| { | |
| "best_metric": 0.0008732560090720654, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-150", | |
| "epoch": 0.32397408207343414, | |
| "eval_steps": 50, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0021598272138228943, | |
| "grad_norm": 1.0908081531524658, | |
| "learning_rate": 1.0060000000000002e-05, | |
| "loss": 0.0443, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0021598272138228943, | |
| "eval_loss": 0.14869527518749237, | |
| "eval_runtime": 38.1641, | |
| "eval_samples_per_second": 5.11, | |
| "eval_steps_per_second": 1.284, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.004319654427645789, | |
| "grad_norm": 0.6184279322624207, | |
| "learning_rate": 2.0120000000000004e-05, | |
| "loss": 0.0345, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.0064794816414686825, | |
| "grad_norm": 0.3461187183856964, | |
| "learning_rate": 3.018e-05, | |
| "loss": 0.0337, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.008639308855291577, | |
| "grad_norm": 0.22748076915740967, | |
| "learning_rate": 4.024000000000001e-05, | |
| "loss": 0.0299, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.01079913606911447, | |
| "grad_norm": 0.11859659850597382, | |
| "learning_rate": 5.03e-05, | |
| "loss": 0.0259, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.012958963282937365, | |
| "grad_norm": 0.1670089215040207, | |
| "learning_rate": 6.036e-05, | |
| "loss": 0.0249, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.01511879049676026, | |
| "grad_norm": 0.1212053894996643, | |
| "learning_rate": 7.042e-05, | |
| "loss": 0.0189, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.017278617710583154, | |
| "grad_norm": 0.08793988823890686, | |
| "learning_rate": 8.048000000000002e-05, | |
| "loss": 0.0179, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.019438444924406047, | |
| "grad_norm": 0.2328803539276123, | |
| "learning_rate": 9.054000000000001e-05, | |
| "loss": 0.027, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.02159827213822894, | |
| "grad_norm": 0.17274922132492065, | |
| "learning_rate": 0.0001006, | |
| "loss": 0.0257, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.023758099352051837, | |
| "grad_norm": 0.28910505771636963, | |
| "learning_rate": 0.00010007052631578948, | |
| "loss": 0.0171, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.02591792656587473, | |
| "grad_norm": 0.13060903549194336, | |
| "learning_rate": 9.954105263157895e-05, | |
| "loss": 0.0085, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.028077753779697623, | |
| "grad_norm": 0.25134187936782837, | |
| "learning_rate": 9.901157894736842e-05, | |
| "loss": 0.0121, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.03023758099352052, | |
| "grad_norm": 0.1418023407459259, | |
| "learning_rate": 9.84821052631579e-05, | |
| "loss": 0.0035, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.032397408207343416, | |
| "grad_norm": 0.2569674253463745, | |
| "learning_rate": 9.795263157894737e-05, | |
| "loss": 0.0033, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.03455723542116631, | |
| "grad_norm": 0.46587738394737244, | |
| "learning_rate": 9.742315789473686e-05, | |
| "loss": 0.008, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.0367170626349892, | |
| "grad_norm": 0.10296951234340668, | |
| "learning_rate": 9.689368421052633e-05, | |
| "loss": 0.0015, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.038876889848812095, | |
| "grad_norm": 0.052655696868896484, | |
| "learning_rate": 9.63642105263158e-05, | |
| "loss": 0.0011, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.04103671706263499, | |
| "grad_norm": 0.1694544404745102, | |
| "learning_rate": 9.583473684210527e-05, | |
| "loss": 0.0028, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.04319654427645788, | |
| "grad_norm": 0.028250250965356827, | |
| "learning_rate": 9.530526315789474e-05, | |
| "loss": 0.0005, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.04535637149028078, | |
| "grad_norm": 0.30077144503593445, | |
| "learning_rate": 9.477578947368422e-05, | |
| "loss": 0.0043, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.047516198704103674, | |
| "grad_norm": 0.3806341886520386, | |
| "learning_rate": 9.424631578947369e-05, | |
| "loss": 0.0021, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.04967602591792657, | |
| "grad_norm": 0.3099500238895416, | |
| "learning_rate": 9.371684210526316e-05, | |
| "loss": 0.0041, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.05183585313174946, | |
| "grad_norm": 0.3820614218711853, | |
| "learning_rate": 9.318736842105263e-05, | |
| "loss": 0.0065, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.05399568034557235, | |
| "grad_norm": 0.3208763003349304, | |
| "learning_rate": 9.26578947368421e-05, | |
| "loss": 0.0044, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.056155507559395246, | |
| "grad_norm": 0.20953547954559326, | |
| "learning_rate": 9.212842105263159e-05, | |
| "loss": 0.0018, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.058315334773218146, | |
| "grad_norm": 0.11007744818925858, | |
| "learning_rate": 9.159894736842107e-05, | |
| "loss": 0.0016, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.06047516198704104, | |
| "grad_norm": 0.005611012689769268, | |
| "learning_rate": 9.106947368421054e-05, | |
| "loss": 0.0004, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.06263498920086392, | |
| "grad_norm": 0.3053535223007202, | |
| "learning_rate": 9.054000000000001e-05, | |
| "loss": 0.0068, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.06479481641468683, | |
| "grad_norm": 0.009484589099884033, | |
| "learning_rate": 9.001052631578948e-05, | |
| "loss": 0.0002, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.06695464362850972, | |
| "grad_norm": 0.025772912427783012, | |
| "learning_rate": 8.948105263157895e-05, | |
| "loss": 0.0008, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.06911447084233262, | |
| "grad_norm": 0.17834728956222534, | |
| "learning_rate": 8.895157894736842e-05, | |
| "loss": 0.0017, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.07127429805615551, | |
| "grad_norm": 0.4382934868335724, | |
| "learning_rate": 8.842210526315789e-05, | |
| "loss": 0.0108, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.0734341252699784, | |
| "grad_norm": 0.029989449307322502, | |
| "learning_rate": 8.789263157894738e-05, | |
| "loss": 0.0003, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.0755939524838013, | |
| "grad_norm": 0.296584814786911, | |
| "learning_rate": 8.736315789473685e-05, | |
| "loss": 0.0062, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.07775377969762419, | |
| "grad_norm": 0.004909132141619921, | |
| "learning_rate": 8.683368421052632e-05, | |
| "loss": 0.0003, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.07991360691144708, | |
| "grad_norm": 0.03281238302588463, | |
| "learning_rate": 8.63042105263158e-05, | |
| "loss": 0.0008, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.08207343412526998, | |
| "grad_norm": 0.2719196379184723, | |
| "learning_rate": 8.577473684210527e-05, | |
| "loss": 0.0061, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.08423326133909287, | |
| "grad_norm": 0.05495695397257805, | |
| "learning_rate": 8.524526315789474e-05, | |
| "loss": 0.0007, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.08639308855291576, | |
| "grad_norm": 0.22600561380386353, | |
| "learning_rate": 8.471578947368421e-05, | |
| "loss": 0.0053, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.08855291576673865, | |
| "grad_norm": 0.08108193427324295, | |
| "learning_rate": 8.41863157894737e-05, | |
| "loss": 0.0013, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.09071274298056156, | |
| "grad_norm": 0.15901321172714233, | |
| "learning_rate": 8.365684210526317e-05, | |
| "loss": 0.0034, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.09287257019438445, | |
| "grad_norm": 0.22143931686878204, | |
| "learning_rate": 8.312736842105264e-05, | |
| "loss": 0.0014, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.09503239740820735, | |
| "grad_norm": 0.18730013072490692, | |
| "learning_rate": 8.259789473684211e-05, | |
| "loss": 0.0019, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.09719222462203024, | |
| "grad_norm": 0.19464093446731567, | |
| "learning_rate": 8.206842105263158e-05, | |
| "loss": 0.0011, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.09935205183585313, | |
| "grad_norm": 0.09799759089946747, | |
| "learning_rate": 8.153894736842105e-05, | |
| "loss": 0.001, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.10151187904967603, | |
| "grad_norm": 0.5721032023429871, | |
| "learning_rate": 8.100947368421053e-05, | |
| "loss": 0.0025, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.10367170626349892, | |
| "grad_norm": 0.11062470078468323, | |
| "learning_rate": 8.048000000000002e-05, | |
| "loss": 0.0007, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.10583153347732181, | |
| "grad_norm": 0.36123791337013245, | |
| "learning_rate": 7.995052631578949e-05, | |
| "loss": 0.0018, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.1079913606911447, | |
| "grad_norm": 0.17662936449050903, | |
| "learning_rate": 7.942105263157896e-05, | |
| "loss": 0.0011, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.1079913606911447, | |
| "eval_loss": 0.003911234904080629, | |
| "eval_runtime": 38.5635, | |
| "eval_samples_per_second": 5.057, | |
| "eval_steps_per_second": 1.271, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.1101511879049676, | |
| "grad_norm": 0.09929417818784714, | |
| "learning_rate": 7.889157894736843e-05, | |
| "loss": 0.0122, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.11231101511879049, | |
| "grad_norm": 0.0626210868358612, | |
| "learning_rate": 7.83621052631579e-05, | |
| "loss": 0.0117, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.11447084233261338, | |
| "grad_norm": 0.08357356488704681, | |
| "learning_rate": 7.783263157894737e-05, | |
| "loss": 0.013, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.11663066954643629, | |
| "grad_norm": 0.35985463857650757, | |
| "learning_rate": 7.730315789473684e-05, | |
| "loss": 0.0162, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.11879049676025918, | |
| "grad_norm": 0.06007225811481476, | |
| "learning_rate": 7.677368421052632e-05, | |
| "loss": 0.0126, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.12095032397408208, | |
| "grad_norm": 0.07496761530637741, | |
| "learning_rate": 7.624421052631579e-05, | |
| "loss": 0.0099, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.12311015118790497, | |
| "grad_norm": 0.05158822238445282, | |
| "learning_rate": 7.571473684210526e-05, | |
| "loss": 0.0044, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.12526997840172785, | |
| "grad_norm": 0.04859482869505882, | |
| "learning_rate": 7.518526315789475e-05, | |
| "loss": 0.0053, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.12742980561555076, | |
| "grad_norm": 0.08607001602649689, | |
| "learning_rate": 7.465578947368422e-05, | |
| "loss": 0.0016, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.12958963282937366, | |
| "grad_norm": 0.19171972572803497, | |
| "learning_rate": 7.412631578947369e-05, | |
| "loss": 0.0017, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.13174946004319654, | |
| "grad_norm": 0.1458485871553421, | |
| "learning_rate": 7.359684210526317e-05, | |
| "loss": 0.0018, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.13390928725701945, | |
| "grad_norm": 0.274966835975647, | |
| "learning_rate": 7.306736842105264e-05, | |
| "loss": 0.0027, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.13606911447084233, | |
| "grad_norm": 0.051928453147411346, | |
| "learning_rate": 7.253789473684211e-05, | |
| "loss": 0.001, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.13822894168466524, | |
| "grad_norm": 0.2043527066707611, | |
| "learning_rate": 7.200842105263158e-05, | |
| "loss": 0.0018, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.14038876889848811, | |
| "grad_norm": 0.0076795765198767185, | |
| "learning_rate": 7.147894736842105e-05, | |
| "loss": 0.0002, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.14254859611231102, | |
| "grad_norm": 0.011842038482427597, | |
| "learning_rate": 7.094947368421052e-05, | |
| "loss": 0.0004, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.1447084233261339, | |
| "grad_norm": 0.0678977221250534, | |
| "learning_rate": 7.042e-05, | |
| "loss": 0.0009, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.1468682505399568, | |
| "grad_norm": 0.5634580254554749, | |
| "learning_rate": 6.989052631578948e-05, | |
| "loss": 0.0021, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.1490280777537797, | |
| "grad_norm": 0.23194696009159088, | |
| "learning_rate": 6.936105263157896e-05, | |
| "loss": 0.0021, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.1511879049676026, | |
| "grad_norm": 0.0037620756775140762, | |
| "learning_rate": 6.883157894736843e-05, | |
| "loss": 0.0002, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.15334773218142547, | |
| "grad_norm": 0.14211751520633698, | |
| "learning_rate": 6.83021052631579e-05, | |
| "loss": 0.0039, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.15550755939524838, | |
| "grad_norm": 0.1930842101573944, | |
| "learning_rate": 6.777263157894737e-05, | |
| "loss": 0.0005, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.15766738660907129, | |
| "grad_norm": 0.004353968892246485, | |
| "learning_rate": 6.724315789473684e-05, | |
| "loss": 0.0002, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.15982721382289417, | |
| "grad_norm": 0.20285466313362122, | |
| "learning_rate": 6.671368421052631e-05, | |
| "loss": 0.0013, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.16198704103671707, | |
| "grad_norm": 0.0263319481164217, | |
| "learning_rate": 6.61842105263158e-05, | |
| "loss": 0.0029, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.16414686825053995, | |
| "grad_norm": 0.013903641141951084, | |
| "learning_rate": 6.565473684210527e-05, | |
| "loss": 0.0002, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.16630669546436286, | |
| "grad_norm": 0.02208566665649414, | |
| "learning_rate": 6.512526315789474e-05, | |
| "loss": 0.0005, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.16846652267818574, | |
| "grad_norm": 0.430324912071228, | |
| "learning_rate": 6.459578947368421e-05, | |
| "loss": 0.0043, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.17062634989200864, | |
| "grad_norm": 0.48172298073768616, | |
| "learning_rate": 6.406631578947369e-05, | |
| "loss": 0.0052, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.17278617710583152, | |
| "grad_norm": 0.02294292487204075, | |
| "learning_rate": 6.353684210526316e-05, | |
| "loss": 0.0003, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.17494600431965443, | |
| "grad_norm": 0.008779804222285748, | |
| "learning_rate": 6.300736842105263e-05, | |
| "loss": 0.0003, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.1771058315334773, | |
| "grad_norm": 0.025301771238446236, | |
| "learning_rate": 6.247789473684212e-05, | |
| "loss": 0.0003, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.17926565874730022, | |
| "grad_norm": 0.010768801905214787, | |
| "learning_rate": 6.194842105263159e-05, | |
| "loss": 0.0002, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.18142548596112312, | |
| "grad_norm": 0.01551789976656437, | |
| "learning_rate": 6.141894736842106e-05, | |
| "loss": 0.0003, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.183585313174946, | |
| "grad_norm": 0.1713397204875946, | |
| "learning_rate": 6.088947368421053e-05, | |
| "loss": 0.0023, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.1857451403887689, | |
| "grad_norm": 0.11895774304866791, | |
| "learning_rate": 6.036e-05, | |
| "loss": 0.001, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.1879049676025918, | |
| "grad_norm": 0.011528117582201958, | |
| "learning_rate": 5.9830526315789475e-05, | |
| "loss": 0.0003, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.1900647948164147, | |
| "grad_norm": 0.0027349735610187054, | |
| "learning_rate": 5.9301052631578946e-05, | |
| "loss": 0.0002, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.19222462203023757, | |
| "grad_norm": 0.0025566022377461195, | |
| "learning_rate": 5.877157894736843e-05, | |
| "loss": 0.0001, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.19438444924406048, | |
| "grad_norm": 0.010126935318112373, | |
| "learning_rate": 5.82421052631579e-05, | |
| "loss": 0.0003, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.19654427645788336, | |
| "grad_norm": 0.0021927962079644203, | |
| "learning_rate": 5.771263157894737e-05, | |
| "loss": 0.0001, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.19870410367170627, | |
| "grad_norm": 0.007295831106603146, | |
| "learning_rate": 5.718315789473685e-05, | |
| "loss": 0.0002, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.20086393088552915, | |
| "grad_norm": 0.3323706388473511, | |
| "learning_rate": 5.665368421052632e-05, | |
| "loss": 0.0008, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.20302375809935205, | |
| "grad_norm": 0.007849380373954773, | |
| "learning_rate": 5.612421052631579e-05, | |
| "loss": 0.0001, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.20518358531317496, | |
| "grad_norm": 0.5382089614868164, | |
| "learning_rate": 5.559473684210527e-05, | |
| "loss": 0.0054, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.20734341252699784, | |
| "grad_norm": 0.004848291631788015, | |
| "learning_rate": 5.506526315789474e-05, | |
| "loss": 0.0003, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.20950323974082075, | |
| "grad_norm": 0.009182745590806007, | |
| "learning_rate": 5.453578947368421e-05, | |
| "loss": 0.0003, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.21166306695464362, | |
| "grad_norm": 0.02678913250565529, | |
| "learning_rate": 5.400631578947369e-05, | |
| "loss": 0.0003, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.21382289416846653, | |
| "grad_norm": 0.10986873507499695, | |
| "learning_rate": 5.347684210526316e-05, | |
| "loss": 0.0009, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.2159827213822894, | |
| "grad_norm": 0.6294146180152893, | |
| "learning_rate": 5.294736842105263e-05, | |
| "loss": 0.0097, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.2159827213822894, | |
| "eval_loss": 0.004535946063697338, | |
| "eval_runtime": 38.2221, | |
| "eval_samples_per_second": 5.102, | |
| "eval_steps_per_second": 1.282, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.21814254859611232, | |
| "grad_norm": 0.23118843138217926, | |
| "learning_rate": 5.24178947368421e-05, | |
| "loss": 0.0068, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.2203023758099352, | |
| "grad_norm": 0.07103104144334793, | |
| "learning_rate": 5.1888421052631585e-05, | |
| "loss": 0.0039, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.2224622030237581, | |
| "grad_norm": 0.03884083777666092, | |
| "learning_rate": 5.135894736842106e-05, | |
| "loss": 0.0021, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.22462203023758098, | |
| "grad_norm": 0.01880479045212269, | |
| "learning_rate": 5.082947368421053e-05, | |
| "loss": 0.0009, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.2267818574514039, | |
| "grad_norm": 0.0531153567135334, | |
| "learning_rate": 5.03e-05, | |
| "loss": 0.001, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.22894168466522677, | |
| "grad_norm": 0.053985074162483215, | |
| "learning_rate": 4.977052631578947e-05, | |
| "loss": 0.0013, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.23110151187904968, | |
| "grad_norm": 0.037867240607738495, | |
| "learning_rate": 4.924105263157895e-05, | |
| "loss": 0.0023, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.23326133909287258, | |
| "grad_norm": 0.11338239908218384, | |
| "learning_rate": 4.871157894736843e-05, | |
| "loss": 0.0024, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.23542116630669546, | |
| "grad_norm": 0.014712098054587841, | |
| "learning_rate": 4.81821052631579e-05, | |
| "loss": 0.0009, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.23758099352051837, | |
| "grad_norm": 0.03086118772625923, | |
| "learning_rate": 4.765263157894737e-05, | |
| "loss": 0.001, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.23974082073434125, | |
| "grad_norm": 0.013892450369894505, | |
| "learning_rate": 4.7123157894736845e-05, | |
| "loss": 0.0005, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.24190064794816415, | |
| "grad_norm": 0.2487848550081253, | |
| "learning_rate": 4.6593684210526316e-05, | |
| "loss": 0.0082, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.24406047516198703, | |
| "grad_norm": 0.11459369212388992, | |
| "learning_rate": 4.606421052631579e-05, | |
| "loss": 0.0037, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.24622030237580994, | |
| "grad_norm": 0.07744896411895752, | |
| "learning_rate": 4.553473684210527e-05, | |
| "loss": 0.0023, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.24838012958963282, | |
| "grad_norm": 0.10101820528507233, | |
| "learning_rate": 4.500526315789474e-05, | |
| "loss": 0.0034, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.2505399568034557, | |
| "grad_norm": 0.007767634466290474, | |
| "learning_rate": 4.447578947368421e-05, | |
| "loss": 0.0003, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.2526997840172786, | |
| "grad_norm": 0.10766242444515228, | |
| "learning_rate": 4.394631578947369e-05, | |
| "loss": 0.0012, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.2548596112311015, | |
| "grad_norm": 0.008573450148105621, | |
| "learning_rate": 4.341684210526316e-05, | |
| "loss": 0.0003, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.2570194384449244, | |
| "grad_norm": 0.0011212700046598911, | |
| "learning_rate": 4.2887368421052636e-05, | |
| "loss": 0.0001, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.2591792656587473, | |
| "grad_norm": 0.005315995309501886, | |
| "learning_rate": 4.2357894736842106e-05, | |
| "loss": 0.0002, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.2613390928725702, | |
| "grad_norm": 0.0370321199297905, | |
| "learning_rate": 4.182842105263158e-05, | |
| "loss": 0.0002, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.2634989200863931, | |
| "grad_norm": 0.0062796249985694885, | |
| "learning_rate": 4.1298947368421053e-05, | |
| "loss": 0.0002, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.265658747300216, | |
| "grad_norm": 0.06505784392356873, | |
| "learning_rate": 4.0769473684210524e-05, | |
| "loss": 0.0004, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.2678185745140389, | |
| "grad_norm": 0.11662036925554276, | |
| "learning_rate": 4.024000000000001e-05, | |
| "loss": 0.0003, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.26997840172786175, | |
| "grad_norm": 0.01756948232650757, | |
| "learning_rate": 3.971052631578948e-05, | |
| "loss": 0.0003, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.27213822894168466, | |
| "grad_norm": 0.01176679227501154, | |
| "learning_rate": 3.918105263157895e-05, | |
| "loss": 0.0002, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.27429805615550756, | |
| "grad_norm": 0.16879205405712128, | |
| "learning_rate": 3.865157894736842e-05, | |
| "loss": 0.0051, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.27645788336933047, | |
| "grad_norm": 0.015613168478012085, | |
| "learning_rate": 3.8122105263157896e-05, | |
| "loss": 0.0003, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.2786177105831533, | |
| "grad_norm": 0.001986750401556492, | |
| "learning_rate": 3.759263157894737e-05, | |
| "loss": 0.0001, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.28077753779697623, | |
| "grad_norm": 0.008711952716112137, | |
| "learning_rate": 3.7063157894736844e-05, | |
| "loss": 0.0003, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.28293736501079914, | |
| "grad_norm": 0.0025189965963363647, | |
| "learning_rate": 3.653368421052632e-05, | |
| "loss": 0.0001, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.28509719222462204, | |
| "grad_norm": 0.0006644928944297135, | |
| "learning_rate": 3.600421052631579e-05, | |
| "loss": 0.0001, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.28725701943844495, | |
| "grad_norm": 0.0026721965987235308, | |
| "learning_rate": 3.547473684210526e-05, | |
| "loss": 0.0002, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.2894168466522678, | |
| "grad_norm": 0.015398127026855946, | |
| "learning_rate": 3.494526315789474e-05, | |
| "loss": 0.0001, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.2915766738660907, | |
| "grad_norm": 0.014543671160936356, | |
| "learning_rate": 3.4415789473684216e-05, | |
| "loss": 0.0003, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.2937365010799136, | |
| "grad_norm": 0.006670842412859201, | |
| "learning_rate": 3.3886315789473686e-05, | |
| "loss": 0.0001, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.2958963282937365, | |
| "grad_norm": 0.005377969238907099, | |
| "learning_rate": 3.3356842105263156e-05, | |
| "loss": 0.0002, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.2980561555075594, | |
| "grad_norm": 0.0062574828043580055, | |
| "learning_rate": 3.2827368421052634e-05, | |
| "loss": 0.0002, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.3002159827213823, | |
| "grad_norm": 0.14523354172706604, | |
| "learning_rate": 3.2297894736842104e-05, | |
| "loss": 0.0145, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.3023758099352052, | |
| "grad_norm": 0.010900533758103848, | |
| "learning_rate": 3.176842105263158e-05, | |
| "loss": 0.0002, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.3045356371490281, | |
| "grad_norm": 0.010036949999630451, | |
| "learning_rate": 3.123894736842106e-05, | |
| "loss": 0.0001, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.30669546436285094, | |
| "grad_norm": 0.598667562007904, | |
| "learning_rate": 3.070947368421053e-05, | |
| "loss": 0.0035, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.30885529157667385, | |
| "grad_norm": 0.002981688594445586, | |
| "learning_rate": 3.018e-05, | |
| "loss": 0.0001, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.31101511879049676, | |
| "grad_norm": 0.0013599407393485308, | |
| "learning_rate": 2.9650526315789473e-05, | |
| "loss": 0.0001, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.31317494600431967, | |
| "grad_norm": 0.5247439742088318, | |
| "learning_rate": 2.912105263157895e-05, | |
| "loss": 0.0167, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.31533477321814257, | |
| "grad_norm": 0.004649030510336161, | |
| "learning_rate": 2.8591578947368424e-05, | |
| "loss": 0.0002, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.3174946004319654, | |
| "grad_norm": 0.0023146297316998243, | |
| "learning_rate": 2.8062105263157894e-05, | |
| "loss": 0.0001, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.31965442764578833, | |
| "grad_norm": 0.4256044328212738, | |
| "learning_rate": 2.753263157894737e-05, | |
| "loss": 0.0243, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.32181425485961124, | |
| "grad_norm": 0.22870223224163055, | |
| "learning_rate": 2.7003157894736845e-05, | |
| "loss": 0.0005, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.32397408207343414, | |
| "grad_norm": 5.829990386962891, | |
| "learning_rate": 2.6473684210526315e-05, | |
| "loss": 0.0135, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.32397408207343414, | |
| "eval_loss": 0.0008732560090720654, | |
| "eval_runtime": 38.606, | |
| "eval_samples_per_second": 5.051, | |
| "eval_steps_per_second": 1.269, | |
| "step": 150 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 200, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 50, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 5, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 0 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 7.467142214791987e+16, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |