{
  "apical_warmup": {
    "last": {
      "calibration_mse": 518.2958984375,
      "prediction_target_cosine": 0.0001586056141805217,
      "target_power": 518.2958984375
    },
    "steps": 100
  },
  "architecture": {
    "adaptive_apical_parameters": 2260992,
    "base_width": 16,
    "blocks_per_stage": 3,
    "bn_eps": 1e-05,
    "bn_momentum": 0.1,
    "depth": 20,
    "family": "CIFAR 6n+2 ResNet, option-A shortcuts",
    "fixed_traffic_coefficients": 188416,
    "forward_parameters": 269722,
    "hidden_shapes": [
      [
        16,
        32,
        32
      ],
      [
        16,
        32,
        32
      ],
      [
        16,
        32,
        32
      ],
      [
        16,
        32,
        32
      ],
      [
        16,
        32,
        32
      ],
      [
        16,
        32,
        32
      ],
      [
        16,
        32,
        32
      ],
      [
        32,
        16,
        16
      ],
      [
        32,
        16,
        16
      ],
      [
        32,
        16,
        16
      ],
      [
        32,
        16,
        16
      ],
      [
        32,
        16,
        16
      ],
      [
        32,
        16,
        16
      ],
      [
        64,
        8,
        8
      ],
      [
        64,
        8,
        8
      ],
      [
        64,
        8,
        8
      ],
      [
        64,
        8,
        8
      ],
      [
        64,
        8,
        8
      ],
      [
        64,
        8,
        8
      ]
    ],
    "normalization": "batchnorm",
    "predictor_parameters": 376832,
    "residual_scale": 1.0,
    "vectorizer_mode": "spatial_template",
    "vectorizer_parameters": 1884160
  },
  "args": {
    "a_scale": 0.0,
    "a_warmup_steps": 100,
    "alignment_probe": 32,
    "apical_seed": null,
    "augment_train": 1,
    "batch_size": 128,
    "bn_eps": 1e-05,
    "bn_momentum": 0.1,
    "data_dir": "/home/yurenh2/sdrn/data",
    "depth": 20,
    "device": "cuda:0",
    "epochs": 20,
    "eta_A": 0.001,
    "eta_P": 0.01,
    "eval_every": 0,
    "eval_split": "validation",
    "learn_P": 0,
    "loader_seed": 0,
    "lr": 0.03,
    "lr_gamma": 0.1,
    "lr_milestones": "100,150",
    "lr_schedule": "cosine",
    "max_steps": 0,
    "mode": "sdil",
    "momentum": 0.9,
    "normalization": "batchnorm",
    "nuisance_scale": 0.0,
    "out": "results/oral_a_short/sdil_spatial_template_lr0.03.json",
    "output_lr": 0.1,
    "pert_directions": 1,
    "pert_every": 4,
    "pert_sigma": 0.01,
    "perturb_seed": 1000,
    "predictor_warmup_steps": 0,
    "residual_scale": null,
    "seed": 0,
    "split_seed": 2027,
    "train_limit": 10000,
    "use_residual": 1,
    "val_examples": 5000,
    "vectorizer_mode": "spatial_template",
    "warmup_epochs": 0,
    "weight_decay": 0.0001,
    "weight_scale": 1.0,
    "width": 16
  },
  "counters": {
    "apical_warmup_examples": 12688,
    "calibration_event_examples": 62688,
    "causal_scalar_observations": 990,
    "logical_batch_loss_queries": 990,
    "ordinary_examples": 200000,
    "per_example_loss_terms": 125376,
    "perturbation_events": 495,
    "perturbation_forward_examples": 125376,
    "predictor_warmup_examples": 0
  },
  "diagnostics": {
    "early_third_mean": 0.0016261294658761472,
    "innovation_negative_gradient_cosine": [
      0.0026925508864223957,
      0.002747426275163889,
      0.0011170203797519207,
      2.4994893465191126e-05,
      0.002543856156989932,
      0.0006309282034635544,
      0.0015769501915201545,
      -0.0002410386805422604,
      0.002343986416235566,
      0.0025179297663271427,
      0.0012802392011508346,
      0.0014369012787938118,
      0.001235027564689517,
      0.01105789840221405,
      0.01401814166456461,
      0.020832587033510208,
      0.011961638927459717,
      0.017622990533709526,
      0.012485330924391747
    ],
    "normalization_state": "training_batch_stats_without_running_update",
    "raw_negative_gradient_cosine": [
      0.0026925508864223957,
      0.002747426275163889,
      0.0011170203797519207,
      2.4994893465191126e-05,
      0.002543856156989932,
      0.0006309282034635544,
      0.0015769501915201545,
      -0.0002410386805422604,
      0.002343986416235566,
      0.0025179297663271427,
      0.0012802392011508346,
      0.0014369012787938118,
      0.001235027564689517,
      0.01105789840221405,
      0.01401814166456461,
      0.020832587033510208,
      0.011961638927459717,
      0.017622990533709526,
      0.012485330924391747
    ],
    "teaching_negative_gradient_cosine": [
      0.0026925508864223957,
      0.002747426275163889,
      0.0011170203797519207,
      2.4994893465191126e-05,
      0.002543856156989932,
      0.0006309282034635544,
      0.0015769501915201545,
      -0.0002410386805422604,
      0.002343986416235566,
      0.0025179297663271427,
      0.0012802392011508346,
      0.0014369012787938118,
      0.001235027564689517,
      0.01105789840221405,
      0.01401814166456461,
      0.020832587033510208,
      0.011961638927459717,
      0.017622990533709526,
      0.012485330924391747
    ],
    "wall_s": 0.10700726509094238
  },
  "epochs": [
    {
      "calibration": {
        "calibration_mse": 225.0854659847186,
        "prediction_target_cosine": -1.1367990201770532e-05,
        "target_power": 225.08545032376824
      },
      "epoch": 1,
      "lr": 0.03,
      "output_lr": 0.1,
      "step": 79,
      "train_examples": 10000,
      "train_loss": 2.779825601196289
    },
    {
      "calibration": {
        "calibration_mse": 224.05881381941876,
        "prediction_target_cosine": 7.364900904852228e-05,
        "target_power": 224.05880134805392
      },
      "epoch": 2,
      "lr": 0.029815325108927065,
      "output_lr": 0.0993844170297569,
      "step": 158,
      "train_examples": 10000,
      "train_loss": 2.464528601074219
    },
    {
      "calibration": {
        "calibration_mse": 189.00379390978944,
        "prediction_target_cosine": -4.6249213236814e-05,
        "target_power": 189.00376264984192
      },
      "epoch": 3,
      "lr": 0.029265847744427302,
      "output_lr": 0.09755282581475769,
      "step": 237,
      "train_examples": 10000,
      "train_loss": 2.608198607635498
    },
    {
      "calibration": {
        "calibration_mse": 252.0002302787942,
        "prediction_target_cosine": 2.8914304700664197e-05,
        "target_power": 252.00018578426793
      },
      "epoch": 4,
      "lr": 0.028365097862825515,
      "output_lr": 0.09455032620941839,
      "step": 316,
      "train_examples": 10000,
      "train_loss": 2.8782430679321287
    },
    {
      "calibration": {
        "calibration_mse": 162.07468072448088,
        "prediction_target_cosine": -9.906694745418392e-05,
        "target_power": 162.07462500592936
      },
      "epoch": 5,
      "lr": 0.02713525491562421,
      "output_lr": 0.09045084971874738,
      "step": 395,
      "train_examples": 10000,
      "train_loss": 3.600461712646484
    },
    {
      "calibration": {
        "calibration_mse": 339.4159528536642,
        "prediction_target_cosine": -3.538781021778639e-05,
        "target_power": 339.415888260888
      },
      "epoch": 6,
      "lr": 0.02560660171779821,
      "output_lr": 0.08535533905932738,
      "step": 474,
      "train_examples": 10000,
      "train_loss": 4.972540879821778
    },
    {
      "calibration": {
        "calibration_mse": 320.14482944957587,
        "prediction_target_cosine": -5.400665091043201e-05,
        "target_power": 320.14475205922895
      },
      "epoch": 7,
      "lr": 0.023816778784387097,
      "output_lr": 0.07938926261462366,
      "step": 553,
      "train_examples": 10000,
      "train_loss": 6.406736657714844
    },
    {
      "calibration": {
        "calibration_mse": 239.223344233768,
        "prediction_target_cosine": 1.9879424781897673e-05,
        "target_power": 239.2232916822968
      },
      "epoch": 8,
      "lr": 0.0218098574960932,
      "output_lr": 0.07269952498697733,
      "step": 632,
      "train_examples": 10000,
      "train_loss": 6.1722028503417965
    },
    {
      "calibration": {
        "calibration_mse": 463.84863981811895,
        "prediction_target_cosine": -1.875769559546624e-05,
        "target_power": 463.8485612288765
      },
      "epoch": 9,
      "lr": 0.01963525491562421,
      "output_lr": 0.06545084971874737,
      "step": 711,
      "train_examples": 10000,
      "train_loss": 6.8595925750732425
    },
    {
      "calibration": {
        "calibration_mse": 453.42677381142323,
        "prediction_target_cosine": -3.32086027929901e-05,
        "target_power": 453.42670377959377
      },
      "epoch": 10,
      "lr": 0.017346516975603462,
      "output_lr": 0.05782172325201155,
      "step": 790,
      "train_examples": 10000,
      "train_loss": 5.333097433471679
    },
    {
      "calibration": {
        "calibration_mse": 309.16671639810437,
        "prediction_target_cosine": -5.810984922454561e-05,
        "target_power": 309.1665808511818
      },
      "epoch": 11,
      "lr": 0.015,
      "output_lr": 0.05,
      "step": 869,
      "train_examples": 10000,
      "train_loss": 5.10923678855896
    },
    {
      "calibration": {
        "calibration_mse": 159.81345182357174,
        "prediction_target_cosine": 5.906782660362616e-06,
        "target_power": 159.8133521187769
      },
      "epoch": 12,
      "lr": 0.012653483024396533,
      "output_lr": 0.04217827674798845,
      "step": 948,
      "train_examples": 10000,
      "train_loss": 3.1977948154449463
    },
    {
      "calibration": {
        "calibration_mse": 228.8384463015944,
        "prediction_target_cosine": 4.724701258411892e-05,
        "target_power": 228.83836178986923
      },
      "epoch": 13,
      "lr": 0.01036474508437579,
      "output_lr": 0.03454915028125263,
      "step": 1027,
      "train_examples": 10000,
      "train_loss": 2.952687503814697
    },
    {
      "calibration": {
        "calibration_mse": 138.64461772111446,
        "prediction_target_cosine": -4.9393437286832497e-05,
        "target_power": 138.6445097280585
      },
      "epoch": 14,
      "lr": 0.0081901425039068,
      "output_lr": 0.027300475013022664,
      "step": 1106,
      "train_examples": 10000,
      "train_loss": 2.335374161529541
    },
    {
      "calibration": {
        "calibration_mse": 162.95293996288723,
        "prediction_target_cosine": -0.00010373396979351931,
        "target_power": 162.9528389635293
      },
      "epoch": 15,
      "lr": 0.006183221215612904,
      "output_lr": 0.02061073738537635,
      "step": 1185,
      "train_examples": 10000,
      "train_loss": 2.05778305015564
    },
    {
      "calibration": {
        "calibration_mse": 296.1335524799349,
        "prediction_target_cosine": 5.441280496664627e-05,
        "target_power": 296.13348323311357
      },
      "epoch": 16,
      "lr": 0.004393398282201788,
      "output_lr": 0.014644660940672627,
      "step": 1264,
      "train_examples": 10000,
      "train_loss": 2.0930088500976565
    },
    {
      "calibration": {
        "calibration_mse": 151.49618704825153,
        "prediction_target_cosine": -6.845286494395057e-05,
        "target_power": 151.4960966487294
      },
      "epoch": 17,
      "lr": 0.00286474508437579,
      "output_lr": 0.009549150281252633,
      "step": 1343,
      "train_examples": 10000,
      "train_loss": 1.9203613328933715
    },
    {
      "calibration": {
        "calibration_mse": 172.30423948797198,
        "prediction_target_cosine": -4.908389648638358e-05,
        "target_power": 172.3041283291319
      },
      "epoch": 18,
      "lr": 0.001634902137174483,
      "output_lr": 0.005449673790581611,
      "step": 1422,
      "train_examples": 10000,
      "train_loss": 1.8738044273376464
    },
    {
      "calibration": {
        "calibration_mse": 204.4578604910361,
        "prediction_target_cosine": -6.160888634974665e-05,
        "target_power": 204.45776410757202
      },
      "epoch": 19,
      "lr": 0.000734152255572697,
      "output_lr": 0.0024471741852423235,
      "step": 1501,
      "train_examples": 10000,
      "train_loss": 1.8021988306045533
    },
    {
      "calibration": {
        "calibration_mse": 258.4718296773374,
        "prediction_target_cosine": 5.921584174947326e-05,
        "target_power": 258.47175269814323
      },
      "epoch": 20,
      "lr": 0.0001846748910729351,
      "output_lr": 0.0006155829702431171,
      "step": 1580,
      "train_examples": 10000,
      "train_loss": 1.787876174545288
    }
  ],
  "evaluation_protocol": {
    "test_evaluations": 0,
    "test_used_for_selection": false,
    "validation_evaluations": 1
  },
  "final": {
    "accuracy": 0.3596,
    "epoch": 20,
    "evaluation_split": "validation",
    "finite": true,
    "loss": 1.777641943359375,
    "step": 1580
  },
  "hardware": {
    "cuda_device_name": "NVIDIA GeForce GTX 1080",
    "cuda_visible_devices": "5",
    "device": "cuda:0",
    "device_total_memory_bytes": 8507949056,
    "peak_memory_allocated_bytes": 1740965888,
    "peak_memory_reserved_bytes": 2866806784,
    "torch_version": "2.3.1+cu118"
  },
  "protocol_family": "oral_a_cifar_local_resnet_development",
  "provenance": {
    "git_commit": "a1d060e6d96d4ab973d16cd75149dd61899c7c95",
    "git_tracked_dirty": false
  },
  "schema_version": 1,
  "split": {
    "cifar_source_files": [
      {
        "bytes": 31035704,
        "path": "/home/yurenh2/sdrn/data/cifar-10-batches-py/data_batch_1",
        "sha256": "54636561a3ce25bd3e19253c6b0d8538147b0ae398331ac4a2d86c6d987368cd"
      },
      {
        "bytes": 31035320,
        "path": "/home/yurenh2/sdrn/data/cifar-10-batches-py/data_batch_2",
        "sha256": "766b2cef9fbc745cf056b3152224f7cf77163b330ea9a15f9392beb8b89bc5a8"
      },
      {
        "bytes": 31035999,
        "path": "/home/yurenh2/sdrn/data/cifar-10-batches-py/data_batch_3",
        "sha256": "0f00d98ebfb30b3ec0ad19f9756dc2630b89003e10525f5e148445e82aa6a1f9"
      },
      {
        "bytes": 31035696,
        "path": "/home/yurenh2/sdrn/data/cifar-10-batches-py/data_batch_4",
        "sha256": "3f7bb240661948b8f4d53e36ec720d8306f5668bd0071dcb4e6c947f78e9682b"
      },
      {
        "bytes": 31035623,
        "path": "/home/yurenh2/sdrn/data/cifar-10-batches-py/data_batch_5",
        "sha256": "d91802434d8376bbaeeadf58a737e3a1b12ac839077e931237e0dcd43adcb154"
      },
      {
        "bytes": 31035526,
        "path": "/home/yurenh2/sdrn/data/cifar-10-batches-py/test_batch",
        "sha256": "f53d8d457504f7cff4ea9e021afcf0e0ad8e24a91f3fc42091b8adef61157831"
      }
    ],
    "dataset": "cifar10",
    "evaluation_split": "validation",
    "input_layout": "NCHW",
    "input_shape": [
      3,
      32,
      32
    ],
    "loader_seed": 0,
    "normalization_mean": [
      0.49140000343322754,
      0.4821999967098236,
      0.4465000033378601
    ],
    "normalization_std": [
      0.24699999392032623,
      0.2434999942779541,
      0.26159998774528503
    ],
    "split_from_training_only": true,
    "split_seed": 2027,
    "test_examples": 10000,
    "train_examples": 10000,
    "training_augmentation": "random_crop_32_padding_4_zero_then_horizontal_flip_p0.5",
    "validation_class_counts": {
      "0": 500,
      "1": 500,
      "2": 500,
      "3": 500,
      "4": 500,
      "5": 500,
      "6": 500,
      "7": 500,
      "8": 500,
      "9": 500
    },
    "validation_examples": 5000,
    "validation_index_sha256": "8328b206a97c420e49e54e3eca4abe3274c4756b084355784ea3fb8059e4515b"
  },
  "timing": {
    "apical_warmup_wall_s": 5.937381267547607,
    "evaluation_wall_s": 0.39031529426574707,
    "predictor_warmup_wall_s": 0.0,
    "timing_excludes_data_loading_hashing_and_model_construction": true,
    "total_timed_wall_s": 95.62685227394104,
    "train_wall_s": 89.29021406173706
  },
  "work": {
    "apical_macs_per_example": 1884160,
    "causal_scalar_observations": 990,
    "components": {
      "apical_projection_macs": 400738222080,
      "apical_regression_macs": 118114222080,
      "bp_reverse_macs_estimate": 0,
      "local_weight_correlation_macs": 8110208000000,
      "ordinary_forward_macs": 8110208000000,
      "perturbation_forward_macs": 5084127191040,
      "warmup_clean_forward_macs": 514511595520
    },
    "definition": "multiply-accumulates in conv/linear maps; one local weight correlation equals one forward-weight MAC count; BP reverse is estimated as one weight-gradient plus one activation-gradient convolution per forward convolution; elementwise nonlinearities and optimizer arithmetic excluded",
    "forward_macs_per_example": 40551040,
    "logical_batch_loss_queries": 990,
    "per_example_cross_entropy_terms": 125376,
    "total_clean_forward_examples": 212688,
    "total_forward_equivalent_examples": 338064,
    "total_macs_estimate": 22337907230720
  }
}