{
  "architecture": {
    "adaptive_apical_parameters": 0,
    "adaptive_feedback_parameters": 267904,
    "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_feedback_parameters": 0,
    "fixed_traffic_coefficients": 0,
    "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": 0,
    "residual_scale": 1.0,
    "vectorizer_mode": null,
    "vectorizer_parameters": 0
  },
  "args": {
    "a_scale": 1.0,
    "a_warmup_steps": 0,
    "alignment_probe": 32,
    "apical_calibration_mode": "unit_targets",
    "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",
    "epochs": 20,
    "eta_A": 0.01,
    "eta_P": 0.01,
    "eval_every": 0,
    "eval_split": "validation",
    "learn_P": 0,
    "loader_seed": 0,
    "lr": 0.1,
    "lr_gamma": 0.1,
    "lr_milestones": "100,150",
    "lr_schedule": "cosine",
    "max_steps": 0,
    "mirror_batch_size": 1,
    "mirror_eta": 0.1,
    "mirror_every": 16,
    "mirror_noise_std": 1.0,
    "mirror_seed": 3000,
    "mirror_warmup_steps": 20,
    "mode": "rrm",
    "momentum": 0.9,
    "normalization": "batchnorm",
    "nuisance_scale": 0.0,
    "out": "results/residual_mirror_short/rrm_lr0.1.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
  },
  "calibration_metric_space": "local_parent_child_response_residual",
  "counters": {
    "apical_warmup_examples": 0,
    "calibration_event_examples": 0,
    "causal_scalar_observations": 0,
    "logical_batch_loss_queries": 0,
    "mirror_conv_examples": 119,
    "mirror_events": 119,
    "mirror_readout_examples": 121856,
    "ordinary_examples": 200000,
    "per_example_loss_terms": 0,
    "perturbation_events": 0,
    "perturbation_forward_examples": 0,
    "predictor_warmup_examples": 0
  },
  "diagnostics": {
    "early_third_mean": 0.9991278549035391,
    "feedback_forward_cosine": [
      0.9999154806137085,
      0.9999306797981262,
      0.9999545812606812,
      0.9999644160270691,
      0.9999690055847168,
      0.9999793767929077,
      0.9999752044677734,
      0.9999589920043945,
      0.9999452233314514,
      0.9999603033065796,
      0.9999622702598572,
      0.9999719858169556,
      0.9999478459358215,
      0.9997169375419617,
      0.9997090101242065,
      0.9997721910476685,
      0.9997538328170776,
      0.9998348951339722,
      0.9999200105667114
    ],
    "feedback_forward_norm_ratio": [
      0.9980469942092896,
      0.9989184737205505,
      0.9994211196899414,
      0.9998089671134949,
      0.9999980926513672,
      1.0001499652862549,
      1.0000603199005127,
      1.0000534057617188,
      0.9997336864471436,
      1.0000892877578735,
      1.000082015991211,
      1.000295639038086,
      1.0003461837768555,
      1.0006914138793945,
      1.0006706714630127,
      1.0003938674926758,
      1.0005216598510742,
      1.0005003213882446,
      0.9909704923629761
    ],
    "innovation_negative_gradient_cosine": [
      0.9988824725151062,
      0.9990805387496948,
      0.9991225004196167,
      0.9991109371185303,
      0.9992891550064087,
      0.9992815256118774,
      0.9993298649787903,
      0.9993308782577515,
      0.999434232711792,
      0.9993816614151001,
      0.9995160102844238,
      0.9995108842849731,
      0.9995355010032654,
      0.9993031024932861,
      0.99957275390625,
      0.9993000030517578,
      0.9997174739837646,
      0.9995701313018799,
      0.9999008178710938
    ],
    "normalization_state": "training_batch_stats_without_running_update",
    "raw_negative_gradient_cosine": [
      0.9988824725151062,
      0.9990805387496948,
      0.9991225004196167,
      0.9991109371185303,
      0.9992891550064087,
      0.9992815256118774,
      0.9993298649787903,
      0.9993308782577515,
      0.999434232711792,
      0.9993816614151001,
      0.9995160102844238,
      0.9995108842849731,
      0.9995355010032654,
      0.9993031024932861,
      0.99957275390625,
      0.9993000030517578,
      0.9997174739837646,
      0.9995701313018799,
      0.9999008178710938
    ],
    "teaching_negative_gradient_cosine": [
      0.9988824725151062,
      0.9990805387496948,
      0.9991225004196167,
      0.9991109371185303,
      0.9992891550064087,
      0.9992815256118774,
      0.9993298649787903,
      0.9993308782577515,
      0.999434232711792,
      0.9993816614151001,
      0.9995160102844238,
      0.9995108842849731,
      0.9995355010032654,
      0.9993031024932861,
      0.99957275390625,
      0.9993000030517578,
      0.9997174739837646,
      0.9995701313018799,
      0.9999008178710938
    ],
    "wall_s": 0.11099910736083984
  },
  "epochs": [
    {
      "epoch": 1,
      "lr": 0.1,
      "mirror": {
        "conv_batch_size": 1.0,
        "mirror_response_residual_fraction": 0.2601846737040344,
        "mirror_response_residual_rms": 0.3535099094948956,
        "mirror_update_rms": 0.0794724507423343,
        "readout_batch_size": 1024.0
      },
      "output_lr": 0.1,
      "step": 79,
      "train_examples": 10000,
      "train_loss": 2.6401633117675782
    },
    {
      "epoch": 2,
      "lr": 0.0993844170297569,
      "mirror": {
        "conv_batch_size": 1.0,
        "mirror_response_residual_fraction": 0.2320270030739035,
        "mirror_response_residual_rms": 0.3178695360002074,
        "mirror_update_rms": 0.06566347399745359,
        "readout_batch_size": 1024.0
      },
      "output_lr": 0.0993844170297569,
      "step": 158,
      "train_examples": 10000,
      "train_loss": 2.428320087814331
    },
    {
      "epoch": 3,
      "lr": 0.09755282581475769,
      "mirror": {
        "conv_batch_size": 1.0,
        "mirror_response_residual_fraction": 0.1902612302737704,
        "mirror_response_residual_rms": 0.261214385527155,
        "mirror_update_rms": 0.052451028617311414,
        "readout_batch_size": 1024.0
      },
      "output_lr": 0.09755282581475769,
      "step": 237,
      "train_examples": 10000,
      "train_loss": 2.5726836837768556
    },
    {
      "epoch": 4,
      "lr": 0.09455032620941839,
      "mirror": {
        "conv_batch_size": 1.0,
        "mirror_response_residual_fraction": 0.17018607732039362,
        "mirror_response_residual_rms": 0.23495350486028804,
        "mirror_update_rms": 0.042288586572611774,
        "readout_batch_size": 1024.0
      },
      "output_lr": 0.09455032620941839,
      "step": 316,
      "train_examples": 10000,
      "train_loss": 2.528295587730408
    },
    {
      "epoch": 5,
      "lr": 0.09045084971874738,
      "mirror": {
        "conv_batch_size": 1.0,
        "mirror_response_residual_fraction": 0.15231621565559716,
        "mirror_response_residual_rms": 0.21180007147453325,
        "mirror_update_rms": 0.03544802639828195,
        "readout_batch_size": 1024.0
      },
      "output_lr": 0.09045084971874738,
      "step": 395,
      "train_examples": 10000,
      "train_loss": 2.218377275466919
    },
    {
      "epoch": 6,
      "lr": 0.08535533905932738,
      "mirror": {
        "conv_batch_size": 1.0,
        "mirror_response_residual_fraction": 0.14108329109632886,
        "mirror_response_residual_rms": 0.19726323418619918,
        "mirror_update_rms": 0.02993521390942853,
        "readout_batch_size": 1024.0
      },
      "output_lr": 0.08535533905932738,
      "step": 474,
      "train_examples": 10000,
      "train_loss": 2.0937036754608154
    },
    {
      "epoch": 7,
      "lr": 0.07938926261462366,
      "mirror": {
        "conv_batch_size": 1.0,
        "mirror_response_residual_fraction": 0.13834582525482075,
        "mirror_response_residual_rms": 0.19534085520608319,
        "mirror_update_rms": 0.025985602855510263,
        "readout_batch_size": 1024.0
      },
      "output_lr": 0.07938926261462366,
      "step": 553,
      "train_examples": 10000,
      "train_loss": 2.218831299972534
    },
    {
      "epoch": 8,
      "lr": 0.07269952498697733,
      "mirror": {
        "conv_batch_size": 1.0,
        "mirror_response_residual_fraction": 0.13345085752752778,
        "mirror_response_residual_rms": 0.1909307264239541,
        "mirror_update_rms": 0.0240043080261321,
        "readout_batch_size": 1024.0
      },
      "output_lr": 0.07269952498697733,
      "step": 632,
      "train_examples": 10000,
      "train_loss": 2.274434162902832
    },
    {
      "epoch": 9,
      "lr": 0.06545084971874737,
      "mirror": {
        "conv_batch_size": 1.0,
        "mirror_response_residual_fraction": 0.12035600897902161,
        "mirror_response_residual_rms": 0.17304416237519363,
        "mirror_update_rms": 0.02157844918317908,
        "readout_batch_size": 1024.0
      },
      "output_lr": 0.06545084971874737,
      "step": 711,
      "train_examples": 10000,
      "train_loss": 1.8082092502593994
    },
    {
      "epoch": 10,
      "lr": 0.05782172325201155,
      "mirror": {
        "conv_batch_size": 1.0,
        "mirror_response_residual_fraction": 0.1096965221826743,
        "mirror_response_residual_rms": 0.15870117165633163,
        "mirror_update_rms": 0.01960604709124913,
        "readout_batch_size": 1024.0
      },
      "output_lr": 0.05782172325201155,
      "step": 790,
      "train_examples": 10000,
      "train_loss": 1.6496932834625244
    },
    {
      "epoch": 11,
      "lr": 0.05,
      "mirror": {
        "conv_batch_size": 1.0,
        "mirror_response_residual_fraction": 0.10060992128674659,
        "mirror_response_residual_rms": 0.14640697910423445,
        "mirror_update_rms": 0.017701212220667454,
        "readout_batch_size": 1024.0
      },
      "output_lr": 0.05,
      "step": 869,
      "train_examples": 10000,
      "train_loss": 1.5478480306625366
    },
    {
      "epoch": 12,
      "lr": 0.04217827674798845,
      "mirror": {
        "conv_batch_size": 1.0,
        "mirror_response_residual_fraction": 0.09085091944110543,
        "mirror_response_residual_rms": 0.13333263756453764,
        "mirror_update_rms": 0.01605992521153775,
        "readout_batch_size": 1024.0
      },
      "output_lr": 0.04217827674798845,
      "step": 948,
      "train_examples": 10000,
      "train_loss": 1.3823965042114257
    },
    {
      "epoch": 13,
      "lr": 0.03454915028125263,
      "mirror": {
        "conv_batch_size": 1.0,
        "mirror_response_residual_fraction": 0.07902996300142369,
        "mirror_response_residual_rms": 0.11615116335525824,
        "mirror_update_rms": 0.014287858257776787,
        "readout_batch_size": 1024.0
      },
      "output_lr": 0.03454915028125263,
      "step": 1027,
      "train_examples": 10000,
      "train_loss": 1.34172273979187
    },
    {
      "epoch": 14,
      "lr": 0.027300475013022664,
      "mirror": {
        "conv_batch_size": 1.0,
        "mirror_response_residual_fraction": 0.06905828831434452,
        "mirror_response_residual_rms": 0.10180783214913795,
        "mirror_update_rms": 0.012587746597357776,
        "readout_batch_size": 1024.0
      },
      "output_lr": 0.027300475013022664,
      "step": 1106,
      "train_examples": 10000,
      "train_loss": 1.2290157025337218
    },
    {
      "epoch": 15,
      "lr": 0.02061073738537635,
      "mirror": {
        "conv_batch_size": 1.0,
        "mirror_response_residual_fraction": 0.05757729123066844,
        "mirror_response_residual_rms": 0.08502710898907911,
        "mirror_update_rms": 0.01099315299825323,
        "readout_batch_size": 1024.0
      },
      "output_lr": 0.02061073738537635,
      "step": 1185,
      "train_examples": 10000,
      "train_loss": 1.1287344594955444
    },
    {
      "epoch": 16,
      "lr": 0.014644660940672627,
      "mirror": {
        "conv_batch_size": 1.0,
        "mirror_response_residual_fraction": 0.04826044807946187,
        "mirror_response_residual_rms": 0.07131268925507378,
        "mirror_update_rms": 0.009481739617492971,
        "readout_batch_size": 1024.0
      },
      "output_lr": 0.014644660940672627,
      "step": 1264,
      "train_examples": 10000,
      "train_loss": 1.0884003732681276
    },
    {
      "epoch": 17,
      "lr": 0.009549150281252633,
      "mirror": {
        "conv_batch_size": 1.0,
        "mirror_response_residual_fraction": 0.03974752585157935,
        "mirror_response_residual_rms": 0.05869449700782102,
        "mirror_update_rms": 0.008195443513961818,
        "readout_batch_size": 1024.0
      },
      "output_lr": 0.009549150281252633,
      "step": 1343,
      "train_examples": 10000,
      "train_loss": 1.0395743354797364
    },
    {
      "epoch": 18,
      "lr": 0.005449673790581611,
      "mirror": {
        "conv_batch_size": 1.0,
        "mirror_response_residual_fraction": 0.0302168111419576,
        "mirror_response_residual_rms": 0.044779826338416255,
        "mirror_update_rms": 0.006658613012222196,
        "readout_batch_size": 1024.0
      },
      "output_lr": 0.005449673790581611,
      "step": 1422,
      "train_examples": 10000,
      "train_loss": 0.9841103965759277
    },
    {
      "epoch": 19,
      "lr": 0.0024471741852423235,
      "mirror": {
        "conv_batch_size": 1.0,
        "mirror_response_residual_fraction": 0.02212630175876517,
        "mirror_response_residual_rms": 0.03275805770984623,
        "mirror_update_rms": 0.00537444741444945,
        "readout_batch_size": 1024.0
      },
      "output_lr": 0.0024471741852423235,
      "step": 1501,
      "train_examples": 10000,
      "train_loss": 0.9521100430488586
    },
    {
      "epoch": 20,
      "lr": 0.0006155829702431171,
      "mirror": {
        "conv_batch_size": 1.0,
        "mirror_response_residual_fraction": 0.015118729649761752,
        "mirror_response_residual_rms": 0.022381774527825898,
        "mirror_update_rms": 0.004188984685734049,
        "readout_batch_size": 1024.0
      },
      "output_lr": 0.0006155829702431171,
      "step": 1580,
      "train_examples": 10000,
      "train_loss": 0.938697664642334
    }
  ],
  "evaluation_protocol": {
    "test_evaluations": 0,
    "test_used_for_selection": false,
    "validation_evaluations": 1
  },
  "final": {
    "accuracy": 0.6508,
    "epoch": 20,
    "evaluation_split": "validation",
    "finite": true,
    "loss": 1.01295,
    "step": 1580
  },
  "hardware": {
    "cuda_device_name": "NVIDIA GeForce GTX 1080",
    "cuda_visible_devices": "7",
    "device": "cuda",
    "device_total_memory_bytes": 8507949056,
    "peak_memory_allocated_bytes": 1730831872,
    "peak_memory_reserved_bytes": 2501902336,
    "torch_version": "2.3.1+cu118"
  },
  "mirror_warmup": {
    "first": {
      "conv_batch_size": 1,
      "mirror_response_residual_fraction": 1.4124661316345333,
      "mirror_response_residual_rms": 1.90927576279183,
      "mirror_update_rms": 0.25520769706611623,
      "readout_batch_size": 1024
    },
    "last": {
      "conv_batch_size": 1,
      "mirror_response_residual_fraction": 0.26271250078043745,
      "mirror_response_residual_rms": 0.3557588547676734,
      "mirror_update_rms": 0.0907495165488641,
      "readout_batch_size": 1024
    },
    "mean": {
      "conv_batch_size": 1.0,
      "mirror_response_residual_fraction": 0.6719071790962137,
      "mirror_response_residual_rms": 0.9100978838204868,
      "mirror_update_rms": 0.15685383656211757,
      "readout_batch_size": 1024.0
    },
    "steps": 20
  },
  "protocol_family": "oral_a_cifar_local_resnet_development",
  "provenance": {
    "git_commit": "6206483780d75ecff61c396b7f1bf7e23e5b1ca5",
    "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": 0.0,
    "evaluation_wall_s": 0.36290621757507324,
    "mirror_warmup_wall_s": 0.630126953125,
    "predictor_warmup_wall_s": 0.0,
    "timing_excludes_data_loading_hashing_and_model_construction": true,
    "total_timed_wall_s": 93.04161047935486,
    "train_wall_s": 92.04554319381714
  },
  "work": {
    "apical_macs_per_example": 40108672,
    "causal_scalar_observations": 0,
    "components": {
      "apical_projection_macs": 8021734400000,
      "apical_regression_macs": 0,
      "bp_reverse_macs_estimate": 0,
      "local_weight_correlation_macs": 8110208000000,
      "mirror_feedback_prediction_macs": 4850843648,
      "mirror_local_correlation_macs": 4850843648,
      "mirror_response_macs": 4850843648,
      "ordinary_forward_macs": 8110208000000,
      "perturbation_forward_macs": 0,
      "warmup_clean_forward_macs": 0
    },
    "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": 0,
    "mirror_probe_examples": 119,
    "mirror_readout_probe_examples": 121856,
    "per_example_cross_entropy_terms": 0,
    "total_clean_forward_examples": 200000,
    "total_forward_equivalent_examples": 200000,
    "total_macs_estimate": 24256702930944
  }
}