{ "architecture": { "base_width": 16, "depth": 32, "family": "CIFAR 6n+2 ResNet, option-A shortcuts", "forward_parameter_count": 464154, "normalization": "batchnorm", "residual_scale": 1.0 }, "args": { "augment_train": 1, "batch_size": 128, "data_dir": "/orion/u/oscarwan/scratch/sdil-data", "depth": 32, "device": "cuda", "dp_alpha": 0.0, "dp_beta": 0.1, "dp_inference_passes": 16, "ep_beta": 0.5, "ep_dt": 0.5, "ep_free_steps": 20, "ep_nudge_steps": 4, "epochs": 100, "eval_every": 1, "eval_split": "validation", "feedback_seed": 1729, "ff_score_from_layer": 1, "ff_threshold": 2.0, "loader_seed": 0, "lr": 0.003, "lr_gamma": 0.1, "lr_milestones": "60,90", "lr_schedule": "pepita", "max_steps": 0, "method": "pepita", "momentum": 0.9, "out": "/orion/u/oscarwan/scratch/sdil-work/sdil-a6000-full-rerun/results/resnet_crossover/r2/resnet-r2-pepita-d32-lr0p003.json", "output_lr": 0.003, "pepita_projection_scale": 0.05, "seed": 0, "split_seed": 2027, "train_limit": 0, "val_examples": 5000, "weight_decay": 0.0001, "width": 16 }, "epochs": [ { "epoch": 1, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.2907867482503255, "train_seconds": 9.360012769699097, "validation": { "accuracy": 0.181, "examples": 5000, "loss": 2.190810546875 }, "validation_seconds": 0.9673621654510498 }, { "epoch": 2, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.1893652502695717, "train_seconds": 8.754016399383545, "validation": { "accuracy": 0.1966, "examples": 5000, "loss": 2.148013818359375 }, "validation_seconds": 0.169417142868042 }, { "epoch": 3, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.161732895490858, "train_seconds": 8.748339891433716, "validation": { "accuracy": 0.204, "examples": 5000, "loss": 2.12567041015625 }, "validation_seconds": 0.1697547435760498 }, { "epoch": 4, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.146513925764296, "train_seconds": 8.752492904663086, "validation": { "accuracy": 0.2226, "examples": 5000, "loss": 2.121567333984375 }, "validation_seconds": 0.17016267776489258 }, { "epoch": 5, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.1360204981486004, "train_seconds": 8.763130903244019, "validation": { "accuracy": 0.235, "examples": 5000, "loss": 2.091363623046875 }, "validation_seconds": 0.1700141429901123 }, { "epoch": 6, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.1298021606021456, "train_seconds": 8.769134998321533, "validation": { "accuracy": 0.2434, "examples": 5000, "loss": 2.0821609375 }, "validation_seconds": 0.17003583908081055 }, { "epoch": 7, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.123042589187622, "train_seconds": 8.775540590286255, "validation": { "accuracy": 0.2244, "examples": 5000, "loss": 2.086983447265625 }, "validation_seconds": 0.16979694366455078 }, { "epoch": 8, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.118494323179457, "train_seconds": 8.769744873046875, "validation": { "accuracy": 0.2466, "examples": 5000, "loss": 2.068868017578125 }, "validation_seconds": 0.17055869102478027 }, { "epoch": 9, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.1129601199256047, "train_seconds": 8.778553485870361, "validation": { "accuracy": 0.2376, "examples": 5000, "loss": 2.071684130859375 }, "validation_seconds": 0.17051935195922852 }, { "epoch": 10, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.1110865395439995, "train_seconds": 8.786123275756836, "validation": { "accuracy": 0.2464, "examples": 5000, "loss": 2.063233203125 }, "validation_seconds": 0.17072343826293945 }, { "epoch": 11, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.1061982627020943, "train_seconds": 8.78468656539917, "validation": { "accuracy": 0.2348, "examples": 5000, "loss": 2.0644167724609375 }, "validation_seconds": 0.17047452926635742 }, { "epoch": 12, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.107127010981242, "train_seconds": 8.783576250076294, "validation": { "accuracy": 0.25, "examples": 5000, "loss": 2.0543634765625 }, "validation_seconds": 0.17060136795043945 }, { "epoch": 13, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.1037176490783693, "train_seconds": 8.785626411437988, "validation": { "accuracy": 0.231, "examples": 5000, "loss": 2.064813330078125 }, "validation_seconds": 0.17042326927185059 }, { "epoch": 14, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.099651092529297, "train_seconds": 8.749484539031982, "validation": { "accuracy": 0.2542, "examples": 5000, "loss": 2.048787548828125 }, "validation_seconds": 0.17104148864746094 }, { "epoch": 15, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.1024867318047415, "train_seconds": 8.785282135009766, "validation": { "accuracy": 0.263, "examples": 5000, "loss": 2.0486366455078127 }, "validation_seconds": 0.17087507247924805 }, { "epoch": 16, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.096451896031698, "train_seconds": 8.786890745162964, "validation": { "accuracy": 0.2558, "examples": 5000, "loss": 2.0489378662109377 }, "validation_seconds": 0.17048311233520508 }, { "epoch": 17, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0968072445339625, "train_seconds": 8.784305572509766, "validation": { "accuracy": 0.2578, "examples": 5000, "loss": 2.034791357421875 }, "validation_seconds": 0.17049431800842285 }, { "epoch": 18, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0995056786007353, "train_seconds": 8.785536050796509, "validation": { "accuracy": 0.2574, "examples": 5000, "loss": 2.04157490234375 }, "validation_seconds": 0.17114782333374023 }, { "epoch": 19, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0956084409925673, "train_seconds": 8.790813207626343, "validation": { "accuracy": 0.256, "examples": 5000, "loss": 2.0414225830078125 }, "validation_seconds": 0.17094159126281738 }, { "epoch": 20, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.093731904051039, "train_seconds": 8.78307318687439, "validation": { "accuracy": 0.2562, "examples": 5000, "loss": 2.0405240966796874 }, "validation_seconds": 0.17032456398010254 }, { "epoch": 21, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0907077075534395, "train_seconds": 8.78516411781311, "validation": { "accuracy": 0.252, "examples": 5000, "loss": 2.054002294921875 }, "validation_seconds": 0.17098021507263184 }, { "epoch": 22, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0961777904086643, "train_seconds": 8.780037879943848, "validation": { "accuracy": 0.264, "examples": 5000, "loss": 2.04380771484375 }, "validation_seconds": 0.17042255401611328 }, { "epoch": 23, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.093424896197849, "train_seconds": 8.782354831695557, "validation": { "accuracy": 0.2602, "examples": 5000, "loss": 2.0328253173828124 }, "validation_seconds": 0.17090868949890137 }, { "epoch": 24, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.089472702874078, "train_seconds": 8.781430721282959, "validation": { "accuracy": 0.2324, "examples": 5000, "loss": 2.0593703125 }, "validation_seconds": 0.17047786712646484 }, { "epoch": 25, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0928057125091555, "train_seconds": 8.784738302230835, "validation": { "accuracy": 0.2614, "examples": 5000, "loss": 2.0374564697265627 }, "validation_seconds": 0.17136812210083008 }, { "epoch": 26, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0903255541059704, "train_seconds": 8.7819344997406, "validation": { "accuracy": 0.2682, "examples": 5000, "loss": 2.030443994140625 }, "validation_seconds": 0.17069506645202637 }, { "epoch": 27, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0905855367448596, "train_seconds": 8.782570362091064, "validation": { "accuracy": 0.2552, "examples": 5000, "loss": 2.038783154296875 }, "validation_seconds": 0.17075705528259277 }, { "epoch": 28, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0896595900429618, "train_seconds": 8.781562089920044, "validation": { "accuracy": 0.24, "examples": 5000, "loss": 2.05001591796875 }, "validation_seconds": 0.17116379737854004 }, { "epoch": 29, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.088815062628852, "train_seconds": 8.783803462982178, "validation": { "accuracy": 0.2532, "examples": 5000, "loss": 2.038257421875 }, "validation_seconds": 0.17058849334716797 }, { "epoch": 30, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0916393721262616, "train_seconds": 8.789171695709229, "validation": { "accuracy": 0.2564, "examples": 5000, "loss": 2.0344188232421874 }, "validation_seconds": 0.1707472801208496 }, { "epoch": 31, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0937151625739205, "train_seconds": 8.783771753311157, "validation": { "accuracy": 0.264, "examples": 5000, "loss": 2.0369871826171875 }, "validation_seconds": 0.1707751750946045 }, { "epoch": 32, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0915020135243734, "train_seconds": 8.784525871276855, "validation": { "accuracy": 0.2614, "examples": 5000, "loss": 2.0325482177734373 }, "validation_seconds": 0.1706712245941162 }, { "epoch": 33, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.089619413333469, "train_seconds": 8.782733917236328, "validation": { "accuracy": 0.2568, "examples": 5000, "loss": 2.0395154541015623 }, "validation_seconds": 0.1710975170135498 }, { "epoch": 34, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.09081546113756, "train_seconds": 8.780365705490112, "validation": { "accuracy": 0.248, "examples": 5000, "loss": 2.03733984375 }, "validation_seconds": 0.17079472541809082 }, { "epoch": 35, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.088437536536323, "train_seconds": 8.78589916229248, "validation": { "accuracy": 0.258, "examples": 5000, "loss": 2.0394189697265626 }, "validation_seconds": 0.17055249214172363 }, { "epoch": 36, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0895603517744274, "train_seconds": 8.778956413269043, "validation": { "accuracy": 0.2592, "examples": 5000, "loss": 2.0462224609375 }, "validation_seconds": 0.17077350616455078 }, { "epoch": 37, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.091470686488681, "train_seconds": 8.784783363342285, "validation": { "accuracy": 0.2478, "examples": 5000, "loss": 2.0410172119140624 }, "validation_seconds": 0.1713721752166748 }, { "epoch": 38, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.088535167863634, "train_seconds": 8.789288759231567, "validation": { "accuracy": 0.256, "examples": 5000, "loss": 2.0440474853515624 }, "validation_seconds": 0.17089271545410156 }, { "epoch": 39, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.088415269978841, "train_seconds": 8.792225360870361, "validation": { "accuracy": 0.2636, "examples": 5000, "loss": 2.0371565673828127 }, "validation_seconds": 0.17076539993286133 }, { "epoch": 40, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0889872140248618, "train_seconds": 8.789459705352783, "validation": { "accuracy": 0.2554, "examples": 5000, "loss": 2.0380864013671873 }, "validation_seconds": 0.17058658599853516 }, { "epoch": 41, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0890390762329103, "train_seconds": 8.786790609359741, "validation": { "accuracy": 0.231, "examples": 5000, "loss": 2.0498098876953126 }, "validation_seconds": 0.17036080360412598 }, { "epoch": 42, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.08899807349311, "train_seconds": 8.78852891921997, "validation": { "accuracy": 0.2488, "examples": 5000, "loss": 2.037922509765625 }, "validation_seconds": 0.17090797424316406 }, { "epoch": 43, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0921546830071343, "train_seconds": 8.785106897354126, "validation": { "accuracy": 0.2548, "examples": 5000, "loss": 2.02869716796875 }, "validation_seconds": 0.17080354690551758 }, { "epoch": 44, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.092005069584317, "train_seconds": 8.787235975265503, "validation": { "accuracy": 0.2488, "examples": 5000, "loss": 2.049959033203125 }, "validation_seconds": 0.17099785804748535 }, { "epoch": 45, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0909548716227215, "train_seconds": 8.785946130752563, "validation": { "accuracy": 0.2476, "examples": 5000, "loss": 2.048082666015625 }, "validation_seconds": 0.1708078384399414 }, { "epoch": 46, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0915795023176407, "train_seconds": 8.780594825744629, "validation": { "accuracy": 0.2412, "examples": 5000, "loss": 2.0391725341796874 }, "validation_seconds": 0.17049098014831543 }, { "epoch": 47, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0890224663416546, "train_seconds": 8.786250352859497, "validation": { "accuracy": 0.2608, "examples": 5000, "loss": 2.0262034912109375 }, "validation_seconds": 0.17060184478759766 }, { "epoch": 48, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0947676996866864, "train_seconds": 8.783189058303833, "validation": { "accuracy": 0.2566, "examples": 5000, "loss": 2.039441650390625 }, "validation_seconds": 0.17084026336669922 }, { "epoch": 49, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.095011107550727, "train_seconds": 8.784818172454834, "validation": { "accuracy": 0.2562, "examples": 5000, "loss": 2.0363927734375 }, "validation_seconds": 0.17116999626159668 }, { "epoch": 50, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0891846113416883, "train_seconds": 8.784857749938965, "validation": { "accuracy": 0.2578, "examples": 5000, "loss": 2.0316771484375 }, "validation_seconds": 0.17063379287719727 }, { "epoch": 51, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0929476915359495, "train_seconds": 8.783042430877686, "validation": { "accuracy": 0.2558, "examples": 5000, "loss": 2.035330224609375 }, "validation_seconds": 0.17069602012634277 }, { "epoch": 52, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.088564098273383, "train_seconds": 8.782349824905396, "validation": { "accuracy": 0.2532, "examples": 5000, "loss": 2.04925380859375 }, "validation_seconds": 0.17074131965637207 }, { "epoch": 53, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0938075734880237, "train_seconds": 8.777364492416382, "validation": { "accuracy": 0.2616, "examples": 5000, "loss": 2.029446142578125 }, "validation_seconds": 0.1706075668334961 }, { "epoch": 54, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0936699249479505, "train_seconds": 8.779646873474121, "validation": { "accuracy": 0.2338, "examples": 5000, "loss": 2.0565670166015626 }, "validation_seconds": 0.1709299087524414 }, { "epoch": 55, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.094998370022244, "train_seconds": 8.779564380645752, "validation": { "accuracy": 0.2512, "examples": 5000, "loss": 2.0400283447265624 }, "validation_seconds": 0.1704554557800293 }, { "epoch": 56, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0949601395924886, "train_seconds": 8.783570289611816, "validation": { "accuracy": 0.2318, "examples": 5000, "loss": 2.05338671875 }, "validation_seconds": 0.17037272453308105 }, { "epoch": 57, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0950200890858968, "train_seconds": 8.780302047729492, "validation": { "accuracy": 0.238, "examples": 5000, "loss": 2.0578427001953123 }, "validation_seconds": 0.17057228088378906 }, { "epoch": 58, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0979971645143296, "train_seconds": 8.781599044799805, "validation": { "accuracy": 0.2442, "examples": 5000, "loss": 2.053140576171875 }, "validation_seconds": 0.17058086395263672 }, { "epoch": 59, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0933812481774225, "train_seconds": 8.783581972122192, "validation": { "accuracy": 0.2446, "examples": 5000, "loss": 2.039527001953125 }, "validation_seconds": 0.17032074928283691 }, { "epoch": 60, "lr": 0.003, "output_lr": 0.003, "train_examples": 45000, "train_loss": 2.0970389923095705, "train_seconds": 8.782969951629639, "validation": { "accuracy": 0.2502, "examples": 5000, "loss": 2.035926953125 }, "validation_seconds": 0.17087292671203613 }, { "epoch": 61, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.082931641260783, "train_seconds": 8.786428213119507, "validation": { "accuracy": 0.2638, "examples": 5000, "loss": 2.027929052734375 }, "validation_seconds": 0.17071866989135742 }, { "epoch": 62, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.0817930608537463, "train_seconds": 8.783912420272827, "validation": { "accuracy": 0.261, "examples": 5000, "loss": 2.02908349609375 }, "validation_seconds": 0.17039942741394043 }, { "epoch": 63, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.082688669925266, "train_seconds": 8.776290655136108, "validation": { "accuracy": 0.2552, "examples": 5000, "loss": 2.0279300537109375 }, "validation_seconds": 0.17133569717407227 }, { "epoch": 64, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.079216177495321, "train_seconds": 8.777651309967041, "validation": { "accuracy": 0.256, "examples": 5000, "loss": 2.03120791015625 }, "validation_seconds": 0.17051076889038086 }, { "epoch": 65, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.080554505962796, "train_seconds": 8.78281545639038, "validation": { "accuracy": 0.2608, "examples": 5000, "loss": 2.02845283203125 }, "validation_seconds": 0.1705918312072754 }, { "epoch": 66, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.082145662413703, "train_seconds": 8.77703309059143, "validation": { "accuracy": 0.2616, "examples": 5000, "loss": 2.034113623046875 }, "validation_seconds": 0.17079496383666992 }, { "epoch": 67, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.085471007792155, "train_seconds": 8.780488729476929, "validation": { "accuracy": 0.2596, "examples": 5000, "loss": 2.02885205078125 }, "validation_seconds": 0.17075657844543457 }, { "epoch": 68, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.082207935248481, "train_seconds": 8.777437210083008, "validation": { "accuracy": 0.2646, "examples": 5000, "loss": 2.02878359375 }, "validation_seconds": 0.17030858993530273 }, { "epoch": 69, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.0829909870147705, "train_seconds": 8.775349855422974, "validation": { "accuracy": 0.2608, "examples": 5000, "loss": 2.0296310302734377 }, "validation_seconds": 0.17086458206176758 }, { "epoch": 70, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.087993453131782, "train_seconds": 8.779688596725464, "validation": { "accuracy": 0.2556, "examples": 5000, "loss": 2.0325456298828124 }, "validation_seconds": 0.1709308624267578 }, { "epoch": 71, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.0826953997717963, "train_seconds": 8.779837608337402, "validation": { "accuracy": 0.2628, "examples": 5000, "loss": 2.03113203125 }, "validation_seconds": 0.17030596733093262 }, { "epoch": 72, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.0797708891126843, "train_seconds": 8.777223110198975, "validation": { "accuracy": 0.2604, "examples": 5000, "loss": 2.028413623046875 }, "validation_seconds": 0.17042040824890137 }, { "epoch": 73, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.082765061314901, "train_seconds": 8.776772260665894, "validation": { "accuracy": 0.262, "examples": 5000, "loss": 2.0308223388671873 }, "validation_seconds": 0.17062807083129883 }, { "epoch": 74, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.0817778525034587, "train_seconds": 8.77632451057434, "validation": { "accuracy": 0.2552, "examples": 5000, "loss": 2.030654541015625 }, "validation_seconds": 0.17026042938232422 }, { "epoch": 75, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.0830021292792424, "train_seconds": 8.779340982437134, "validation": { "accuracy": 0.261, "examples": 5000, "loss": 2.0272352783203127 }, "validation_seconds": 0.17082738876342773 }, { "epoch": 76, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.0825558311886256, "train_seconds": 8.77163314819336, "validation": { "accuracy": 0.2564, "examples": 5000, "loss": 2.027567041015625 }, "validation_seconds": 0.17051172256469727 }, { "epoch": 77, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.0800666684468587, "train_seconds": 8.77224326133728, "validation": { "accuracy": 0.2624, "examples": 5000, "loss": 2.0338967041015623 }, "validation_seconds": 0.17057394981384277 }, { "epoch": 78, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.0857824336157904, "train_seconds": 8.738943576812744, "validation": { "accuracy": 0.2544, "examples": 5000, "loss": 2.028851806640625 }, "validation_seconds": 0.17060256004333496 }, { "epoch": 79, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.084057655037774, "train_seconds": 8.777712106704712, "validation": { "accuracy": 0.2604, "examples": 5000, "loss": 2.0281642822265624 }, "validation_seconds": 0.17107605934143066 }, { "epoch": 80, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.0811874683803984, "train_seconds": 8.779063940048218, "validation": { "accuracy": 0.253, "examples": 5000, "loss": 2.0347160400390627 }, "validation_seconds": 0.17075872421264648 }, { "epoch": 81, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.0836294103834363, "train_seconds": 8.772765398025513, "validation": { "accuracy": 0.2602, "examples": 5000, "loss": 2.0277829833984375 }, "validation_seconds": 0.1703035831451416 }, { "epoch": 82, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.0816474262237548, "train_seconds": 8.77937364578247, "validation": { "accuracy": 0.2562, "examples": 5000, "loss": 2.032101416015625 }, "validation_seconds": 0.17008304595947266 }, { "epoch": 83, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.079528014585707, "train_seconds": 8.772222995758057, "validation": { "accuracy": 0.2612, "examples": 5000, "loss": 2.0306248291015625 }, "validation_seconds": 0.17063570022583008 }, { "epoch": 84, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.086056492869059, "train_seconds": 8.776498079299927, "validation": { "accuracy": 0.2606, "examples": 5000, "loss": 2.02940244140625 }, "validation_seconds": 0.17087340354919434 }, { "epoch": 85, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.084182815806071, "train_seconds": 8.77548599243164, "validation": { "accuracy": 0.2572, "examples": 5000, "loss": 2.033000146484375 }, "validation_seconds": 0.1705179214477539 }, { "epoch": 86, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.082096115875244, "train_seconds": 8.770426034927368, "validation": { "accuracy": 0.2636, "examples": 5000, "loss": 2.0297025390625 }, "validation_seconds": 0.17105412483215332 }, { "epoch": 87, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.079115822007921, "train_seconds": 8.773576021194458, "validation": { "accuracy": 0.2562, "examples": 5000, "loss": 2.03552451171875 }, "validation_seconds": 0.1705305576324463 }, { "epoch": 88, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.0829235925462513, "train_seconds": 8.77567744255066, "validation": { "accuracy": 0.2576, "examples": 5000, "loss": 2.0320380126953124 }, "validation_seconds": 0.17091631889343262 }, { "epoch": 89, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.0760149488237167, "train_seconds": 8.781469583511353, "validation": { "accuracy": 0.2592, "examples": 5000, "loss": 2.0298075439453127 }, "validation_seconds": 0.17047929763793945 }, { "epoch": 90, "lr": 0.00030000000000000003, "output_lr": 0.00030000000000000003, "train_examples": 45000, "train_loss": 2.081927407285902, "train_seconds": 8.773695468902588, "validation": { "accuracy": 0.2602, "examples": 5000, "loss": 2.0261443359375 }, "validation_seconds": 0.17052292823791504 }, { "epoch": 91, "lr": 3.0000000000000004e-05, "output_lr": 3.0000000000000004e-05, "train_examples": 45000, "train_loss": 2.0803893885294595, "train_seconds": 8.772611379623413, "validation": { "accuracy": 0.258, "examples": 5000, "loss": 2.0315213623046877 }, "validation_seconds": 0.1705324649810791 }, { "epoch": 92, "lr": 3.0000000000000004e-05, "output_lr": 3.0000000000000004e-05, "train_examples": 45000, "train_loss": 2.082957981236776, "train_seconds": 8.775248765945435, "validation": { "accuracy": 0.2594, "examples": 5000, "loss": 2.03054775390625 }, "validation_seconds": 0.17024850845336914 }, { "epoch": 93, "lr": 3.0000000000000004e-05, "output_lr": 3.0000000000000004e-05, "train_examples": 45000, "train_loss": 2.0807154247283934, "train_seconds": 8.77206015586853, "validation": { "accuracy": 0.2604, "examples": 5000, "loss": 2.029746728515625 }, "validation_seconds": 0.17039227485656738 }, { "epoch": 94, "lr": 3.0000000000000004e-05, "output_lr": 3.0000000000000004e-05, "train_examples": 45000, "train_loss": 2.08182001953125, "train_seconds": 8.77450442314148, "validation": { "accuracy": 0.2586, "examples": 5000, "loss": 2.0316868408203126 }, "validation_seconds": 0.1709284782409668 }, { "epoch": 95, "lr": 3.0000000000000004e-05, "output_lr": 3.0000000000000004e-05, "train_examples": 45000, "train_loss": 2.077259977383084, "train_seconds": 8.733667135238647, "validation": { "accuracy": 0.2606, "examples": 5000, "loss": 2.028377978515625 }, "validation_seconds": 0.17060089111328125 }, { "epoch": 96, "lr": 3.0000000000000004e-05, "output_lr": 3.0000000000000004e-05, "train_examples": 45000, "train_loss": 2.0803276078542075, "train_seconds": 8.77278447151184, "validation": { "accuracy": 0.2576, "examples": 5000, "loss": 2.0310527099609375 }, "validation_seconds": 0.17107558250427246 }, { "epoch": 97, "lr": 3.0000000000000004e-05, "output_lr": 3.0000000000000004e-05, "train_examples": 45000, "train_loss": 2.0836526218838163, "train_seconds": 8.760288715362549, "validation": { "accuracy": 0.257, "examples": 5000, "loss": 2.0310808349609375 }, "validation_seconds": 0.17074060440063477 }, { "epoch": 98, "lr": 3.0000000000000004e-05, "output_lr": 3.0000000000000004e-05, "train_examples": 45000, "train_loss": 2.0804748294406465, "train_seconds": 8.81708836555481, "validation": { "accuracy": 0.2586, "examples": 5000, "loss": 2.0306814208984374 }, "validation_seconds": 0.17048335075378418 }, { "epoch": 99, "lr": 3.0000000000000004e-05, "output_lr": 3.0000000000000004e-05, "train_examples": 45000, "train_loss": 2.0814546067555746, "train_seconds": 8.773549795150757, "validation": { "accuracy": 0.2598, "examples": 5000, "loss": 2.028793017578125 }, "validation_seconds": 0.17067503929138184 }, { "epoch": 100, "lr": 3.0000000000000004e-05, "output_lr": 3.0000000000000004e-05, "train_examples": 45000, "train_loss": 2.081185230784946, "train_seconds": 8.774254560470581, "validation": { "accuracy": 0.2618, "examples": 5000, "loss": 2.029934326171875 }, "validation_seconds": 0.17056846618652344 } ], "evaluation_protocol": { "split": "validation", "test_evaluations": 0, "test_used_for_selection": false }, "final": { "accuracy": 0.2618, "evaluation_seconds": 0.17047953605651855, "examples": 5000, "finite": true, "loss": 2.029934326171875 }, "first_nonfinite_step": null, "hardware": { "cuda_device_name": "NVIDIA RTX A6000", "cuda_visible_devices": "0", "device": "cuda", "device_total_memory_bytes": 50896568320, "peak_memory_allocated_bytes": 2322423808, "peak_memory_reserved_bytes": 3162505216, "torch_version": "2.3.1+cu118" }, "layers": [], "protocol_family": "resnet_local_learning_crossover", "provenance": { "git_commit": "f55872beaee658ff76ac15fd421b73017820c320", "git_tracked_dirty": false }, "schema_version": 1, "split": { "cifar_source_files": [ { "bytes": 31035704, "path": "/orion/u/oscarwan/scratch/sdil-data/cifar-10-batches-py/data_batch_1", "sha256": "54636561a3ce25bd3e19253c6b0d8538147b0ae398331ac4a2d86c6d987368cd" }, { "bytes": 31035320, "path": "/orion/u/oscarwan/scratch/sdil-data/cifar-10-batches-py/data_batch_2", "sha256": "766b2cef9fbc745cf056b3152224f7cf77163b330ea9a15f9392beb8b89bc5a8" }, { "bytes": 31035999, "path": "/orion/u/oscarwan/scratch/sdil-data/cifar-10-batches-py/data_batch_3", "sha256": "0f00d98ebfb30b3ec0ad19f9756dc2630b89003e10525f5e148445e82aa6a1f9" }, { "bytes": 31035696, "path": "/orion/u/oscarwan/scratch/sdil-data/cifar-10-batches-py/data_batch_4", "sha256": "3f7bb240661948b8f4d53e36ec720d8306f5668bd0071dcb4e6c947f78e9682b" }, { "bytes": 31035623, "path": "/orion/u/oscarwan/scratch/sdil-data/cifar-10-batches-py/data_batch_5", "sha256": "d91802434d8376bbaeeadf58a737e3a1b12ac839077e931237e0dcd43adcb154" }, { "bytes": 31035526, "path": "/orion/u/oscarwan/scratch/sdil-data/cifar-10-batches-py/test_batch", "sha256": "f53d8d457504f7cff4ea9e021afcf0e0ad8e24a91f3fc42091b8adef61157831" } ], "dataset": "cifar10", "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": 45000, "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" }, "total_wall_seconds": 896.4814369678497, "work": { "candidate_label_evaluation_presentations": 0, "completed_global_epochs": 100, "feedforward_example_passes": 14005000, "forward_macs_per_example": 68862592, "forward_parameter_count": 464154, "local_target_backward_example_evaluations": 0, "local_vjp_example_evaluations": 0, "logical_task_loss_queries": 0, "num_trainable_layers": 32, "ordinary_training_examples": 4500000, "ordinary_validation_examples": 505000, "relaxation_example_passes": 0, "training_example_presentations": 9000000 } }