
"pearson": 0.10414473995437332,
"spearman": 0.10361959048976226
}
},
"error_b": {
"chunk": {
"n": 739,
"pearson": 0.4486463768789744,
"spearman": 0.5830387374788694
},
"trajectory": {
"n": 200,
"pearson": 0.6200008592889438,
"spearman": 0.6070036750918775
}
},
"mean_error": {
"chunk": {
"n": 739,
"pearson": 0.3790466308313798,
"spearman": 0.6039247849384533
},
"trajectory": {
"n": 200,
"pearson": 0.41933898635329725,
"spearman": 0.4685667141678543
}
}
}
},
"overall": {
"means": {
"disagreement": 0.09905400250055567,
"error_a": 0.11649886180063289,
"error_b": 0.14684851254754952,
"mean_error": 0.13167368717409123
},
"correlations": {
"error_a": {
"chunk": {
"n": 739,
"pearson": 0.21070899739306065,
"spearman": 0.3333790934734861
},
"trajectory": {
"n": 200,
"pearson": 0.09531514807835734,
"spearman": 0.09430435760894024
}
},
"error_b": {
"chunk": {
"n": 739,
"pearson": 0.44758837539622703,
"spearman": 0.5208907676612882
},
"trajectory": {
"n": 200,
"pearson": 0.6183365622017337,
"spearman": 0.6050596264906623
}
},
"mean_error": {
"chunk": {
"n": 739,
"pearson": 0.37721046206740283,
"spearman": 0.533760755045286
},
"trajectory": {
"n": 200,
"pearson": 0.41390879551606896,
"spearman": 0.46195454886372167
}
}
}
}
}
}
],
"protocol": {
"action_space": "normalized abs_joint; shared StackCube DATA_STAT",
"target": "executed rollout actions, not expert optimal actions",
"observations": "stored real start/end observations; not online WM target images",
"rng": "idm-rollout-chunk-v2; same root/trajectory/step streams across checkpoints and modalities; shared noise for both pair types",
"sampling": "mean scalar MSE/disagreement across inference seeds; ACT zero latent evaluated once",
"trajectory": "equal mean of chunk metrics within each trajectory",
"selection": "fixed user checkpoints; no fitting or checkpoint selection on this rollout set",
"caveat": "chunks within a trajectory are dependent; only 20 trajectory samples"
}
}