diff options
Diffstat (limited to 'results/resnet_crossover/r2/resnet-r2-pepita-d56-lr0p003.json')
| -rw-r--r-- | results/resnet_crossover/r2/resnet-r2-pepita-d56-lr0p003.json | 1565 |
1 files changed, 1565 insertions, 0 deletions
diff --git a/results/resnet_crossover/r2/resnet-r2-pepita-d56-lr0p003.json b/results/resnet_crossover/r2/resnet-r2-pepita-d56-lr0p003.json new file mode 100644 index 0000000..08c2291 --- /dev/null +++ b/results/resnet_crossover/r2/resnet-r2-pepita-d56-lr0p003.json @@ -0,0 +1,1565 @@ +{ + "architecture": { + "base_width": 16, + "depth": 56, + "family": "CIFAR 6n+2 ResNet, option-A shortcuts", + "forward_parameter_count": 853018, + "normalization": "batchnorm", + "residual_scale": 1.0 + }, + "args": { + "augment_train": 1, + "batch_size": 128, + "data_dir": "/orion/u/oscarwan/scratch/sdil-data", + "depth": 56, + "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-d56-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": 2.3475769726223414, + "train_seconds": 16.284608602523804, + "validation": { + "accuracy": 0.1734, + "examples": 5000, + "loss": 2.24925087890625 + }, + "validation_seconds": 0.31821584701538086 + }, + { + "epoch": 2, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.2237265483432345, + "train_seconds": 15.681286573410034, + "validation": { + "accuracy": 0.157, + "examples": 5000, + "loss": 2.225540087890625 + }, + "validation_seconds": 0.2987985610961914 + }, + { + "epoch": 3, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.2015728338453506, + "train_seconds": 15.6518235206604, + "validation": { + "accuracy": 0.2018, + "examples": 5000, + "loss": 2.17788203125 + }, + "validation_seconds": 0.29958510398864746 + }, + { + "epoch": 4, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1849158473968506, + "train_seconds": 15.668159246444702, + "validation": { + "accuracy": 0.1936, + "examples": 5000, + "loss": 2.168885498046875 + }, + "validation_seconds": 0.29885411262512207 + }, + { + "epoch": 5, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1747074312422012, + "train_seconds": 15.680987358093262, + "validation": { + "accuracy": 0.1752, + "examples": 5000, + "loss": 2.192222607421875 + }, + "validation_seconds": 0.29928016662597656 + }, + { + "epoch": 6, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1689188510047064, + "train_seconds": 15.674059391021729, + "validation": { + "accuracy": 0.2234, + "examples": 5000, + "loss": 2.135821533203125 + }, + "validation_seconds": 0.2990145683288574 + }, + { + "epoch": 7, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.164094039789836, + "train_seconds": 15.639453887939453, + "validation": { + "accuracy": 0.2108, + "examples": 5000, + "loss": 2.135561669921875 + }, + "validation_seconds": 0.2990138530731201 + }, + { + "epoch": 8, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.163451834954156, + "train_seconds": 15.729496479034424, + "validation": { + "accuracy": 0.1988, + "examples": 5000, + "loss": 2.18637685546875 + }, + "validation_seconds": 0.29883646965026855 + }, + { + "epoch": 9, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1566063519795735, + "train_seconds": 15.675715208053589, + "validation": { + "accuracy": 0.2076, + "examples": 5000, + "loss": 2.162432666015625 + }, + "validation_seconds": 0.2995271682739258 + }, + { + "epoch": 10, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.161393331866794, + "train_seconds": 15.64734959602356, + "validation": { + "accuracy": 0.2054, + "examples": 5000, + "loss": 2.134214453125 + }, + "validation_seconds": 0.2993965148925781 + }, + { + "epoch": 11, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.157396332973904, + "train_seconds": 15.720870018005371, + "validation": { + "accuracy": 0.2262, + "examples": 5000, + "loss": 2.12485478515625 + }, + "validation_seconds": 0.2997002601623535 + }, + { + "epoch": 12, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1606002636379666, + "train_seconds": 15.66697907447815, + "validation": { + "accuracy": 0.226, + "examples": 5000, + "loss": 2.121274462890625 + }, + "validation_seconds": 0.2992849349975586 + }, + { + "epoch": 13, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1571049402448867, + "train_seconds": 15.68711233139038, + "validation": { + "accuracy": 0.216, + "examples": 5000, + "loss": 2.131235693359375 + }, + "validation_seconds": 0.2990860939025879 + }, + { + "epoch": 14, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1594723821428086, + "train_seconds": 15.663890361785889, + "validation": { + "accuracy": 0.2184, + "examples": 5000, + "loss": 2.133489697265625 + }, + "validation_seconds": 0.29947352409362793 + }, + { + "epoch": 15, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.156481715689765, + "train_seconds": 15.68401551246643, + "validation": { + "accuracy": 0.2246, + "examples": 5000, + "loss": 2.123956591796875 + }, + "validation_seconds": 0.29934096336364746 + }, + { + "epoch": 16, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1623080538855657, + "train_seconds": 15.691046476364136, + "validation": { + "accuracy": 0.2108, + "examples": 5000, + "loss": 2.1551609375 + }, + "validation_seconds": 0.2992236614227295 + }, + { + "epoch": 17, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1583885418362088, + "train_seconds": 15.685045003890991, + "validation": { + "accuracy": 0.2354, + "examples": 5000, + "loss": 2.107221484375 + }, + "validation_seconds": 0.29936671257019043 + }, + { + "epoch": 18, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.163076057688395, + "train_seconds": 15.706880807876587, + "validation": { + "accuracy": 0.2152, + "examples": 5000, + "loss": 2.1394583984375 + }, + "validation_seconds": 0.29956936836242676 + }, + { + "epoch": 19, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1610479665544298, + "train_seconds": 15.694411993026733, + "validation": { + "accuracy": 0.218, + "examples": 5000, + "loss": 2.13543974609375 + }, + "validation_seconds": 0.2992556095123291 + }, + { + "epoch": 20, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1602279196421303, + "train_seconds": 15.671279191970825, + "validation": { + "accuracy": 0.2136, + "examples": 5000, + "loss": 2.14320869140625 + }, + "validation_seconds": 0.29967164993286133 + }, + { + "epoch": 21, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.159357769478692, + "train_seconds": 15.638594388961792, + "validation": { + "accuracy": 0.2064, + "examples": 5000, + "loss": 2.159037548828125 + }, + "validation_seconds": 0.2990906238555908 + }, + { + "epoch": 22, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1632483467525905, + "train_seconds": 15.674551486968994, + "validation": { + "accuracy": 0.2232, + "examples": 5000, + "loss": 2.1262857421875 + }, + "validation_seconds": 0.2990391254425049 + }, + { + "epoch": 23, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.161040240605672, + "train_seconds": 15.671793460845947, + "validation": { + "accuracy": 0.219, + "examples": 5000, + "loss": 2.143005712890625 + }, + "validation_seconds": 0.29958534240722656 + }, + { + "epoch": 24, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.157951952955458, + "train_seconds": 15.660223960876465, + "validation": { + "accuracy": 0.2102, + "examples": 5000, + "loss": 2.15192724609375 + }, + "validation_seconds": 0.29978346824645996 + }, + { + "epoch": 25, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1627659381866455, + "train_seconds": 15.678407192230225, + "validation": { + "accuracy": 0.1964, + "examples": 5000, + "loss": 2.164074560546875 + }, + "validation_seconds": 0.29997992515563965 + }, + { + "epoch": 26, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.163635923576355, + "train_seconds": 15.671875715255737, + "validation": { + "accuracy": 0.2158, + "examples": 5000, + "loss": 2.151888037109375 + }, + "validation_seconds": 0.2996237277984619 + }, + { + "epoch": 27, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1635207350413004, + "train_seconds": 15.662642240524292, + "validation": { + "accuracy": 0.2136, + "examples": 5000, + "loss": 2.125229248046875 + }, + "validation_seconds": 0.2993924617767334 + }, + { + "epoch": 28, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.162565950393677, + "train_seconds": 15.636061668395996, + "validation": { + "accuracy": 0.1964, + "examples": 5000, + "loss": 2.16964326171875 + }, + "validation_seconds": 0.29975390434265137 + }, + { + "epoch": 29, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.163283192952474, + "train_seconds": 15.683935642242432, + "validation": { + "accuracy": 0.2076, + "examples": 5000, + "loss": 2.151294189453125 + }, + "validation_seconds": 0.29951977729797363 + }, + { + "epoch": 30, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.161205266825358, + "train_seconds": 15.69214129447937, + "validation": { + "accuracy": 0.2086, + "examples": 5000, + "loss": 2.14878486328125 + }, + "validation_seconds": 0.2994678020477295 + }, + { + "epoch": 31, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1700854666815865, + "train_seconds": 15.638405084609985, + "validation": { + "accuracy": 0.21, + "examples": 5000, + "loss": 2.168007275390625 + }, + "validation_seconds": 0.2993042469024658 + }, + { + "epoch": 32, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1673946551005048, + "train_seconds": 15.769407749176025, + "validation": { + "accuracy": 0.1872, + "examples": 5000, + "loss": 2.15320751953125 + }, + "validation_seconds": 0.2998044490814209 + }, + { + "epoch": 33, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.168034863026937, + "train_seconds": 15.698111772537231, + "validation": { + "accuracy": 0.2076, + "examples": 5000, + "loss": 2.149951318359375 + }, + "validation_seconds": 0.2999296188354492 + }, + { + "epoch": 34, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1725848429785835, + "train_seconds": 15.684096813201904, + "validation": { + "accuracy": 0.2216, + "examples": 5000, + "loss": 2.138531201171875 + }, + "validation_seconds": 0.2992548942565918 + }, + { + "epoch": 35, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.169652655537923, + "train_seconds": 15.643929243087769, + "validation": { + "accuracy": 0.2218, + "examples": 5000, + "loss": 2.138718603515625 + }, + "validation_seconds": 0.29943323135375977 + }, + { + "epoch": 36, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1699724278767905, + "train_seconds": 15.66665768623352, + "validation": { + "accuracy": 0.2084, + "examples": 5000, + "loss": 2.14305009765625 + }, + "validation_seconds": 0.29984354972839355 + }, + { + "epoch": 37, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1717837972429064, + "train_seconds": 15.693549633026123, + "validation": { + "accuracy": 0.2022, + "examples": 5000, + "loss": 2.137519580078125 + }, + "validation_seconds": 0.2995481491088867 + }, + { + "epoch": 38, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1724463059743244, + "train_seconds": 15.658748149871826, + "validation": { + "accuracy": 0.2186, + "examples": 5000, + "loss": 2.131788037109375 + }, + "validation_seconds": 0.29953789710998535 + }, + { + "epoch": 39, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.168097107272678, + "train_seconds": 15.651520252227783, + "validation": { + "accuracy": 0.2226, + "examples": 5000, + "loss": 2.13143837890625 + }, + "validation_seconds": 0.29965829849243164 + }, + { + "epoch": 40, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.170875356547038, + "train_seconds": 15.668555498123169, + "validation": { + "accuracy": 0.222, + "examples": 5000, + "loss": 2.124761669921875 + }, + "validation_seconds": 0.2994194030761719 + }, + { + "epoch": 41, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1716907165103487, + "train_seconds": 15.660097360610962, + "validation": { + "accuracy": 0.203, + "examples": 5000, + "loss": 2.168586181640625 + }, + "validation_seconds": 0.29941606521606445 + }, + { + "epoch": 42, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1702115387386747, + "train_seconds": 15.656601190567017, + "validation": { + "accuracy": 0.2056, + "examples": 5000, + "loss": 2.140987890625 + }, + "validation_seconds": 0.299454927444458 + }, + { + "epoch": 43, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1729301143222384, + "train_seconds": 15.676163911819458, + "validation": { + "accuracy": 0.1914, + "examples": 5000, + "loss": 2.18998251953125 + }, + "validation_seconds": 0.299652099609375 + }, + { + "epoch": 44, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1765141748216417, + "train_seconds": 15.67884373664856, + "validation": { + "accuracy": 0.209, + "examples": 5000, + "loss": 2.13410859375 + }, + "validation_seconds": 0.299485445022583 + }, + { + "epoch": 45, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.181572889751858, + "train_seconds": 15.63332986831665, + "validation": { + "accuracy": 0.2036, + "examples": 5000, + "loss": 2.15123642578125 + }, + "validation_seconds": 0.29934072494506836 + }, + { + "epoch": 46, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.175901057518853, + "train_seconds": 15.680678606033325, + "validation": { + "accuracy": 0.2054, + "examples": 5000, + "loss": 2.168797021484375 + }, + "validation_seconds": 0.29965710639953613 + }, + { + "epoch": 47, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1773429202185737, + "train_seconds": 15.656840324401855, + "validation": { + "accuracy": 0.2014, + "examples": 5000, + "loss": 2.13987265625 + }, + "validation_seconds": 0.2994050979614258 + }, + { + "epoch": 48, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1835591296725805, + "train_seconds": 15.684873819351196, + "validation": { + "accuracy": 0.2052, + "examples": 5000, + "loss": 2.15433330078125 + }, + "validation_seconds": 0.299971342086792 + }, + { + "epoch": 49, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.188736485120985, + "train_seconds": 15.647130250930786, + "validation": { + "accuracy": 0.2238, + "examples": 5000, + "loss": 2.130195849609375 + }, + "validation_seconds": 0.2995433807373047 + }, + { + "epoch": 50, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.180174428134494, + "train_seconds": 15.650701761245728, + "validation": { + "accuracy": 0.1848, + "examples": 5000, + "loss": 2.22315712890625 + }, + "validation_seconds": 0.2997779846191406 + }, + { + "epoch": 51, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.182067480383979, + "train_seconds": 15.674840927124023, + "validation": { + "accuracy": 0.1828, + "examples": 5000, + "loss": 2.165380908203125 + }, + "validation_seconds": 0.2998027801513672 + }, + { + "epoch": 52, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.179020531166924, + "train_seconds": 15.649744749069214, + "validation": { + "accuracy": 0.1946, + "examples": 5000, + "loss": 2.224890478515625 + }, + "validation_seconds": 0.29945802688598633 + }, + { + "epoch": 53, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.182496388244629, + "train_seconds": 15.68449592590332, + "validation": { + "accuracy": 0.2044, + "examples": 5000, + "loss": 2.159840478515625 + }, + "validation_seconds": 0.299649715423584 + }, + { + "epoch": 54, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.176760282389323, + "train_seconds": 15.676718711853027, + "validation": { + "accuracy": 0.2154, + "examples": 5000, + "loss": 2.154713623046875 + }, + "validation_seconds": 0.30002808570861816 + }, + { + "epoch": 55, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.181396118672689, + "train_seconds": 15.64739179611206, + "validation": { + "accuracy": 0.2118, + "examples": 5000, + "loss": 2.15625439453125 + }, + "validation_seconds": 0.299640417098999 + }, + { + "epoch": 56, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1928697071499297, + "train_seconds": 15.674021244049072, + "validation": { + "accuracy": 0.1838, + "examples": 5000, + "loss": 2.1972078125 + }, + "validation_seconds": 0.29950594902038574 + }, + { + "epoch": 57, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1952037180158825, + "train_seconds": 15.704137325286865, + "validation": { + "accuracy": 0.2096, + "examples": 5000, + "loss": 2.15042080078125 + }, + "validation_seconds": 0.29971742630004883 + }, + { + "epoch": 58, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.187411170917087, + "train_seconds": 15.676226615905762, + "validation": { + "accuracy": 0.2104, + "examples": 5000, + "loss": 2.142286865234375 + }, + "validation_seconds": 0.29947590827941895 + }, + { + "epoch": 59, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.1929678960164387, + "train_seconds": 15.638705253601074, + "validation": { + "accuracy": 0.1936, + "examples": 5000, + "loss": 2.195626708984375 + }, + "validation_seconds": 0.2998230457305908 + }, + { + "epoch": 60, + "lr": 0.003, + "output_lr": 0.003, + "train_examples": 45000, + "train_loss": 2.187026701990763, + "train_seconds": 15.711337566375732, + "validation": { + "accuracy": 0.2156, + "examples": 5000, + "loss": 2.13201728515625 + }, + "validation_seconds": 0.3002026081085205 + }, + { + "epoch": 61, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.153464202711317, + "train_seconds": 15.665104866027832, + "validation": { + "accuracy": 0.2306, + "examples": 5000, + "loss": 2.118577685546875 + }, + "validation_seconds": 0.29940152168273926 + }, + { + "epoch": 62, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.153363601769341, + "train_seconds": 15.632834196090698, + "validation": { + "accuracy": 0.2176, + "examples": 5000, + "loss": 2.125661083984375 + }, + "validation_seconds": 0.29975414276123047 + }, + { + "epoch": 63, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.1547847679562038, + "train_seconds": 15.688811302185059, + "validation": { + "accuracy": 0.2246, + "examples": 5000, + "loss": 2.1178529296875 + }, + "validation_seconds": 0.2998015880584717 + }, + { + "epoch": 64, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.1549198490990533, + "train_seconds": 15.655038595199585, + "validation": { + "accuracy": 0.2172, + "examples": 5000, + "loss": 2.1274083984375 + }, + "validation_seconds": 0.2996184825897217 + }, + { + "epoch": 65, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.1497549418555364, + "train_seconds": 15.679463386535645, + "validation": { + "accuracy": 0.2244, + "examples": 5000, + "loss": 2.1182056640625 + }, + "validation_seconds": 0.29967188835144043 + }, + { + "epoch": 66, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.1515708785163032, + "train_seconds": 15.667905807495117, + "validation": { + "accuracy": 0.2192, + "examples": 5000, + "loss": 2.13347197265625 + }, + "validation_seconds": 0.2998011112213135 + }, + { + "epoch": 67, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.150513512632582, + "train_seconds": 15.695557594299316, + "validation": { + "accuracy": 0.232, + "examples": 5000, + "loss": 2.12719873046875 + }, + "validation_seconds": 0.2998483180999756 + }, + { + "epoch": 68, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.156382405641344, + "train_seconds": 15.711182832717896, + "validation": { + "accuracy": 0.2216, + "examples": 5000, + "loss": 2.131969189453125 + }, + "validation_seconds": 0.29973673820495605 + }, + { + "epoch": 69, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.1527923811594647, + "train_seconds": 15.65618896484375, + "validation": { + "accuracy": 0.2242, + "examples": 5000, + "loss": 2.128789599609375 + }, + "validation_seconds": 0.2994709014892578 + }, + { + "epoch": 70, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.153493192715115, + "train_seconds": 15.669008016586304, + "validation": { + "accuracy": 0.2196, + "examples": 5000, + "loss": 2.13042177734375 + }, + "validation_seconds": 0.29961538314819336 + }, + { + "epoch": 71, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.150450718943278, + "train_seconds": 15.71059536933899, + "validation": { + "accuracy": 0.2334, + "examples": 5000, + "loss": 2.118706591796875 + }, + "validation_seconds": 0.2995724678039551 + }, + { + "epoch": 72, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.152825714450412, + "train_seconds": 15.683857917785645, + "validation": { + "accuracy": 0.2224, + "examples": 5000, + "loss": 2.129543701171875 + }, + "validation_seconds": 0.2995285987854004 + }, + { + "epoch": 73, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.1515955153571236, + "train_seconds": 15.6649649143219, + "validation": { + "accuracy": 0.2194, + "examples": 5000, + "loss": 2.12808408203125 + }, + "validation_seconds": 0.299257755279541 + }, + { + "epoch": 74, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.1496700807359486, + "train_seconds": 15.722174882888794, + "validation": { + "accuracy": 0.216, + "examples": 5000, + "loss": 2.12822939453125 + }, + "validation_seconds": 0.29958152770996094 + }, + { + "epoch": 75, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.1503506173027884, + "train_seconds": 15.714526891708374, + "validation": { + "accuracy": 0.224, + "examples": 5000, + "loss": 2.119987060546875 + }, + "validation_seconds": 0.2995455265045166 + }, + { + "epoch": 76, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.1522617220136855, + "train_seconds": 15.697211027145386, + "validation": { + "accuracy": 0.2248, + "examples": 5000, + "loss": 2.12939140625 + }, + "validation_seconds": 0.2992062568664551 + }, + { + "epoch": 77, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.154192639456855, + "train_seconds": 15.73182463645935, + "validation": { + "accuracy": 0.2216, + "examples": 5000, + "loss": 2.131347509765625 + }, + "validation_seconds": 0.29962968826293945 + }, + { + "epoch": 78, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.1487903494940865, + "train_seconds": 15.704017400741577, + "validation": { + "accuracy": 0.22, + "examples": 5000, + "loss": 2.124372314453125 + }, + "validation_seconds": 0.2995433807373047 + }, + { + "epoch": 79, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.1514638863457574, + "train_seconds": 15.68510890007019, + "validation": { + "accuracy": 0.2244, + "examples": 5000, + "loss": 2.12896591796875 + }, + "validation_seconds": 0.2998325824737549 + }, + { + "epoch": 80, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.1498017944335936, + "train_seconds": 15.664867162704468, + "validation": { + "accuracy": 0.2254, + "examples": 5000, + "loss": 2.124163525390625 + }, + "validation_seconds": 0.2989673614501953 + }, + { + "epoch": 81, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.1490917604658337, + "train_seconds": 15.712203979492188, + "validation": { + "accuracy": 0.2296, + "examples": 5000, + "loss": 2.11083935546875 + }, + "validation_seconds": 0.29970264434814453 + }, + { + "epoch": 82, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.1534319928487142, + "train_seconds": 15.689087390899658, + "validation": { + "accuracy": 0.2166, + "examples": 5000, + "loss": 2.13138134765625 + }, + "validation_seconds": 0.2988622188568115 + }, + { + "epoch": 83, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.150700324376424, + "train_seconds": 15.66288447380066, + "validation": { + "accuracy": 0.2264, + "examples": 5000, + "loss": 2.124049951171875 + }, + "validation_seconds": 0.2999234199523926 + }, + { + "epoch": 84, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.1545015112559, + "train_seconds": 15.717640161514282, + "validation": { + "accuracy": 0.2256, + "examples": 5000, + "loss": 2.120006201171875 + }, + "validation_seconds": 0.2995924949645996 + }, + { + "epoch": 85, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.1520926057603624, + "train_seconds": 15.6857271194458, + "validation": { + "accuracy": 0.217, + "examples": 5000, + "loss": 2.123595703125 + }, + "validation_seconds": 0.29950642585754395 + }, + { + "epoch": 86, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.1539180565304226, + "train_seconds": 15.702425003051758, + "validation": { + "accuracy": 0.2266, + "examples": 5000, + "loss": 2.127067041015625 + }, + "validation_seconds": 0.299513578414917 + }, + { + "epoch": 87, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.1541437976413302, + "train_seconds": 15.651525497436523, + "validation": { + "accuracy": 0.2174, + "examples": 5000, + "loss": 2.12785693359375 + }, + "validation_seconds": 0.2997150421142578 + }, + { + "epoch": 88, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.150302347056071, + "train_seconds": 15.696409463882446, + "validation": { + "accuracy": 0.221, + "examples": 5000, + "loss": 2.1273095703125 + }, + "validation_seconds": 0.2993946075439453 + }, + { + "epoch": 89, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.152333891635471, + "train_seconds": 15.691116094589233, + "validation": { + "accuracy": 0.2258, + "examples": 5000, + "loss": 2.12480576171875 + }, + "validation_seconds": 0.29930853843688965 + }, + { + "epoch": 90, + "lr": 0.00030000000000000003, + "output_lr": 0.00030000000000000003, + "train_examples": 45000, + "train_loss": 2.1507328274620905, + "train_seconds": 15.65622353553772, + "validation": { + "accuracy": 0.2304, + "examples": 5000, + "loss": 2.118491552734375 + }, + "validation_seconds": 0.2990148067474365 + }, + { + "epoch": 91, + "lr": 3.0000000000000004e-05, + "output_lr": 3.0000000000000004e-05, + "train_examples": 45000, + "train_loss": 2.1486516242980955, + "train_seconds": 15.699949741363525, + "validation": { + "accuracy": 0.2204, + "examples": 5000, + "loss": 2.12403779296875 + }, + "validation_seconds": 0.2998237609863281 + }, + { + "epoch": 92, + "lr": 3.0000000000000004e-05, + "output_lr": 3.0000000000000004e-05, + "train_examples": 45000, + "train_loss": 2.1497891513400607, + "train_seconds": 15.700400590896606, + "validation": { + "accuracy": 0.2206, + "examples": 5000, + "loss": 2.1224546875 + }, + "validation_seconds": 0.2998065948486328 + }, + { + "epoch": 93, + "lr": 3.0000000000000004e-05, + "output_lr": 3.0000000000000004e-05, + "train_examples": 45000, + "train_loss": 2.1481778487735324, + "train_seconds": 15.68268871307373, + "validation": { + "accuracy": 0.2306, + "examples": 5000, + "loss": 2.119259814453125 + }, + "validation_seconds": 0.29954075813293457 + }, + { + "epoch": 94, + "lr": 3.0000000000000004e-05, + "output_lr": 3.0000000000000004e-05, + "train_examples": 45000, + "train_loss": 2.1484259122212728, + "train_seconds": 15.650254726409912, + "validation": { + "accuracy": 0.2268, + "examples": 5000, + "loss": 2.122284521484375 + }, + "validation_seconds": 0.2995462417602539 + }, + { + "epoch": 95, + "lr": 3.0000000000000004e-05, + "output_lr": 3.0000000000000004e-05, + "train_examples": 45000, + "train_loss": 2.1485114953782825, + "train_seconds": 15.716254949569702, + "validation": { + "accuracy": 0.2312, + "examples": 5000, + "loss": 2.1186466796875 + }, + "validation_seconds": 0.2994718551635742 + }, + { + "epoch": 96, + "lr": 3.0000000000000004e-05, + "output_lr": 3.0000000000000004e-05, + "train_examples": 45000, + "train_loss": 2.14853645807902, + "train_seconds": 15.704525470733643, + "validation": { + "accuracy": 0.2256, + "examples": 5000, + "loss": 2.120378759765625 + }, + "validation_seconds": 0.2997438907623291 + }, + { + "epoch": 97, + "lr": 3.0000000000000004e-05, + "output_lr": 3.0000000000000004e-05, + "train_examples": 45000, + "train_loss": 2.1462308433532713, + "train_seconds": 15.661826848983765, + "validation": { + "accuracy": 0.2294, + "examples": 5000, + "loss": 2.120892578125 + }, + "validation_seconds": 0.2998523712158203 + }, + { + "epoch": 98, + "lr": 3.0000000000000004e-05, + "output_lr": 3.0000000000000004e-05, + "train_examples": 45000, + "train_loss": 2.149182392586602, + "train_seconds": 15.695777177810669, + "validation": { + "accuracy": 0.229, + "examples": 5000, + "loss": 2.1194109375 + }, + "validation_seconds": 0.29935646057128906 + }, + { + "epoch": 99, + "lr": 3.0000000000000004e-05, + "output_lr": 3.0000000000000004e-05, + "train_examples": 45000, + "train_loss": 2.149133202785916, + "train_seconds": 15.691457509994507, + "validation": { + "accuracy": 0.2268, + "examples": 5000, + "loss": 2.116005859375 + }, + "validation_seconds": 0.29990196228027344 + }, + { + "epoch": 100, + "lr": 3.0000000000000004e-05, + "output_lr": 3.0000000000000004e-05, + "train_examples": 45000, + "train_loss": 2.1499680579715306, + "train_seconds": 15.672620058059692, + "validation": { + "accuracy": 0.228, + "examples": 5000, + "loss": 2.124036181640625 + }, + "validation_seconds": 0.29943299293518066 + } + ], + "evaluation_protocol": { + "split": "validation", + "test_evaluations": 0, + "test_used_for_selection": false + }, + "final": { + "accuracy": 0.228, + "evaluation_seconds": 0.2996077537536621, + "examples": 5000, + "finite": true, + "loss": 2.124036181640625 + }, + "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": 3249364480, + "peak_memory_reserved_bytes": 4500488192, + "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": 1598.730625152588, + "work": { + "candidate_label_evaluation_presentations": 0, + "completed_global_epochs": 100, + "feedforward_example_passes": 14005000, + "forward_macs_per_example": 125485696, + "forward_parameter_count": 853018, + "local_target_backward_example_evaluations": 0, + "local_vjp_example_evaluations": 0, + "logical_task_loss_queries": 0, + "num_trainable_layers": 56, + "ordinary_training_examples": 4500000, + "ordinary_validation_examples": 505000, + "relaxation_example_passes": 0, + "training_example_presentations": 9000000 + } +} |
