| { | |
| "best_metric": 0.4497845768928528, | |
| "best_model_checkpoint": "/home1/datahome/villien/project_hub/DinoVdeau/models/DinoVdrone-large-2025_02_03_31850-bs32_freeze_probs/checkpoint-2052", | |
| "epoch": 37.0, | |
| "eval_steps": 500, | |
| "global_step": 2812, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "eval_explained_variance": 0.0698663592338562, | |
| "eval_kl_divergence": 1.791170597076416, | |
| "eval_loss": 0.5543646216392517, | |
| "eval_mae": 0.22781437635421753, | |
| "eval_rmse": 0.26052138209342957, | |
| "eval_runtime": 36.0517, | |
| "eval_samples_per_second": 22.218, | |
| "eval_steps_per_second": 0.721, | |
| "learning_rate": 0.001, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_explained_variance": 0.07257537543773651, | |
| "eval_kl_divergence": 1.0344740152359009, | |
| "eval_loss": 0.48171326518058777, | |
| "eval_mae": 0.16019925475120544, | |
| "eval_rmse": 0.20066045224666595, | |
| "eval_runtime": 32.1858, | |
| "eval_samples_per_second": 24.887, | |
| "eval_steps_per_second": 0.808, | |
| "learning_rate": 0.001, | |
| "step": 152 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_explained_variance": 0.27069091796875, | |
| "eval_kl_divergence": 0.6063743233680725, | |
| "eval_loss": 0.4615386724472046, | |
| "eval_mae": 0.1370290368795395, | |
| "eval_rmse": 0.18013149499893188, | |
| "eval_runtime": 22.3828, | |
| "eval_samples_per_second": 35.786, | |
| "eval_steps_per_second": 1.162, | |
| "learning_rate": 0.001, | |
| "step": 228 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_explained_variance": 0.08923790603876114, | |
| "eval_kl_divergence": 0.6576936841011047, | |
| "eval_loss": 0.4632064700126648, | |
| "eval_mae": 0.1390942633152008, | |
| "eval_rmse": 0.18370357155799866, | |
| "eval_runtime": 22.4997, | |
| "eval_samples_per_second": 35.601, | |
| "eval_steps_per_second": 1.156, | |
| "learning_rate": 0.001, | |
| "step": 304 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_explained_variance": 0.2999000549316406, | |
| "eval_kl_divergence": 0.6677561402320862, | |
| "eval_loss": 0.4579373300075531, | |
| "eval_mae": 0.13632091879844666, | |
| "eval_rmse": 0.17685972154140472, | |
| "eval_runtime": 22.4993, | |
| "eval_samples_per_second": 35.601, | |
| "eval_steps_per_second": 1.156, | |
| "learning_rate": 0.001, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_explained_variance": 0.30495360493659973, | |
| "eval_kl_divergence": 0.7895806431770325, | |
| "eval_loss": 0.4571229815483093, | |
| "eval_mae": 0.13304883241653442, | |
| "eval_rmse": 0.176561176776886, | |
| "eval_runtime": 22.7, | |
| "eval_samples_per_second": 35.286, | |
| "eval_steps_per_second": 1.145, | |
| "learning_rate": 0.001, | |
| "step": 456 | |
| }, | |
| { | |
| "epoch": 6.578947368421053, | |
| "grad_norm": 0.27282145619392395, | |
| "learning_rate": 0.001, | |
| "loss": 0.4966, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_explained_variance": 0.3103345036506653, | |
| "eval_kl_divergence": 0.6492787599563599, | |
| "eval_loss": 0.4586440622806549, | |
| "eval_mae": 0.13074660301208496, | |
| "eval_rmse": 0.17730861902236938, | |
| "eval_runtime": 22.4946, | |
| "eval_samples_per_second": 35.608, | |
| "eval_steps_per_second": 1.156, | |
| "learning_rate": 0.001, | |
| "step": 532 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_explained_variance": 0.30848240852355957, | |
| "eval_kl_divergence": 0.9475247263908386, | |
| "eval_loss": 0.45786425471305847, | |
| "eval_mae": 0.13187319040298462, | |
| "eval_rmse": 0.17719128727912903, | |
| "eval_runtime": 22.513, | |
| "eval_samples_per_second": 35.579, | |
| "eval_steps_per_second": 1.155, | |
| "learning_rate": 0.001, | |
| "step": 608 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_explained_variance": 0.3135569095611572, | |
| "eval_kl_divergence": 0.7270792126655579, | |
| "eval_loss": 0.4551210105419159, | |
| "eval_mae": 0.13063107430934906, | |
| "eval_rmse": 0.17461702227592468, | |
| "eval_runtime": 22.4949, | |
| "eval_samples_per_second": 35.608, | |
| "eval_steps_per_second": 1.156, | |
| "learning_rate": 0.001, | |
| "step": 684 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_explained_variance": 0.29183268547058105, | |
| "eval_kl_divergence": 0.6882225275039673, | |
| "eval_loss": 0.45817646384239197, | |
| "eval_mae": 0.13160471618175507, | |
| "eval_rmse": 0.17741131782531738, | |
| "eval_runtime": 22.508, | |
| "eval_samples_per_second": 35.587, | |
| "eval_steps_per_second": 1.155, | |
| "learning_rate": 0.001, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_explained_variance": 0.25859755277633667, | |
| "eval_kl_divergence": 0.371459424495697, | |
| "eval_loss": 0.46834859251976013, | |
| "eval_mae": 0.13717466592788696, | |
| "eval_rmse": 0.18415462970733643, | |
| "eval_runtime": 22.4535, | |
| "eval_samples_per_second": 35.674, | |
| "eval_steps_per_second": 1.158, | |
| "learning_rate": 0.001, | |
| "step": 836 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_explained_variance": 0.308597594499588, | |
| "eval_kl_divergence": 0.5270651578903198, | |
| "eval_loss": 0.4578668475151062, | |
| "eval_mae": 0.13155478239059448, | |
| "eval_rmse": 0.17636829614639282, | |
| "eval_runtime": 22.5868, | |
| "eval_samples_per_second": 35.463, | |
| "eval_steps_per_second": 1.151, | |
| "learning_rate": 0.001, | |
| "step": 912 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_explained_variance": 0.3100161552429199, | |
| "eval_kl_divergence": 0.9167731404304504, | |
| "eval_loss": 0.4558842182159424, | |
| "eval_mae": 0.1300584226846695, | |
| "eval_rmse": 0.1756095588207245, | |
| "eval_runtime": 22.7991, | |
| "eval_samples_per_second": 35.133, | |
| "eval_steps_per_second": 1.14, | |
| "learning_rate": 0.001, | |
| "step": 988 | |
| }, | |
| { | |
| "epoch": 13.157894736842104, | |
| "grad_norm": 0.26468560099601746, | |
| "learning_rate": 0.001, | |
| "loss": 0.4448, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_explained_variance": 0.3228832185268402, | |
| "eval_kl_divergence": 0.8826888799667358, | |
| "eval_loss": 0.4555540680885315, | |
| "eval_mae": 0.12918463349342346, | |
| "eval_rmse": 0.17491525411605835, | |
| "eval_runtime": 22.7093, | |
| "eval_samples_per_second": 35.272, | |
| "eval_steps_per_second": 1.145, | |
| "learning_rate": 0.001, | |
| "step": 1064 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_explained_variance": 0.34160685539245605, | |
| "eval_kl_divergence": 0.7008672952651978, | |
| "eval_loss": 0.45217740535736084, | |
| "eval_mae": 0.1262015700340271, | |
| "eval_rmse": 0.17165420949459076, | |
| "eval_runtime": 22.6104, | |
| "eval_samples_per_second": 35.426, | |
| "eval_steps_per_second": 1.15, | |
| "learning_rate": 0.001, | |
| "step": 1140 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_explained_variance": 0.31633102893829346, | |
| "eval_kl_divergence": 1.0038130283355713, | |
| "eval_loss": 0.45556434988975525, | |
| "eval_mae": 0.12863175570964813, | |
| "eval_rmse": 0.1752910166978836, | |
| "eval_runtime": 22.5805, | |
| "eval_samples_per_second": 35.473, | |
| "eval_steps_per_second": 1.151, | |
| "learning_rate": 0.001, | |
| "step": 1216 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "eval_explained_variance": 0.3205307126045227, | |
| "eval_kl_divergence": 0.2600082457065582, | |
| "eval_loss": 0.458648681640625, | |
| "eval_mae": 0.13426683843135834, | |
| "eval_rmse": 0.17750133574008942, | |
| "eval_runtime": 22.3997, | |
| "eval_samples_per_second": 35.759, | |
| "eval_steps_per_second": 1.161, | |
| "learning_rate": 0.001, | |
| "step": 1292 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_explained_variance": -4.778772354125977, | |
| "eval_kl_divergence": 2.054769277572632, | |
| "eval_loss": 0.567169725894928, | |
| "eval_mae": 0.16376179456710815, | |
| "eval_rmse": 0.23688165843486786, | |
| "eval_runtime": 22.2481, | |
| "eval_samples_per_second": 36.003, | |
| "eval_steps_per_second": 1.169, | |
| "learning_rate": 0.001, | |
| "step": 1368 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "eval_explained_variance": 0.32792216539382935, | |
| "eval_kl_divergence": 0.7114961743354797, | |
| "eval_loss": 0.45287612080574036, | |
| "eval_mae": 0.12865176796913147, | |
| "eval_rmse": 0.17274516820907593, | |
| "eval_runtime": 22.3304, | |
| "eval_samples_per_second": 35.87, | |
| "eval_steps_per_second": 1.164, | |
| "learning_rate": 0.001, | |
| "step": 1444 | |
| }, | |
| { | |
| "epoch": 19.736842105263158, | |
| "grad_norm": 0.1475011706352234, | |
| "learning_rate": 0.001, | |
| "loss": 0.4406, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_explained_variance": 0.3204135596752167, | |
| "eval_kl_divergence": 0.9694227576255798, | |
| "eval_loss": 0.45518893003463745, | |
| "eval_mae": 0.12852200865745544, | |
| "eval_rmse": 0.17462262511253357, | |
| "eval_runtime": 22.5552, | |
| "eval_samples_per_second": 35.513, | |
| "eval_steps_per_second": 1.153, | |
| "learning_rate": 0.001, | |
| "step": 1520 | |
| }, | |
| { | |
| "epoch": 21.0, | |
| "eval_explained_variance": 0.32996666431427, | |
| "eval_kl_divergence": 0.778915524482727, | |
| "eval_loss": 0.45299893617630005, | |
| "eval_mae": 0.12820784747600555, | |
| "eval_rmse": 0.17243456840515137, | |
| "eval_runtime": 22.5861, | |
| "eval_samples_per_second": 35.464, | |
| "eval_steps_per_second": 1.151, | |
| "learning_rate": 0.001, | |
| "step": 1596 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "eval_explained_variance": 0.34726351499557495, | |
| "eval_kl_divergence": 0.7368760704994202, | |
| "eval_loss": 0.4502638280391693, | |
| "eval_mae": 0.12613575160503387, | |
| "eval_rmse": 0.17001575231552124, | |
| "eval_runtime": 22.7271, | |
| "eval_samples_per_second": 35.244, | |
| "eval_steps_per_second": 1.144, | |
| "learning_rate": 0.0001, | |
| "step": 1672 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "eval_explained_variance": 0.34029728174209595, | |
| "eval_kl_divergence": 0.5027008056640625, | |
| "eval_loss": 0.453466534614563, | |
| "eval_mae": 0.12802833318710327, | |
| "eval_rmse": 0.17160943150520325, | |
| "eval_runtime": 22.4563, | |
| "eval_samples_per_second": 35.669, | |
| "eval_steps_per_second": 1.158, | |
| "learning_rate": 0.0001, | |
| "step": 1748 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "eval_explained_variance": 0.3511368930339813, | |
| "eval_kl_divergence": 0.5968054533004761, | |
| "eval_loss": 0.4502425491809845, | |
| "eval_mae": 0.12641073763370514, | |
| "eval_rmse": 0.16971111297607422, | |
| "eval_runtime": 22.592, | |
| "eval_samples_per_second": 35.455, | |
| "eval_steps_per_second": 1.151, | |
| "learning_rate": 0.0001, | |
| "step": 1824 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "eval_explained_variance": 0.3504308760166168, | |
| "eval_kl_divergence": 0.621475100517273, | |
| "eval_loss": 0.45040303468704224, | |
| "eval_mae": 0.12673497200012207, | |
| "eval_rmse": 0.1699284017086029, | |
| "eval_runtime": 22.3202, | |
| "eval_samples_per_second": 35.887, | |
| "eval_steps_per_second": 1.165, | |
| "learning_rate": 0.0001, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 26.0, | |
| "eval_explained_variance": 0.34598857164382935, | |
| "eval_kl_divergence": 0.6567814350128174, | |
| "eval_loss": 0.4509589374065399, | |
| "eval_mae": 0.12596669793128967, | |
| "eval_rmse": 0.17043226957321167, | |
| "eval_runtime": 22.4313, | |
| "eval_samples_per_second": 35.709, | |
| "eval_steps_per_second": 1.159, | |
| "learning_rate": 0.0001, | |
| "step": 1976 | |
| }, | |
| { | |
| "epoch": 26.31578947368421, | |
| "grad_norm": 0.16591614484786987, | |
| "learning_rate": 0.0001, | |
| "loss": 0.4334, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 27.0, | |
| "eval_explained_variance": 0.35463404655456543, | |
| "eval_kl_divergence": 0.5748001337051392, | |
| "eval_loss": 0.4497845768928528, | |
| "eval_mae": 0.1262420266866684, | |
| "eval_rmse": 0.1693224012851715, | |
| "eval_runtime": 22.3409, | |
| "eval_samples_per_second": 35.854, | |
| "eval_steps_per_second": 1.164, | |
| "learning_rate": 0.0001, | |
| "step": 2052 | |
| }, | |
| { | |
| "epoch": 28.0, | |
| "eval_explained_variance": 0.34665071964263916, | |
| "eval_kl_divergence": 0.7001035809516907, | |
| "eval_loss": 0.45060041546821594, | |
| "eval_mae": 0.12559720873832703, | |
| "eval_rmse": 0.17011338472366333, | |
| "eval_runtime": 22.4894, | |
| "eval_samples_per_second": 35.617, | |
| "eval_steps_per_second": 1.156, | |
| "learning_rate": 0.0001, | |
| "step": 2128 | |
| }, | |
| { | |
| "epoch": 29.0, | |
| "eval_explained_variance": 0.3531297743320465, | |
| "eval_kl_divergence": 0.5840001702308655, | |
| "eval_loss": 0.4504892826080322, | |
| "eval_mae": 0.12626922130584717, | |
| "eval_rmse": 0.16992022097110748, | |
| "eval_runtime": 22.4286, | |
| "eval_samples_per_second": 35.713, | |
| "eval_steps_per_second": 1.159, | |
| "learning_rate": 0.0001, | |
| "step": 2204 | |
| }, | |
| { | |
| "epoch": 30.0, | |
| "eval_explained_variance": 0.34863966703414917, | |
| "eval_kl_divergence": 0.8101097345352173, | |
| "eval_loss": 0.45060065388679504, | |
| "eval_mae": 0.12516793608665466, | |
| "eval_rmse": 0.1702672839164734, | |
| "eval_runtime": 22.3007, | |
| "eval_samples_per_second": 35.918, | |
| "eval_steps_per_second": 1.166, | |
| "learning_rate": 0.0001, | |
| "step": 2280 | |
| }, | |
| { | |
| "epoch": 31.0, | |
| "eval_explained_variance": 0.3488965630531311, | |
| "eval_kl_divergence": 0.7415657043457031, | |
| "eval_loss": 0.45080825686454773, | |
| "eval_mae": 0.12486829608678818, | |
| "eval_rmse": 0.1701475977897644, | |
| "eval_runtime": 22.3895, | |
| "eval_samples_per_second": 35.776, | |
| "eval_steps_per_second": 1.161, | |
| "learning_rate": 0.0001, | |
| "step": 2356 | |
| }, | |
| { | |
| "epoch": 32.0, | |
| "eval_explained_variance": 0.3523526191711426, | |
| "eval_kl_divergence": 0.6401851177215576, | |
| "eval_loss": 0.4501984417438507, | |
| "eval_mae": 0.12540514767169952, | |
| "eval_rmse": 0.16971096396446228, | |
| "eval_runtime": 22.3438, | |
| "eval_samples_per_second": 35.849, | |
| "eval_steps_per_second": 1.164, | |
| "learning_rate": 0.0001, | |
| "step": 2432 | |
| }, | |
| { | |
| "epoch": 32.89473684210526, | |
| "grad_norm": 0.1912919282913208, | |
| "learning_rate": 0.0001, | |
| "loss": 0.4289, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 33.0, | |
| "eval_explained_variance": 0.33917075395584106, | |
| "eval_kl_divergence": 0.8411455154418945, | |
| "eval_loss": 0.4510658085346222, | |
| "eval_mae": 0.12500226497650146, | |
| "eval_rmse": 0.1709534227848053, | |
| "eval_runtime": 22.56, | |
| "eval_samples_per_second": 35.505, | |
| "eval_steps_per_second": 1.152, | |
| "learning_rate": 0.0001, | |
| "step": 2508 | |
| }, | |
| { | |
| "epoch": 34.0, | |
| "eval_explained_variance": 0.33831754326820374, | |
| "eval_kl_divergence": 0.7203648686408997, | |
| "eval_loss": 0.45148056745529175, | |
| "eval_mae": 0.12593072652816772, | |
| "eval_rmse": 0.17108403146266937, | |
| "eval_runtime": 22.505, | |
| "eval_samples_per_second": 35.592, | |
| "eval_steps_per_second": 1.155, | |
| "learning_rate": 1e-05, | |
| "step": 2584 | |
| }, | |
| { | |
| "epoch": 35.0, | |
| "eval_explained_variance": 0.34977057576179504, | |
| "eval_kl_divergence": 0.7354820966720581, | |
| "eval_loss": 0.4502483904361725, | |
| "eval_mae": 0.12473371624946594, | |
| "eval_rmse": 0.16982755064964294, | |
| "eval_runtime": 22.3912, | |
| "eval_samples_per_second": 35.773, | |
| "eval_steps_per_second": 1.161, | |
| "learning_rate": 1e-05, | |
| "step": 2660 | |
| }, | |
| { | |
| "epoch": 36.0, | |
| "eval_explained_variance": 0.3486325144767761, | |
| "eval_kl_divergence": 0.49899470806121826, | |
| "eval_loss": 0.4508889615535736, | |
| "eval_mae": 0.1260843575000763, | |
| "eval_rmse": 0.1702878773212433, | |
| "eval_runtime": 22.8537, | |
| "eval_samples_per_second": 35.049, | |
| "eval_steps_per_second": 1.138, | |
| "learning_rate": 1e-05, | |
| "step": 2736 | |
| }, | |
| { | |
| "epoch": 37.0, | |
| "eval_explained_variance": 0.3535779118537903, | |
| "eval_kl_divergence": 0.5451197624206543, | |
| "eval_loss": 0.44998663663864136, | |
| "eval_mae": 0.12602195143699646, | |
| "eval_rmse": 0.16962522268295288, | |
| "eval_runtime": 22.2216, | |
| "eval_samples_per_second": 36.046, | |
| "eval_steps_per_second": 1.17, | |
| "learning_rate": 1e-05, | |
| "step": 2812 | |
| }, | |
| { | |
| "epoch": 37.0, | |
| "learning_rate": 1e-05, | |
| "step": 2812, | |
| "total_flos": 1.318896404308369e+20, | |
| "train_loss": 0.4464974437295797, | |
| "train_runtime": 4164.1679, | |
| "train_samples_per_second": 86.596, | |
| "train_steps_per_second": 2.738 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 11400, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 150, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 10, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 10 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.318896404308369e+20, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |