1
|
{"args": {"method": "pepita", "dataset": "mnist", "depth": 3, "width": 256, "epochs": 60, "batch_size": 64, "max_batches": 0, "eta": 0.001, "momentum": 0.9, "act": "tanh", "residual": 0, "feedback_scale": 1.0, "pepita_f_scale": 0.05, "keep_prob": 0.9, "ff_threshold": 2.0, "ep_beta": 0.5, "ep_dt": 0.5, "ep_free_steps": 20, "ep_nudge_steps": 4, "canonical_preprocess": 1, "seed": 1, "device": "cuda", "outdir": "results", "tag": "pepita_tuned_v1_mnist_pepita_w256_d3_s1"}, "method_source": {"paper": "https://proceedings.mlr.press/v162/dellaferrera22a.html", "code": "https://github.com/GiorgiaD/PEPITA", "protocol": "ERIN two-forward-pass rule; He-uniform; ReLU; F scale 0.05"}, "provenance": {"git_commit": "c8a372cae2612839bdbdafcf9c114f52cb4a6de3", "git_dirty": false}, "steps": [{"epoch_end": 0, "test_acc": 0.6576, "test_loss": 1.564298486328125, "train_loss": 1.947989583015442}, {"epoch_end": 1, "test_acc": 0.7354, "test_loss": 1.163722137451172, "train_loss": 1.1030917167663574}, {"epoch_end": 2, "test_acc": 0.7769, "test_loss": 0.9230647705078125, "train_loss": 1.266661524772644}, {"epoch_end": 3, "test_acc": 0.798, "test_loss": 0.7730442993164063, "train_loss": 0.8979678153991699}, {"epoch_end": 4, "test_acc": 0.8201, "test_loss": 0.6726371032714844, "train_loss": 1.0183836221694946}, {"epoch_end": 5, "test_acc": 0.8352, "test_loss": 0.6087283172607422, "train_loss": 0.7167853116989136}, {"epoch_end": 6, "test_acc": 0.8459, "test_loss": 0.5613793762207031, "train_loss": 0.9687842130661011}, {"epoch_end": 7, "test_acc": 0.8532, "test_loss": 0.5256933166503907, "train_loss": 0.9201305508613586}, {"epoch_end": 8, "test_acc": 0.8584, "test_loss": 0.4994579284667969, "train_loss": 0.7166524529457092}, {"epoch_end": 9, "test_acc": 0.8604, "test_loss": 0.4805103973388672, "train_loss": 0.8213995695114136}, {"epoch_end": 10, "test_acc": 0.8647, "test_loss": 0.4627833618164062, "train_loss": 1.2773122787475586}, {"epoch_end": 11, "test_acc": 0.871, "test_loss": 0.44426866455078123, "train_loss": 0.7805830240249634}, {"epoch_end": 12, "test_acc": 0.8748, "test_loss": 0.4328979797363281, "train_loss": 0.6343363523483276}, {"epoch_end": 13, "test_acc": 0.8762, "test_loss": 0.4268895965576172, "train_loss": 0.784195065498352}, {"epoch_end": 14, "test_acc": 0.8808, "test_loss": 0.4103630859375, "train_loss": 0.7097253799438477}, {"epoch_end": 15, "test_acc": 0.8836, "test_loss": 0.4051292022705078, "train_loss": 0.6262856721878052}, {"epoch_end": 16, "test_acc": 0.8844, "test_loss": 0.3944833221435547, "train_loss": 0.38616836071014404}, {"epoch_end": 17, "test_acc": 0.8804, "test_loss": 0.4045672668457031, "train_loss": 0.7168458104133606}, {"epoch_end": 18, "test_acc": 0.8888, "test_loss": 0.38385306854248047, "train_loss": 1.1206046342849731}, {"epoch_end": 19, "test_acc": 0.8854, "test_loss": 0.3848447036743164, "train_loss": 0.45324984192848206}, {"epoch_end": 20, "test_acc": 0.8852, "test_loss": 0.3811326904296875, "train_loss": 0.8075611591339111}, {"epoch_end": 21, "test_acc": 0.8876, "test_loss": 0.3744813674926758, "train_loss": 0.5532623529434204}, {"epoch_end": 22, "test_acc": 0.8953, "test_loss": 0.3603331268310547, "train_loss": 0.344644695520401}, {"epoch_end": 23, "test_acc": 0.8902, "test_loss": 0.3664552688598633, "train_loss": 1.017719030380249}, {"epoch_end": 24, "test_acc": 0.8939, "test_loss": 0.35598382720947264, "train_loss": 0.47657108306884766}, {"epoch_end": 25, "test_acc": 0.8945, "test_loss": 0.3587079376220703, "train_loss": 0.38839417695999146}, {"epoch_end": 26, "test_acc": 0.8965, "test_loss": 0.35173089294433596, "train_loss": 0.7471293807029724}, {"epoch_end": 27, "test_acc": 0.8968, "test_loss": 0.3482848815917969, "train_loss": 0.4068065583705902}, {"epoch_end": 28, "test_acc": 0.8983, "test_loss": 0.34574397888183594, "train_loss": 0.6031719446182251}, {"epoch_end": 29, "test_acc": 0.8996, "test_loss": 0.3458175537109375, "train_loss": 0.7321828007698059}, {"epoch_end": 30, "test_acc": 0.9028, "test_loss": 0.33481695709228515, "train_loss": 0.7975817322731018}, {"epoch_end": 31, "test_acc": 0.9038, "test_loss": 0.33111589584350587, "train_loss": 0.9323378801345825}, {"epoch_end": 32, "test_acc": 0.8991, "test_loss": 0.3399761764526367, "train_loss": 0.5482977032661438}, {"epoch_end": 33, "test_acc": 0.9028, "test_loss": 0.3293631561279297, "train_loss": 0.5658836364746094}, {"epoch_end": 34, "test_acc": 0.9014, "test_loss": 0.33089072647094725, "train_loss": 0.7559365630149841}, {"epoch_end": 35, "test_acc": 0.9024, "test_loss": 0.33261343688964845, "train_loss": 0.18419679999351501}, {"epoch_end": 36, "test_acc": 0.908, "test_loss": 0.3201718803405762, "train_loss": 0.43125277757644653}, {"epoch_end": 37, "test_acc": 0.9051, "test_loss": 0.32171831130981443, "train_loss": 0.333583265542984}, {"epoch_end": 38, "test_acc": 0.9072, "test_loss": 0.31539949951171875, "train_loss": 0.6099624037742615}, {"epoch_end": 39, "test_acc": 0.9037, "test_loss": 0.32065243225097656, "train_loss": 0.4864993691444397}, {"epoch_end": 40, "test_acc": 0.9084, "test_loss": 0.31198085174560547, "train_loss": 0.33445459604263306}, {"epoch_end": 41, "test_acc": 0.91, "test_loss": 0.3080244026184082, "train_loss": 0.3257055878639221}, {"epoch_end": 42, "test_acc": 0.9037, "test_loss": 0.3229133834838867, "train_loss": 0.3919094204902649}, {"epoch_end": 43, "test_acc": 0.906, "test_loss": 0.31626682891845703, "train_loss": 0.35287898778915405}, {"epoch_end": 44, "test_acc": 0.9077, "test_loss": 0.3151045967102051, "train_loss": 0.16899406909942627}, {"epoch_end": 45, "test_acc": 0.9129, "test_loss": 0.2987576126098633, "train_loss": 0.7801811695098877}, {"epoch_end": 46, "test_acc": 0.9105, "test_loss": 0.3051134475708008, "train_loss": 0.2572104036808014}, {"epoch_end": 47, "test_acc": 0.9089, "test_loss": 0.30757727508544924, "train_loss": 0.5740845203399658}, {"epoch_end": 48, "test_acc": 0.9138, "test_loss": 0.29410076599121093, "train_loss": 0.34150099754333496}, {"epoch_end": 49, "test_acc": 0.9104, "test_loss": 0.30366666870117187, "train_loss": 0.6901942491531372}, {"epoch_end": 50, "test_acc": 0.916, "test_loss": 0.2919005424499512, "train_loss": 0.5166893005371094}, {"epoch_end": 51, "test_acc": 0.9144, "test_loss": 0.29178643798828124, "train_loss": 0.8644915819168091}, {"epoch_end": 52, "test_acc": 0.9143, "test_loss": 0.293666389465332, "train_loss": 0.3433041572570801}, {"epoch_end": 53, "test_acc": 0.9128, "test_loss": 0.2942913429260254, "train_loss": 0.26649895310401917}, {"epoch_end": 54, "test_acc": 0.9107, "test_loss": 0.3001371025085449, "train_loss": 0.22972865402698517}, {"epoch_end": 55, "test_acc": 0.9157, "test_loss": 0.2882266075134277, "train_loss": 0.3866208493709564}, {"epoch_end": 56, "test_acc": 0.9172, "test_loss": 0.288204443359375, "train_loss": 0.5994341969490051}, {"epoch_end": 57, "test_acc": 0.9195, "test_loss": 0.2820064727783203, "train_loss": 0.6545388698577881}, {"epoch_end": 58, "test_acc": 0.9127, "test_loss": 0.2988055702209473, "train_loss": 0.5011070966720581}, {"epoch_end": 59, "test_acc": 0.91, "test_loss": 0.3045978668212891, "train_loss": 0.6184585690498352}], "final": {"test_acc": 0.91, "test_loss": 0.3045978668212891, "wall_s": 94.28006768226624}}
|
