summaryrefslogtreecommitdiff
path: root/results/resnet_crossover/r2/resnet-r2-dualprop-d32-lr0p05.json
diff options
context:
space:
mode:
authorOscar Wan <oscarwan@orionp2.stanford.edu>2026-08-26 08:50:07 -0700
committerOscar Wan <oscarwan@orionp2.stanford.edu>2026-08-26 08:50:07 -0700
commit3c7c221ae6cfcdf6840da1d5262d94c75d56288b (patch)
treee13adbf676e3b5e1dc57fcf26dcc153c7983abc4 /results/resnet_crossover/r2/resnet-r2-dualprop-d32-lr0p05.json
parentf55872beaee658ff76ac15fd421b73017820c320 (diff)
experiment: add completed A6000 crossover resultsa6000-x1-results
Diffstat (limited to 'results/resnet_crossover/r2/resnet-r2-dualprop-d32-lr0p05.json')
-rw-r--r--results/resnet_crossover/r2/resnet-r2-dualprop-d32-lr0p05.json2965
1 files changed, 2965 insertions, 0 deletions
diff --git a/results/resnet_crossover/r2/resnet-r2-dualprop-d32-lr0p05.json b/results/resnet_crossover/r2/resnet-r2-dualprop-d32-lr0p05.json
new file mode 100644
index 0000000..31332f5
--- /dev/null
+++ b/results/resnet_crossover/r2/resnet-r2-dualprop-d32-lr0p05.json
@@ -0,0 +1,2965 @@
+{
+ "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": 200,
+ "eval_every": 1,
+ "eval_split": "validation",
+ "feedback_seed": 1729,
+ "ff_score_from_layer": 1,
+ "ff_threshold": 2.0,
+ "loader_seed": 0,
+ "lr": 0.05,
+ "lr_gamma": 0.1,
+ "lr_milestones": "100,150",
+ "lr_schedule": "step",
+ "max_steps": 0,
+ "method": "dualprop",
+ "momentum": 0.9,
+ "out": "/orion/u/oscarwan/scratch/sdil-work/sdil-a6000-full-rerun/results/resnet_crossover/r2/resnet-r2-dualprop-d32-lr0p05.json",
+ "output_lr": 0.05,
+ "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": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 1.8004995579401653,
+ "train_seconds": 563.7639303207397,
+ "validation": {
+ "accuracy": 0.4328,
+ "examples": 5000,
+ "loss": 1.557549609375
+ },
+ "validation_seconds": 0.19327449798583984
+ },
+ {
+ "epoch": 2,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 1.4176780060662164,
+ "train_seconds": 558.1841831207275,
+ "validation": {
+ "accuracy": 0.4978,
+ "examples": 5000,
+ "loss": 1.464878515625
+ },
+ "validation_seconds": 0.17017483711242676
+ },
+ {
+ "epoch": 3,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 1.2581868873596191,
+ "train_seconds": 558.1408138275146,
+ "validation": {
+ "accuracy": 0.5734,
+ "examples": 5000,
+ "loss": 1.216384326171875
+ },
+ "validation_seconds": 0.17085552215576172
+ },
+ {
+ "epoch": 4,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 1.1536001446617974,
+ "train_seconds": 558.4430482387543,
+ "validation": {
+ "accuracy": 0.605,
+ "examples": 5000,
+ "loss": 1.1569667236328125
+ },
+ "validation_seconds": 0.17072129249572754
+ },
+ {
+ "epoch": 5,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 1.0694282343122694,
+ "train_seconds": 558.2857687473297,
+ "validation": {
+ "accuracy": 0.5586,
+ "examples": 5000,
+ "loss": 1.3289294189453125
+ },
+ "validation_seconds": 0.1704881191253662
+ },
+ {
+ "epoch": 6,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.9969480787065294,
+ "train_seconds": 558.2259268760681,
+ "validation": {
+ "accuracy": 0.6168,
+ "examples": 5000,
+ "loss": 1.1593406494140626
+ },
+ "validation_seconds": 0.17039990425109863
+ },
+ {
+ "epoch": 7,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.9467010908550686,
+ "train_seconds": 558.310210943222,
+ "validation": {
+ "accuracy": 0.672,
+ "examples": 5000,
+ "loss": 0.9539438110351562
+ },
+ "validation_seconds": 0.17075109481811523
+ },
+ {
+ "epoch": 8,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.9019010075887044,
+ "train_seconds": 558.3344357013702,
+ "validation": {
+ "accuracy": 0.6492,
+ "examples": 5000,
+ "loss": 0.9964987182617188
+ },
+ "validation_seconds": 0.1706087589263916
+ },
+ {
+ "epoch": 9,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.8510134382459852,
+ "train_seconds": 558.2307453155518,
+ "validation": {
+ "accuracy": 0.6824,
+ "examples": 5000,
+ "loss": 0.8969764892578125
+ },
+ "validation_seconds": 0.17046070098876953
+ },
+ {
+ "epoch": 10,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.8201242864820693,
+ "train_seconds": 558.233719587326,
+ "validation": {
+ "accuracy": 0.6836,
+ "examples": 5000,
+ "loss": 0.9454015991210938
+ },
+ "validation_seconds": 0.170670747756958
+ },
+ {
+ "epoch": 11,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.7837703473091125,
+ "train_seconds": 558.2044539451599,
+ "validation": {
+ "accuracy": 0.7118,
+ "examples": 5000,
+ "loss": 0.86032119140625
+ },
+ "validation_seconds": 0.17042040824890137
+ },
+ {
+ "epoch": 12,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.7580736145867242,
+ "train_seconds": 557.9444570541382,
+ "validation": {
+ "accuracy": 0.7284,
+ "examples": 5000,
+ "loss": 0.7895968017578125
+ },
+ "validation_seconds": 0.1704864501953125
+ },
+ {
+ "epoch": 13,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.7373642958958944,
+ "train_seconds": 557.9286274909973,
+ "validation": {
+ "accuracy": 0.679,
+ "examples": 5000,
+ "loss": 0.9507829223632812
+ },
+ "validation_seconds": 0.17061734199523926
+ },
+ {
+ "epoch": 14,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.7064299224853515,
+ "train_seconds": 557.9418158531189,
+ "validation": {
+ "accuracy": 0.7248,
+ "examples": 5000,
+ "loss": 0.7915084350585937
+ },
+ "validation_seconds": 0.17058253288269043
+ },
+ {
+ "epoch": 15,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.6900307644420199,
+ "train_seconds": 557.9299385547638,
+ "validation": {
+ "accuracy": 0.7308,
+ "examples": 5000,
+ "loss": 0.8004146240234375
+ },
+ "validation_seconds": 0.1706540584564209
+ },
+ {
+ "epoch": 16,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.673870362080468,
+ "train_seconds": 557.9290549755096,
+ "validation": {
+ "accuracy": 0.715,
+ "examples": 5000,
+ "loss": 0.8556584228515625
+ },
+ "validation_seconds": 0.17044925689697266
+ },
+ {
+ "epoch": 17,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.6590127436743842,
+ "train_seconds": 557.9305055141449,
+ "validation": {
+ "accuracy": 0.7504,
+ "examples": 5000,
+ "loss": 0.7366830810546875
+ },
+ "validation_seconds": 0.17050766944885254
+ },
+ {
+ "epoch": 18,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.6435344869931539,
+ "train_seconds": 557.9347696304321,
+ "validation": {
+ "accuracy": 0.736,
+ "examples": 5000,
+ "loss": 0.7978809814453125
+ },
+ "validation_seconds": 0.17049336433410645
+ },
+ {
+ "epoch": 19,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.6259501564449734,
+ "train_seconds": 557.9711971282959,
+ "validation": {
+ "accuracy": 0.761,
+ "examples": 5000,
+ "loss": 0.7134017700195312
+ },
+ "validation_seconds": 0.1704099178314209
+ },
+ {
+ "epoch": 20,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.6067184276686775,
+ "train_seconds": 557.9173040390015,
+ "validation": {
+ "accuracy": 0.7484,
+ "examples": 5000,
+ "loss": 0.754651025390625
+ },
+ "validation_seconds": 0.17058968544006348
+ },
+ {
+ "epoch": 21,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.5990790719456143,
+ "train_seconds": 557.9328565597534,
+ "validation": {
+ "accuracy": 0.7502,
+ "examples": 5000,
+ "loss": 0.7732291381835937
+ },
+ "validation_seconds": 0.17064690589904785
+ },
+ {
+ "epoch": 22,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.5874932703706953,
+ "train_seconds": 558.1986618041992,
+ "validation": {
+ "accuracy": 0.7624,
+ "examples": 5000,
+ "loss": 0.7070720458984375
+ },
+ "validation_seconds": 0.17105531692504883
+ },
+ {
+ "epoch": 23,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.5748521372265286,
+ "train_seconds": 558.8032100200653,
+ "validation": {
+ "accuracy": 0.7744,
+ "examples": 5000,
+ "loss": 0.6826259155273438
+ },
+ "validation_seconds": 0.1711256504058838
+ },
+ {
+ "epoch": 24,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.5692189562638601,
+ "train_seconds": 559.0503284931183,
+ "validation": {
+ "accuracy": 0.7674,
+ "examples": 5000,
+ "loss": 0.7043788940429687
+ },
+ "validation_seconds": 0.17097234725952148
+ },
+ {
+ "epoch": 25,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.5517877299308777,
+ "train_seconds": 558.8799324035645,
+ "validation": {
+ "accuracy": 0.7764,
+ "examples": 5000,
+ "loss": 0.6821715209960938
+ },
+ "validation_seconds": 0.17085647583007812
+ },
+ {
+ "epoch": 26,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.5391471517350939,
+ "train_seconds": 558.8605720996857,
+ "validation": {
+ "accuracy": 0.7406,
+ "examples": 5000,
+ "loss": 0.8436376220703125
+ },
+ "validation_seconds": 0.17091798782348633
+ },
+ {
+ "epoch": 27,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.5402941966957516,
+ "train_seconds": 558.9005839824677,
+ "validation": {
+ "accuracy": 0.7802,
+ "examples": 5000,
+ "loss": 0.6626584350585938
+ },
+ "validation_seconds": 0.17087268829345703
+ },
+ {
+ "epoch": 28,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.5282208977381389,
+ "train_seconds": 558.775890827179,
+ "validation": {
+ "accuracy": 0.7828,
+ "examples": 5000,
+ "loss": 0.6616478881835938
+ },
+ "validation_seconds": 0.1705186367034912
+ },
+ {
+ "epoch": 29,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.5246573151694404,
+ "train_seconds": 558.2551503181458,
+ "validation": {
+ "accuracy": 0.7682,
+ "examples": 5000,
+ "loss": 0.7028174072265625
+ },
+ "validation_seconds": 0.17051959037780762
+ },
+ {
+ "epoch": 30,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.5210510110855102,
+ "train_seconds": 558.6386096477509,
+ "validation": {
+ "accuracy": 0.7812,
+ "examples": 5000,
+ "loss": 0.6577355712890625
+ },
+ "validation_seconds": 0.1710948944091797
+ },
+ {
+ "epoch": 31,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.5098838945123885,
+ "train_seconds": 558.713506937027,
+ "validation": {
+ "accuracy": 0.7962,
+ "examples": 5000,
+ "loss": 0.6501040649414063
+ },
+ "validation_seconds": 0.17058014869689941
+ },
+ {
+ "epoch": 32,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.5057243042204115,
+ "train_seconds": 558.045622587204,
+ "validation": {
+ "accuracy": 0.7628,
+ "examples": 5000,
+ "loss": 0.725605029296875
+ },
+ "validation_seconds": 0.17093896865844727
+ },
+ {
+ "epoch": 33,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.49384519081115724,
+ "train_seconds": 558.0422115325928,
+ "validation": {
+ "accuracy": 0.795,
+ "examples": 5000,
+ "loss": 0.6151902099609375
+ },
+ "validation_seconds": 0.17077398300170898
+ },
+ {
+ "epoch": 34,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.4975082278781467,
+ "train_seconds": 558.0692346096039,
+ "validation": {
+ "accuracy": 0.7904,
+ "examples": 5000,
+ "loss": 0.6437848266601562
+ },
+ "validation_seconds": 0.17070817947387695
+ },
+ {
+ "epoch": 35,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.4914001432524787,
+ "train_seconds": 558.0532667636871,
+ "validation": {
+ "accuracy": 0.7844,
+ "examples": 5000,
+ "loss": 0.6282599609375
+ },
+ "validation_seconds": 0.1710524559020996
+ },
+ {
+ "epoch": 36,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.48799099100960625,
+ "train_seconds": 558.0403628349304,
+ "validation": {
+ "accuracy": 0.7858,
+ "examples": 5000,
+ "loss": 0.664295068359375
+ },
+ "validation_seconds": 0.17072486877441406
+ },
+ {
+ "epoch": 37,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.48119476664861044,
+ "train_seconds": 558.0373411178589,
+ "validation": {
+ "accuracy": 0.7984,
+ "examples": 5000,
+ "loss": 0.6241986450195313
+ },
+ "validation_seconds": 0.17064595222473145
+ },
+ {
+ "epoch": 38,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.4726532228257921,
+ "train_seconds": 558.2699501514435,
+ "validation": {
+ "accuracy": 0.755,
+ "examples": 5000,
+ "loss": 0.750040283203125
+ },
+ "validation_seconds": 0.17100858688354492
+ },
+ {
+ "epoch": 39,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.45894244771533543,
+ "train_seconds": 558.2855563163757,
+ "validation": {
+ "accuracy": 0.8098,
+ "examples": 5000,
+ "loss": 0.5754578002929688
+ },
+ "validation_seconds": 0.1708691120147705
+ },
+ {
+ "epoch": 40,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.4561460572242737,
+ "train_seconds": 558.2716417312622,
+ "validation": {
+ "accuracy": 0.7994,
+ "examples": 5000,
+ "loss": 0.6093698974609375
+ },
+ "validation_seconds": 0.17070937156677246
+ },
+ {
+ "epoch": 41,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.45154505163828534,
+ "train_seconds": 558.3074398040771,
+ "validation": {
+ "accuracy": 0.8016,
+ "examples": 5000,
+ "loss": 0.58370234375
+ },
+ "validation_seconds": 0.17050838470458984
+ },
+ {
+ "epoch": 42,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.44932130374908447,
+ "train_seconds": 558.3013200759888,
+ "validation": {
+ "accuracy": 0.7796,
+ "examples": 5000,
+ "loss": 0.6969513793945312
+ },
+ "validation_seconds": 0.17067265510559082
+ },
+ {
+ "epoch": 43,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.4478407272444831,
+ "train_seconds": 558.3348424434662,
+ "validation": {
+ "accuracy": 0.794,
+ "examples": 5000,
+ "loss": 0.6354197021484375
+ },
+ "validation_seconds": 0.1706705093383789
+ },
+ {
+ "epoch": 44,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.4487147048632304,
+ "train_seconds": 558.3120241165161,
+ "validation": {
+ "accuracy": 0.7848,
+ "examples": 5000,
+ "loss": 0.6651848510742188
+ },
+ "validation_seconds": 0.17068982124328613
+ },
+ {
+ "epoch": 45,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.44171588740878637,
+ "train_seconds": 558.7507667541504,
+ "validation": {
+ "accuracy": 0.7996,
+ "examples": 5000,
+ "loss": 0.6317643798828125
+ },
+ "validation_seconds": 0.17115306854248047
+ },
+ {
+ "epoch": 46,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.44115461015701296,
+ "train_seconds": 558.9619481563568,
+ "validation": {
+ "accuracy": 0.8178,
+ "examples": 5000,
+ "loss": 0.5661478759765625
+ },
+ "validation_seconds": 0.17086172103881836
+ },
+ {
+ "epoch": 47,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.42366895433531865,
+ "train_seconds": 558.5156924724579,
+ "validation": {
+ "accuracy": 0.8084,
+ "examples": 5000,
+ "loss": 0.6045284912109375
+ },
+ "validation_seconds": 0.17072129249572754
+ },
+ {
+ "epoch": 48,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.4294327813837263,
+ "train_seconds": 558.0816631317139,
+ "validation": {
+ "accuracy": 0.8078,
+ "examples": 5000,
+ "loss": 0.6047329711914062
+ },
+ "validation_seconds": 0.17075610160827637
+ },
+ {
+ "epoch": 49,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.42285685943497553,
+ "train_seconds": 558.0882160663605,
+ "validation": {
+ "accuracy": 0.7926,
+ "examples": 5000,
+ "loss": 0.6357721313476562
+ },
+ "validation_seconds": 0.17081046104431152
+ },
+ {
+ "epoch": 50,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.42114123702579076,
+ "train_seconds": 558.0787107944489,
+ "validation": {
+ "accuracy": 0.7966,
+ "examples": 5000,
+ "loss": 0.62530205078125
+ },
+ "validation_seconds": 0.17049264907836914
+ },
+ {
+ "epoch": 51,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.4282205765194363,
+ "train_seconds": 558.0904269218445,
+ "validation": {
+ "accuracy": 0.7872,
+ "examples": 5000,
+ "loss": 0.6708918823242187
+ },
+ "validation_seconds": 0.17061495780944824
+ },
+ {
+ "epoch": 52,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.4141006850772434,
+ "train_seconds": 558.082554101944,
+ "validation": {
+ "accuracy": 0.8066,
+ "examples": 5000,
+ "loss": 0.5993174438476563
+ },
+ "validation_seconds": 0.17061448097229004
+ },
+ {
+ "epoch": 53,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.4119393820762634,
+ "train_seconds": 558.0787973403931,
+ "validation": {
+ "accuracy": 0.8058,
+ "examples": 5000,
+ "loss": 0.6037708984375
+ },
+ "validation_seconds": 0.1706247329711914
+ },
+ {
+ "epoch": 54,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.4145192951838175,
+ "train_seconds": 558.0709445476532,
+ "validation": {
+ "accuracy": 0.7988,
+ "examples": 5000,
+ "loss": 0.6170823974609375
+ },
+ "validation_seconds": 0.17055344581604004
+ },
+ {
+ "epoch": 55,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.4110671861436632,
+ "train_seconds": 558.0375108718872,
+ "validation": {
+ "accuracy": 0.7954,
+ "examples": 5000,
+ "loss": 0.64954423828125
+ },
+ "validation_seconds": 0.17034626007080078
+ },
+ {
+ "epoch": 56,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.4194034888267517,
+ "train_seconds": 558.2900795936584,
+ "validation": {
+ "accuracy": 0.7774,
+ "examples": 5000,
+ "loss": 0.6947837890625
+ },
+ "validation_seconds": 0.17063689231872559
+ },
+ {
+ "epoch": 57,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.40363140162361993,
+ "train_seconds": 558.2407958507538,
+ "validation": {
+ "accuracy": 0.7928,
+ "examples": 5000,
+ "loss": 0.6374149169921876
+ },
+ "validation_seconds": 0.17067933082580566
+ },
+ {
+ "epoch": 58,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.4071178272035387,
+ "train_seconds": 558.2327244281769,
+ "validation": {
+ "accuracy": 0.8138,
+ "examples": 5000,
+ "loss": 0.5728923095703125
+ },
+ "validation_seconds": 0.17045927047729492
+ },
+ {
+ "epoch": 59,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.40012233917978074,
+ "train_seconds": 558.2001922130585,
+ "validation": {
+ "accuracy": 0.7878,
+ "examples": 5000,
+ "loss": 0.6841786987304688
+ },
+ "validation_seconds": 0.17034292221069336
+ },
+ {
+ "epoch": 60,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.3900605029688941,
+ "train_seconds": 558.22704911232,
+ "validation": {
+ "accuracy": 0.8002,
+ "examples": 5000,
+ "loss": 0.6395064331054687
+ },
+ "validation_seconds": 0.17066001892089844
+ },
+ {
+ "epoch": 61,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.38680489319165545,
+ "train_seconds": 558.1992979049683,
+ "validation": {
+ "accuracy": 0.7962,
+ "examples": 5000,
+ "loss": 0.6516251831054688
+ },
+ "validation_seconds": 0.1707756519317627
+ },
+ {
+ "epoch": 62,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.37478828299310474,
+ "train_seconds": 558.2286744117737,
+ "validation": {
+ "accuracy": 0.813,
+ "examples": 5000,
+ "loss": 0.5861128295898438
+ },
+ "validation_seconds": 0.17053627967834473
+ },
+ {
+ "epoch": 63,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.3731063287470076,
+ "train_seconds": 558.2254464626312,
+ "validation": {
+ "accuracy": 0.8112,
+ "examples": 5000,
+ "loss": 0.6001865600585937
+ },
+ "validation_seconds": 0.17064380645751953
+ },
+ {
+ "epoch": 64,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.3759114433394538,
+ "train_seconds": 558.2426578998566,
+ "validation": {
+ "accuracy": 0.8086,
+ "examples": 5000,
+ "loss": 0.6028027587890625
+ },
+ "validation_seconds": 0.1704556941986084
+ },
+ {
+ "epoch": 65,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.37589247471491494,
+ "train_seconds": 558.2081379890442,
+ "validation": {
+ "accuracy": 0.8192,
+ "examples": 5000,
+ "loss": 0.573157470703125
+ },
+ "validation_seconds": 0.17031216621398926
+ },
+ {
+ "epoch": 66,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.37520403162903254,
+ "train_seconds": 557.9686496257782,
+ "validation": {
+ "accuracy": 0.807,
+ "examples": 5000,
+ "loss": 0.6031291137695313
+ },
+ "validation_seconds": 0.17033934593200684
+ },
+ {
+ "epoch": 67,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.37030814722379046,
+ "train_seconds": 557.9458205699921,
+ "validation": {
+ "accuracy": 0.8084,
+ "examples": 5000,
+ "loss": 0.61266845703125
+ },
+ "validation_seconds": 0.17030119895935059
+ },
+ {
+ "epoch": 68,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.3623477321412828,
+ "train_seconds": 558.0508515834808,
+ "validation": {
+ "accuracy": 0.8238,
+ "examples": 5000,
+ "loss": 0.541462060546875
+ },
+ "validation_seconds": 0.17065834999084473
+ },
+ {
+ "epoch": 69,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.3633347564008501,
+ "train_seconds": 558.0455734729767,
+ "validation": {
+ "accuracy": 0.8018,
+ "examples": 5000,
+ "loss": 0.6289015258789062
+ },
+ "validation_seconds": 0.17070508003234863
+ },
+ {
+ "epoch": 70,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.35279572947820026,
+ "train_seconds": 558.0403199195862,
+ "validation": {
+ "accuracy": 0.8192,
+ "examples": 5000,
+ "loss": 0.5510507446289062
+ },
+ "validation_seconds": 0.1707603931427002
+ },
+ {
+ "epoch": 71,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.35029824435975815,
+ "train_seconds": 558.0362615585327,
+ "validation": {
+ "accuracy": 0.8078,
+ "examples": 5000,
+ "loss": 0.6152504028320313
+ },
+ "validation_seconds": 0.17045378684997559
+ },
+ {
+ "epoch": 72,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.36111382049984403,
+ "train_seconds": 558.0379478931427,
+ "validation": {
+ "accuracy": 0.7906,
+ "examples": 5000,
+ "loss": 0.6859040771484375
+ },
+ "validation_seconds": 0.17042279243469238
+ },
+ {
+ "epoch": 73,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.3539845959769355,
+ "train_seconds": 558.021562576294,
+ "validation": {
+ "accuracy": 0.8212,
+ "examples": 5000,
+ "loss": 0.5775150268554687
+ },
+ "validation_seconds": 0.1705188751220703
+ },
+ {
+ "epoch": 74,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.3474092450671726,
+ "train_seconds": 558.0308134555817,
+ "validation": {
+ "accuracy": 0.8092,
+ "examples": 5000,
+ "loss": 0.6013641967773438
+ },
+ "validation_seconds": 0.17047381401062012
+ },
+ {
+ "epoch": 75,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.35469483111169603,
+ "train_seconds": 558.0403017997742,
+ "validation": {
+ "accuracy": 0.8254,
+ "examples": 5000,
+ "loss": 0.54343388671875
+ },
+ "validation_seconds": 0.17022395133972168
+ },
+ {
+ "epoch": 76,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.3471669368902842,
+ "train_seconds": 558.0478513240814,
+ "validation": {
+ "accuracy": 0.807,
+ "examples": 5000,
+ "loss": 0.631876904296875
+ },
+ "validation_seconds": 0.1706843376159668
+ },
+ {
+ "epoch": 77,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.3457748140070174,
+ "train_seconds": 557.9572274684906,
+ "validation": {
+ "accuracy": 0.8212,
+ "examples": 5000,
+ "loss": 0.5879268798828124
+ },
+ "validation_seconds": 0.17055583000183105
+ },
+ {
+ "epoch": 78,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.33029527435302736,
+ "train_seconds": 557.9223172664642,
+ "validation": {
+ "accuracy": 0.828,
+ "examples": 5000,
+ "loss": 0.5529513061523438
+ },
+ "validation_seconds": 0.1705477237701416
+ },
+ {
+ "epoch": 79,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.33166916541523406,
+ "train_seconds": 557.9095616340637,
+ "validation": {
+ "accuracy": 0.7824,
+ "examples": 5000,
+ "loss": 0.6858224365234376
+ },
+ "validation_seconds": 0.17069387435913086
+ },
+ {
+ "epoch": 80,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.33432486742072637,
+ "train_seconds": 557.9043171405792,
+ "validation": {
+ "accuracy": 0.8154,
+ "examples": 5000,
+ "loss": 0.5964239135742188
+ },
+ "validation_seconds": 0.1705014705657959
+ },
+ {
+ "epoch": 81,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.32564821405410765,
+ "train_seconds": 557.9098527431488,
+ "validation": {
+ "accuracy": 0.8184,
+ "examples": 5000,
+ "loss": 0.564143408203125
+ },
+ "validation_seconds": 0.1705763339996338
+ },
+ {
+ "epoch": 82,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.31590132386419506,
+ "train_seconds": 557.9197072982788,
+ "validation": {
+ "accuracy": 0.8242,
+ "examples": 5000,
+ "loss": 0.5693760864257813
+ },
+ "validation_seconds": 0.17065834999084473
+ },
+ {
+ "epoch": 83,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.3159837660683526,
+ "train_seconds": 557.9388198852539,
+ "validation": {
+ "accuracy": 0.8126,
+ "examples": 5000,
+ "loss": 0.6296174926757813
+ },
+ "validation_seconds": 0.17052793502807617
+ },
+ {
+ "epoch": 84,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.32203944091796877,
+ "train_seconds": 557.9468953609467,
+ "validation": {
+ "accuracy": 0.806,
+ "examples": 5000,
+ "loss": 0.6376169067382812
+ },
+ "validation_seconds": 0.17077326774597168
+ },
+ {
+ "epoch": 85,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.32088813879224987,
+ "train_seconds": 557.9684026241302,
+ "validation": {
+ "accuracy": 0.8202,
+ "examples": 5000,
+ "loss": 0.5882717407226562
+ },
+ "validation_seconds": 0.17091774940490723
+ },
+ {
+ "epoch": 86,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.315234685060713,
+ "train_seconds": 557.9197709560394,
+ "validation": {
+ "accuracy": 0.829,
+ "examples": 5000,
+ "loss": 0.5767408813476562
+ },
+ "validation_seconds": 0.17058467864990234
+ },
+ {
+ "epoch": 87,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.3200781061384413,
+ "train_seconds": 557.932461977005,
+ "validation": {
+ "accuracy": 0.8266,
+ "examples": 5000,
+ "loss": 0.5719342407226562
+ },
+ "validation_seconds": 0.17056751251220703
+ },
+ {
+ "epoch": 88,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.32161954258812797,
+ "train_seconds": 557.9289646148682,
+ "validation": {
+ "accuracy": 0.8242,
+ "examples": 5000,
+ "loss": 0.5660660522460937
+ },
+ "validation_seconds": 0.17087936401367188
+ },
+ {
+ "epoch": 89,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.31135784103075664,
+ "train_seconds": 557.9694781303406,
+ "validation": {
+ "accuracy": 0.7986,
+ "examples": 5000,
+ "loss": 0.6810950317382812
+ },
+ "validation_seconds": 0.17078852653503418
+ },
+ {
+ "epoch": 90,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.31059243212805854,
+ "train_seconds": 557.9276149272919,
+ "validation": {
+ "accuracy": 0.8162,
+ "examples": 5000,
+ "loss": 0.6226161254882813
+ },
+ "validation_seconds": 0.17064809799194336
+ },
+ {
+ "epoch": 91,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.303583119614919,
+ "train_seconds": 557.93887758255,
+ "validation": {
+ "accuracy": 0.8278,
+ "examples": 5000,
+ "loss": 0.575564697265625
+ },
+ "validation_seconds": 0.17057085037231445
+ },
+ {
+ "epoch": 92,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.2989853093253242,
+ "train_seconds": 557.9392566680908,
+ "validation": {
+ "accuracy": 0.8348,
+ "examples": 5000,
+ "loss": 0.5497850830078125
+ },
+ "validation_seconds": 0.17045283317565918
+ },
+ {
+ "epoch": 93,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.30290545211368136,
+ "train_seconds": 558.1350507736206,
+ "validation": {
+ "accuracy": 0.829,
+ "examples": 5000,
+ "loss": 0.564298681640625
+ },
+ "validation_seconds": 0.17111921310424805
+ },
+ {
+ "epoch": 94,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.31313663847181533,
+ "train_seconds": 559.4218375682831,
+ "validation": {
+ "accuracy": 0.8368,
+ "examples": 5000,
+ "loss": 0.5624496704101563
+ },
+ "validation_seconds": 0.17076325416564941
+ },
+ {
+ "epoch": 95,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.29203504636817507,
+ "train_seconds": 558.3938927650452,
+ "validation": {
+ "accuracy": 0.8092,
+ "examples": 5000,
+ "loss": 0.6693233520507812
+ },
+ "validation_seconds": 0.17063307762145996
+ },
+ {
+ "epoch": 96,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.3094335092279646,
+ "train_seconds": 558.233589887619,
+ "validation": {
+ "accuracy": 0.8352,
+ "examples": 5000,
+ "loss": 0.5715654907226563
+ },
+ "validation_seconds": 0.1705608367919922
+ },
+ {
+ "epoch": 97,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.3069044820573595,
+ "train_seconds": 558.2597603797913,
+ "validation": {
+ "accuracy": 0.8294,
+ "examples": 5000,
+ "loss": 0.5501206298828125
+ },
+ "validation_seconds": 0.1707000732421875
+ },
+ {
+ "epoch": 98,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.30671100243462457,
+ "train_seconds": 558.1544392108917,
+ "validation": {
+ "accuracy": 0.8368,
+ "examples": 5000,
+ "loss": 0.5354385192871094
+ },
+ "validation_seconds": 0.17098045349121094
+ },
+ {
+ "epoch": 99,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.30218536727163525,
+ "train_seconds": 557.9338235855103,
+ "validation": {
+ "accuracy": 0.8324,
+ "examples": 5000,
+ "loss": 0.5434078918457032
+ },
+ "validation_seconds": 0.17069149017333984
+ },
+ {
+ "epoch": 100,
+ "lr": 0.05,
+ "output_lr": 0.05,
+ "train_examples": 45000,
+ "train_loss": 0.28684145905971525,
+ "train_seconds": 558.2469670772552,
+ "validation": {
+ "accuracy": 0.8372,
+ "examples": 5000,
+ "loss": 0.5330762023925781
+ },
+ "validation_seconds": 0.17123103141784668
+ },
+ {
+ "epoch": 101,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.20102937622732586,
+ "train_seconds": 560.1686954498291,
+ "validation": {
+ "accuracy": 0.8634,
+ "examples": 5000,
+ "loss": 0.4263051818847656
+ },
+ "validation_seconds": 0.1707172393798828
+ },
+ {
+ "epoch": 102,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.1707019625398848,
+ "train_seconds": 560.6133296489716,
+ "validation": {
+ "accuracy": 0.8636,
+ "examples": 5000,
+ "loss": 0.42772237548828124
+ },
+ "validation_seconds": 0.17063069343566895
+ },
+ {
+ "epoch": 103,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.16009664902024798,
+ "train_seconds": 560.8625993728638,
+ "validation": {
+ "accuracy": 0.8652,
+ "examples": 5000,
+ "loss": 0.4311399169921875
+ },
+ "validation_seconds": 0.17083454132080078
+ },
+ {
+ "epoch": 104,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.1511841663281123,
+ "train_seconds": 559.8500211238861,
+ "validation": {
+ "accuracy": 0.8654,
+ "examples": 5000,
+ "loss": 0.42962298583984376
+ },
+ "validation_seconds": 0.1706254482269287
+ },
+ {
+ "epoch": 105,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.1450319301393297,
+ "train_seconds": 559.5080921649933,
+ "validation": {
+ "accuracy": 0.8668,
+ "examples": 5000,
+ "loss": 0.43597459716796877
+ },
+ "validation_seconds": 0.17025041580200195
+ },
+ {
+ "epoch": 106,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.1385588534726037,
+ "train_seconds": 558.8428866863251,
+ "validation": {
+ "accuracy": 0.8688,
+ "examples": 5000,
+ "loss": 0.44206187744140624
+ },
+ "validation_seconds": 0.17053508758544922
+ },
+ {
+ "epoch": 107,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.1349012790547477,
+ "train_seconds": 558.7124433517456,
+ "validation": {
+ "accuracy": 0.8644,
+ "examples": 5000,
+ "loss": 0.44303953857421874
+ },
+ "validation_seconds": 0.1705188751220703
+ },
+ {
+ "epoch": 108,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.13633077380127376,
+ "train_seconds": 558.8155248165131,
+ "validation": {
+ "accuracy": 0.866,
+ "examples": 5000,
+ "loss": 0.4498537658691406
+ },
+ "validation_seconds": 0.17037415504455566
+ },
+ {
+ "epoch": 109,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.12727142566045124,
+ "train_seconds": 559.8420617580414,
+ "validation": {
+ "accuracy": 0.8666,
+ "examples": 5000,
+ "loss": 0.4562872497558594
+ },
+ "validation_seconds": 0.17095494270324707
+ },
+ {
+ "epoch": 110,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.12273772934277852,
+ "train_seconds": 560.2172212600708,
+ "validation": {
+ "accuracy": 0.869,
+ "examples": 5000,
+ "loss": 0.46257111206054685
+ },
+ "validation_seconds": 0.1706855297088623
+ },
+ {
+ "epoch": 111,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.1186966257651647,
+ "train_seconds": 559.0523889064789,
+ "validation": {
+ "accuracy": 0.8694,
+ "examples": 5000,
+ "loss": 0.46149122924804686
+ },
+ "validation_seconds": 0.1704082489013672
+ },
+ {
+ "epoch": 112,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.1151852872159746,
+ "train_seconds": 558.8536169528961,
+ "validation": {
+ "accuracy": 0.8686,
+ "examples": 5000,
+ "loss": 0.468364404296875
+ },
+ "validation_seconds": 0.17055869102478027
+ },
+ {
+ "epoch": 113,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.11532456085681915,
+ "train_seconds": 558.7769916057587,
+ "validation": {
+ "accuracy": 0.871,
+ "examples": 5000,
+ "loss": 0.46807992553710936
+ },
+ "validation_seconds": 0.17038846015930176
+ },
+ {
+ "epoch": 114,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.11298640092081494,
+ "train_seconds": 558.7878985404968,
+ "validation": {
+ "accuracy": 0.8696,
+ "examples": 5000,
+ "loss": 0.4718349487304688
+ },
+ "validation_seconds": 0.1701805591583252
+ },
+ {
+ "epoch": 115,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.11400147623750899,
+ "train_seconds": 558.6093730926514,
+ "validation": {
+ "accuracy": 0.8706,
+ "examples": 5000,
+ "loss": 0.4778865234375
+ },
+ "validation_seconds": 0.1703321933746338
+ },
+ {
+ "epoch": 116,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.10852504232459598,
+ "train_seconds": 558.69162774086,
+ "validation": {
+ "accuracy": 0.8702,
+ "examples": 5000,
+ "loss": 0.48102982788085935
+ },
+ "validation_seconds": 0.16998052597045898
+ },
+ {
+ "epoch": 117,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.10526954917377895,
+ "train_seconds": 558.6054136753082,
+ "validation": {
+ "accuracy": 0.8664,
+ "examples": 5000,
+ "loss": 0.48121639404296873
+ },
+ "validation_seconds": 0.17051982879638672
+ },
+ {
+ "epoch": 118,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.10441256948245896,
+ "train_seconds": 559.6752879619598,
+ "validation": {
+ "accuracy": 0.8698,
+ "examples": 5000,
+ "loss": 0.4906882080078125
+ },
+ "validation_seconds": 0.17052960395812988
+ },
+ {
+ "epoch": 119,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.10324348965088527,
+ "train_seconds": 560.4540963172913,
+ "validation": {
+ "accuracy": 0.8706,
+ "examples": 5000,
+ "loss": 0.49455729370117185
+ },
+ "validation_seconds": 0.1706228256225586
+ },
+ {
+ "epoch": 120,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.10344991790453593,
+ "train_seconds": 560.4171199798584,
+ "validation": {
+ "accuracy": 0.8678,
+ "examples": 5000,
+ "loss": 0.4905825256347656
+ },
+ "validation_seconds": 0.17073750495910645
+ },
+ {
+ "epoch": 121,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.09832412128316032,
+ "train_seconds": 560.4106814861298,
+ "validation": {
+ "accuracy": 0.871,
+ "examples": 5000,
+ "loss": 0.498076904296875
+ },
+ "validation_seconds": 0.1704578399658203
+ },
+ {
+ "epoch": 122,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.09772718087302314,
+ "train_seconds": 560.4327030181885,
+ "validation": {
+ "accuracy": 0.8688,
+ "examples": 5000,
+ "loss": 0.5086375427246094
+ },
+ "validation_seconds": 0.17103815078735352
+ },
+ {
+ "epoch": 123,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.10162255337900586,
+ "train_seconds": 560.3050816059113,
+ "validation": {
+ "accuracy": 0.8672,
+ "examples": 5000,
+ "loss": 0.5079470642089844
+ },
+ "validation_seconds": 0.17122697830200195
+ },
+ {
+ "epoch": 124,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.09721829018063015,
+ "train_seconds": 560.3977375030518,
+ "validation": {
+ "accuracy": 0.8696,
+ "examples": 5000,
+ "loss": 0.5183314514160157
+ },
+ "validation_seconds": 0.1710338592529297
+ },
+ {
+ "epoch": 125,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.093445472296079,
+ "train_seconds": 560.1785962581635,
+ "validation": {
+ "accuracy": 0.8664,
+ "examples": 5000,
+ "loss": 0.5179838073730468
+ },
+ "validation_seconds": 0.17027044296264648
+ },
+ {
+ "epoch": 126,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.09554820556905534,
+ "train_seconds": 560.3105957508087,
+ "validation": {
+ "accuracy": 0.8654,
+ "examples": 5000,
+ "loss": 0.5202593566894531
+ },
+ "validation_seconds": 0.1709599494934082
+ },
+ {
+ "epoch": 127,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.0952698231961992,
+ "train_seconds": 560.3248987197876,
+ "validation": {
+ "accuracy": 0.8676,
+ "examples": 5000,
+ "loss": 0.5269124206542969
+ },
+ "validation_seconds": 0.17055344581604004
+ },
+ {
+ "epoch": 128,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.08956516365210215,
+ "train_seconds": 560.349618434906,
+ "validation": {
+ "accuracy": 0.8654,
+ "examples": 5000,
+ "loss": 0.5317488220214843
+ },
+ "validation_seconds": 0.17071056365966797
+ },
+ {
+ "epoch": 129,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.09056334316995408,
+ "train_seconds": 560.2911763191223,
+ "validation": {
+ "accuracy": 0.866,
+ "examples": 5000,
+ "loss": 0.5350216003417969
+ },
+ "validation_seconds": 0.1710660457611084
+ },
+ {
+ "epoch": 130,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.08700342583258946,
+ "train_seconds": 560.1963536739349,
+ "validation": {
+ "accuracy": 0.8686,
+ "examples": 5000,
+ "loss": 0.5305935913085937
+ },
+ "validation_seconds": 0.17052793502807617
+ },
+ {
+ "epoch": 131,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.08819638755056593,
+ "train_seconds": 560.2379078865051,
+ "validation": {
+ "accuracy": 0.8654,
+ "examples": 5000,
+ "loss": 0.5385918884277344
+ },
+ "validation_seconds": 0.1703329086303711
+ },
+ {
+ "epoch": 132,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.0854943878531456,
+ "train_seconds": 560.2816424369812,
+ "validation": {
+ "accuracy": 0.8658,
+ "examples": 5000,
+ "loss": 0.5405256225585937
+ },
+ "validation_seconds": 0.17044830322265625
+ },
+ {
+ "epoch": 133,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.08349375572767523,
+ "train_seconds": 560.3219213485718,
+ "validation": {
+ "accuracy": 0.8636,
+ "examples": 5000,
+ "loss": 0.5356781799316406
+ },
+ "validation_seconds": 0.17093181610107422
+ },
+ {
+ "epoch": 134,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.08320598353147507,
+ "train_seconds": 560.3251914978027,
+ "validation": {
+ "accuracy": 0.8656,
+ "examples": 5000,
+ "loss": 0.5479776062011719
+ },
+ "validation_seconds": 0.17060375213623047
+ },
+ {
+ "epoch": 135,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.08140208751625484,
+ "train_seconds": 560.2988884449005,
+ "validation": {
+ "accuracy": 0.8656,
+ "examples": 5000,
+ "loss": 0.5495034912109374
+ },
+ "validation_seconds": 0.1703953742980957
+ },
+ {
+ "epoch": 136,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.08287404755486383,
+ "train_seconds": 559.8910458087921,
+ "validation": {
+ "accuracy": 0.8678,
+ "examples": 5000,
+ "loss": 0.5502274353027343
+ },
+ "validation_seconds": 0.17048001289367676
+ },
+ {
+ "epoch": 137,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.08143085203435686,
+ "train_seconds": 559.9673924446106,
+ "validation": {
+ "accuracy": 0.8664,
+ "examples": 5000,
+ "loss": 0.5595277648925782
+ },
+ "validation_seconds": 0.17099213600158691
+ },
+ {
+ "epoch": 138,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.08175256986684269,
+ "train_seconds": 559.9922814369202,
+ "validation": {
+ "accuracy": 0.865,
+ "examples": 5000,
+ "loss": 0.5596325988769532
+ },
+ "validation_seconds": 0.17058587074279785
+ },
+ {
+ "epoch": 139,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.0801470733165741,
+ "train_seconds": 560.0086665153503,
+ "validation": {
+ "accuracy": 0.865,
+ "examples": 5000,
+ "loss": 0.5506817626953125
+ },
+ "validation_seconds": 0.17049789428710938
+ },
+ {
+ "epoch": 140,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.07910232688056099,
+ "train_seconds": 559.9355964660645,
+ "validation": {
+ "accuracy": 0.8666,
+ "examples": 5000,
+ "loss": 0.5706710083007812
+ },
+ "validation_seconds": 0.1705949306488037
+ },
+ {
+ "epoch": 141,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.07886751850545407,
+ "train_seconds": 559.9587299823761,
+ "validation": {
+ "accuracy": 0.8652,
+ "examples": 5000,
+ "loss": 0.5789751953125
+ },
+ "validation_seconds": 0.17045021057128906
+ },
+ {
+ "epoch": 142,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.07730617905788952,
+ "train_seconds": 560.0075674057007,
+ "validation": {
+ "accuracy": 0.8618,
+ "examples": 5000,
+ "loss": 0.5797901489257813
+ },
+ "validation_seconds": 0.170365571975708
+ },
+ {
+ "epoch": 143,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.07442515372832616,
+ "train_seconds": 560.0367701053619,
+ "validation": {
+ "accuracy": 0.8684,
+ "examples": 5000,
+ "loss": 0.5737571044921875
+ },
+ "validation_seconds": 0.17035984992980957
+ },
+ {
+ "epoch": 144,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.07612440371645822,
+ "train_seconds": 559.9988217353821,
+ "validation": {
+ "accuracy": 0.8664,
+ "examples": 5000,
+ "loss": 0.5886032470703125
+ },
+ "validation_seconds": 0.17034530639648438
+ },
+ {
+ "epoch": 145,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.0729590860320462,
+ "train_seconds": 559.9792611598969,
+ "validation": {
+ "accuracy": 0.8696,
+ "examples": 5000,
+ "loss": 0.5738935424804688
+ },
+ "validation_seconds": 0.17038536071777344
+ },
+ {
+ "epoch": 146,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.07158674071364933,
+ "train_seconds": 559.9309771060944,
+ "validation": {
+ "accuracy": 0.87,
+ "examples": 5000,
+ "loss": 0.5678085327148438
+ },
+ "validation_seconds": 0.17050409317016602
+ },
+ {
+ "epoch": 147,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.07183046047290166,
+ "train_seconds": 559.9765605926514,
+ "validation": {
+ "accuracy": 0.8648,
+ "examples": 5000,
+ "loss": 0.57717275390625
+ },
+ "validation_seconds": 0.17051148414611816
+ },
+ {
+ "epoch": 148,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.07182821125984191,
+ "train_seconds": 559.96338057518,
+ "validation": {
+ "accuracy": 0.8672,
+ "examples": 5000,
+ "loss": 0.5847402465820313
+ },
+ "validation_seconds": 0.1704258918762207
+ },
+ {
+ "epoch": 149,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.0710466639969084,
+ "train_seconds": 559.9064841270447,
+ "validation": {
+ "accuracy": 0.864,
+ "examples": 5000,
+ "loss": 0.58584228515625
+ },
+ "validation_seconds": 0.17047476768493652
+ },
+ {
+ "epoch": 150,
+ "lr": 0.005000000000000001,
+ "output_lr": 0.005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.07163201610512204,
+ "train_seconds": 559.985518693924,
+ "validation": {
+ "accuracy": 0.862,
+ "examples": 5000,
+ "loss": 0.60404404296875
+ },
+ "validation_seconds": 0.1704697608947754
+ },
+ {
+ "epoch": 151,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.06554258665839831,
+ "train_seconds": 559.9303710460663,
+ "validation": {
+ "accuracy": 0.8634,
+ "examples": 5000,
+ "loss": 0.59264501953125
+ },
+ "validation_seconds": 0.17035746574401855
+ },
+ {
+ "epoch": 152,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.06308317802614635,
+ "train_seconds": 560.0243787765503,
+ "validation": {
+ "accuracy": 0.8654,
+ "examples": 5000,
+ "loss": 0.5929781616210937
+ },
+ "validation_seconds": 0.17056035995483398
+ },
+ {
+ "epoch": 153,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.0585381213956409,
+ "train_seconds": 560.1924431324005,
+ "validation": {
+ "accuracy": 0.8636,
+ "examples": 5000,
+ "loss": 0.58992890625
+ },
+ "validation_seconds": 0.17081665992736816
+ },
+ {
+ "epoch": 154,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.057798803909619646,
+ "train_seconds": 559.9080834388733,
+ "validation": {
+ "accuracy": 0.8642,
+ "examples": 5000,
+ "loss": 0.58432783203125
+ },
+ "validation_seconds": 0.17063117027282715
+ },
+ {
+ "epoch": 155,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.05727993083530002,
+ "train_seconds": 559.9425733089447,
+ "validation": {
+ "accuracy": 0.8654,
+ "examples": 5000,
+ "loss": 0.5870113159179687
+ },
+ "validation_seconds": 0.17041635513305664
+ },
+ {
+ "epoch": 156,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.058408254322740764,
+ "train_seconds": 559.9141864776611,
+ "validation": {
+ "accuracy": 0.866,
+ "examples": 5000,
+ "loss": 0.5888887939453125
+ },
+ "validation_seconds": 0.17054033279418945
+ },
+ {
+ "epoch": 157,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.05870731733375126,
+ "train_seconds": 559.904869556427,
+ "validation": {
+ "accuracy": 0.8642,
+ "examples": 5000,
+ "loss": 0.5919999145507813
+ },
+ "validation_seconds": 0.1707630157470703
+ },
+ {
+ "epoch": 158,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.056921730032894344,
+ "train_seconds": 559.9407722949982,
+ "validation": {
+ "accuracy": 0.8642,
+ "examples": 5000,
+ "loss": 0.592196240234375
+ },
+ "validation_seconds": 0.17051410675048828
+ },
+ {
+ "epoch": 159,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.057444885372453264,
+ "train_seconds": 560.0611102581024,
+ "validation": {
+ "accuracy": 0.866,
+ "examples": 5000,
+ "loss": 0.58419599609375
+ },
+ "validation_seconds": 0.1703627109527588
+ },
+ {
+ "epoch": 160,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.05547764422363705,
+ "train_seconds": 560.2958824634552,
+ "validation": {
+ "accuracy": 0.8638,
+ "examples": 5000,
+ "loss": 0.5911913940429687
+ },
+ "validation_seconds": 0.17037415504455566
+ },
+ {
+ "epoch": 161,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.055644365504052905,
+ "train_seconds": 560.2289936542511,
+ "validation": {
+ "accuracy": 0.8638,
+ "examples": 5000,
+ "loss": 0.578126025390625
+ },
+ "validation_seconds": 0.1710362434387207
+ },
+ {
+ "epoch": 162,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.05510484569966793,
+ "train_seconds": 560.129026889801,
+ "validation": {
+ "accuracy": 0.8656,
+ "examples": 5000,
+ "loss": 0.5924788818359376
+ },
+ "validation_seconds": 0.17116570472717285
+ },
+ {
+ "epoch": 163,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.05579974476231469,
+ "train_seconds": 560.2356305122375,
+ "validation": {
+ "accuracy": 0.865,
+ "examples": 5000,
+ "loss": 0.5858259155273438
+ },
+ "validation_seconds": 0.1709749698638916
+ },
+ {
+ "epoch": 164,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.05521068836980396,
+ "train_seconds": 560.1183788776398,
+ "validation": {
+ "accuracy": 0.8666,
+ "examples": 5000,
+ "loss": 0.5908378173828125
+ },
+ "validation_seconds": 0.17043590545654297
+ },
+ {
+ "epoch": 165,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.05445151421162817,
+ "train_seconds": 560.0445311069489,
+ "validation": {
+ "accuracy": 0.8652,
+ "examples": 5000,
+ "loss": 0.586676708984375
+ },
+ "validation_seconds": 0.17090559005737305
+ },
+ {
+ "epoch": 166,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.055989591648843554,
+ "train_seconds": 559.4602563381195,
+ "validation": {
+ "accuracy": 0.8666,
+ "examples": 5000,
+ "loss": 0.5927463989257813
+ },
+ "validation_seconds": 0.17037081718444824
+ },
+ {
+ "epoch": 167,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.05388901816672749,
+ "train_seconds": 558.9057474136353,
+ "validation": {
+ "accuracy": 0.8634,
+ "examples": 5000,
+ "loss": 0.5915533813476562
+ },
+ "validation_seconds": 0.17028379440307617
+ },
+ {
+ "epoch": 168,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.054578808952702416,
+ "train_seconds": 558.9692997932434,
+ "validation": {
+ "accuracy": 0.8652,
+ "examples": 5000,
+ "loss": 0.5952635986328125
+ },
+ "validation_seconds": 0.17091751098632812
+ },
+ {
+ "epoch": 169,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.054068607613775466,
+ "train_seconds": 558.840797662735,
+ "validation": {
+ "accuracy": 0.8646,
+ "examples": 5000,
+ "loss": 0.589640966796875
+ },
+ "validation_seconds": 0.1702558994293213
+ },
+ {
+ "epoch": 170,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.053210495739512975,
+ "train_seconds": 559.8298916816711,
+ "validation": {
+ "accuracy": 0.8664,
+ "examples": 5000,
+ "loss": 0.5905986572265625
+ },
+ "validation_seconds": 0.17023777961730957
+ },
+ {
+ "epoch": 171,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.052864870564805135,
+ "train_seconds": 559.8400263786316,
+ "validation": {
+ "accuracy": 0.8632,
+ "examples": 5000,
+ "loss": 0.5897403686523438
+ },
+ "validation_seconds": 0.17044568061828613
+ },
+ {
+ "epoch": 172,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.052678663874997034,
+ "train_seconds": 559.9006729125977,
+ "validation": {
+ "accuracy": 0.8634,
+ "examples": 5000,
+ "loss": 0.5942950317382812
+ },
+ "validation_seconds": 0.17066121101379395
+ },
+ {
+ "epoch": 173,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.05529955416785346,
+ "train_seconds": 560.0638835430145,
+ "validation": {
+ "accuracy": 0.8628,
+ "examples": 5000,
+ "loss": 0.5956375244140625
+ },
+ "validation_seconds": 0.17058396339416504
+ },
+ {
+ "epoch": 174,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.053033065217733384,
+ "train_seconds": 560.1647348403931,
+ "validation": {
+ "accuracy": 0.864,
+ "examples": 5000,
+ "loss": 0.5948213745117188
+ },
+ "validation_seconds": 0.17058897018432617
+ },
+ {
+ "epoch": 175,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.05241316869656245,
+ "train_seconds": 560.0912938117981,
+ "validation": {
+ "accuracy": 0.8656,
+ "examples": 5000,
+ "loss": 0.5923770812988282
+ },
+ "validation_seconds": 0.17046594619750977
+ },
+ {
+ "epoch": 176,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.05160712323255009,
+ "train_seconds": 560.2796442508698,
+ "validation": {
+ "accuracy": 0.8624,
+ "examples": 5000,
+ "loss": 0.5936520629882812
+ },
+ "validation_seconds": 0.17072343826293945
+ },
+ {
+ "epoch": 177,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.051562978375620315,
+ "train_seconds": 560.2672243118286,
+ "validation": {
+ "accuracy": 0.8648,
+ "examples": 5000,
+ "loss": 0.5952009033203125
+ },
+ "validation_seconds": 0.17060518264770508
+ },
+ {
+ "epoch": 178,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.053683729785018494,
+ "train_seconds": 560.2481393814087,
+ "validation": {
+ "accuracy": 0.8644,
+ "examples": 5000,
+ "loss": 0.5962612548828125
+ },
+ "validation_seconds": 0.1711122989654541
+ },
+ {
+ "epoch": 179,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.05219864928192563,
+ "train_seconds": 560.473662853241,
+ "validation": {
+ "accuracy": 0.8634,
+ "examples": 5000,
+ "loss": 0.5932921630859375
+ },
+ "validation_seconds": 0.1708521842956543
+ },
+ {
+ "epoch": 180,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.053021824158562554,
+ "train_seconds": 560.7065186500549,
+ "validation": {
+ "accuracy": 0.8632,
+ "examples": 5000,
+ "loss": 0.5973969970703125
+ },
+ "validation_seconds": 0.17085528373718262
+ },
+ {
+ "epoch": 181,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.05194481589926614,
+ "train_seconds": 560.7485203742981,
+ "validation": {
+ "accuracy": 0.8646,
+ "examples": 5000,
+ "loss": 0.597285546875
+ },
+ "validation_seconds": 0.17076420783996582
+ },
+ {
+ "epoch": 182,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.0498733964893553,
+ "train_seconds": 560.4968025684357,
+ "validation": {
+ "accuracy": 0.8638,
+ "examples": 5000,
+ "loss": 0.5942079223632812
+ },
+ "validation_seconds": 0.1707749366760254
+ },
+ {
+ "epoch": 183,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.05121098554680745,
+ "train_seconds": 560.602189540863,
+ "validation": {
+ "accuracy": 0.8654,
+ "examples": 5000,
+ "loss": 0.5904847534179688
+ },
+ "validation_seconds": 0.17096948623657227
+ },
+ {
+ "epoch": 184,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.0532610271414121,
+ "train_seconds": 560.6874899864197,
+ "validation": {
+ "accuracy": 0.8646,
+ "examples": 5000,
+ "loss": 0.5949594970703125
+ },
+ "validation_seconds": 0.17095732688903809
+ },
+ {
+ "epoch": 185,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.05066566181447771,
+ "train_seconds": 560.6240856647491,
+ "validation": {
+ "accuracy": 0.866,
+ "examples": 5000,
+ "loss": 0.598093505859375
+ },
+ "validation_seconds": 0.1707146167755127
+ },
+ {
+ "epoch": 186,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.050586108859380086,
+ "train_seconds": 560.5915801525116,
+ "validation": {
+ "accuracy": 0.8656,
+ "examples": 5000,
+ "loss": 0.5995073974609375
+ },
+ "validation_seconds": 0.17087697982788086
+ },
+ {
+ "epoch": 187,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.04926515153249105,
+ "train_seconds": 560.5769941806793,
+ "validation": {
+ "accuracy": 0.8646,
+ "examples": 5000,
+ "loss": 0.6008431518554688
+ },
+ "validation_seconds": 0.1711268424987793
+ },
+ {
+ "epoch": 188,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.05084701956609885,
+ "train_seconds": 560.4484589099884,
+ "validation": {
+ "accuracy": 0.8644,
+ "examples": 5000,
+ "loss": 0.5961071411132812
+ },
+ "validation_seconds": 0.1709892749786377
+ },
+ {
+ "epoch": 189,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.05086444434059991,
+ "train_seconds": 560.4446725845337,
+ "validation": {
+ "accuracy": 0.8664,
+ "examples": 5000,
+ "loss": 0.595338525390625
+ },
+ "validation_seconds": 0.17079830169677734
+ },
+ {
+ "epoch": 190,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.05069162193536758,
+ "train_seconds": 560.5038962364197,
+ "validation": {
+ "accuracy": 0.8638,
+ "examples": 5000,
+ "loss": 0.5992982055664062
+ },
+ "validation_seconds": 0.17066144943237305
+ },
+ {
+ "epoch": 191,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.05099646294448111,
+ "train_seconds": 560.6068978309631,
+ "validation": {
+ "accuracy": 0.8638,
+ "examples": 5000,
+ "loss": 0.6020646362304688
+ },
+ "validation_seconds": 0.17072200775146484
+ },
+ {
+ "epoch": 192,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.05033617229064306,
+ "train_seconds": 560.5441524982452,
+ "validation": {
+ "accuracy": 0.865,
+ "examples": 5000,
+ "loss": 0.6090272827148437
+ },
+ "validation_seconds": 0.1708674430847168
+ },
+ {
+ "epoch": 193,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.05045138215753767,
+ "train_seconds": 560.5373227596283,
+ "validation": {
+ "accuracy": 0.862,
+ "examples": 5000,
+ "loss": 0.6072682006835938
+ },
+ "validation_seconds": 0.17099285125732422
+ },
+ {
+ "epoch": 194,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.049500737303495405,
+ "train_seconds": 560.6162390708923,
+ "validation": {
+ "accuracy": 0.8644,
+ "examples": 5000,
+ "loss": 0.597724462890625
+ },
+ "validation_seconds": 0.17102837562561035
+ },
+ {
+ "epoch": 195,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.04977997776005003,
+ "train_seconds": 560.5924124717712,
+ "validation": {
+ "accuracy": 0.8648,
+ "examples": 5000,
+ "loss": 0.6030875610351563
+ },
+ "validation_seconds": 0.17075753211975098
+ },
+ {
+ "epoch": 196,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.04957166252202458,
+ "train_seconds": 560.5982604026794,
+ "validation": {
+ "accuracy": 0.8652,
+ "examples": 5000,
+ "loss": 0.604028369140625
+ },
+ "validation_seconds": 0.17163372039794922
+ },
+ {
+ "epoch": 197,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.050121073783106276,
+ "train_seconds": 560.0612788200378,
+ "validation": {
+ "accuracy": 0.864,
+ "examples": 5000,
+ "loss": 0.6068662475585938
+ },
+ "validation_seconds": 0.17065048217773438
+ },
+ {
+ "epoch": 198,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.04878396252923541,
+ "train_seconds": 559.8991024494171,
+ "validation": {
+ "accuracy": 0.8662,
+ "examples": 5000,
+ "loss": 0.6060625732421875
+ },
+ "validation_seconds": 0.1706082820892334
+ },
+ {
+ "epoch": 199,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.0506157725599077,
+ "train_seconds": 559.9076845645905,
+ "validation": {
+ "accuracy": 0.8642,
+ "examples": 5000,
+ "loss": 0.6069480834960937
+ },
+ "validation_seconds": 0.17104268074035645
+ },
+ {
+ "epoch": 200,
+ "lr": 0.0005000000000000001,
+ "output_lr": 0.0005000000000000001,
+ "train_examples": 45000,
+ "train_loss": 0.050066802283128105,
+ "train_seconds": 559.9151027202606,
+ "validation": {
+ "accuracy": 0.8654,
+ "examples": 5000,
+ "loss": 0.6110852416992187
+ },
+ "validation_seconds": 0.1712048053741455
+ }
+ ],
+ "evaluation_protocol": {
+ "split": "validation",
+ "test_evaluations": 0,
+ "test_used_for_selection": false
+ },
+ "final": {
+ "accuracy": 0.8654,
+ "evaluation_seconds": 0.1711885929107666,
+ "examples": 5000,
+ "finite": true,
+ "loss": 0.6110852416992187
+ },
+ "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": 2322300928,
+ "peak_memory_reserved_bytes": 3586129920,
+ "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": 111860.19822311401,
+ "work": {
+ "candidate_label_evaluation_presentations": 0,
+ "completed_global_epochs": 200,
+ "feedforward_example_passes": 10005000,
+ "forward_macs_per_example": 68862592,
+ "forward_parameter_count": 464154,
+ "local_target_backward_example_evaluations": 0,
+ "local_vjp_example_evaluations": 144000000,
+ "logical_task_loss_queries": 0,
+ "num_trainable_layers": 32,
+ "ordinary_training_examples": 9000000,
+ "ordinary_validation_examples": 1005000,
+ "relaxation_example_passes": 144000000,
+ "training_example_presentations": 9000000
+ }
+}