| { | |
| "best_metric": 1.5666028261184692, | |
| "best_model_checkpoint": "../../experiments_checkpoints/MAdAiLab/google/flan_t5_small_scotus/checkpoint-450", | |
| "epoch": 3.0, | |
| "eval_steps": 50, | |
| "global_step": 471, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.06, | |
| "grad_norm": 3.344521999359131, | |
| "learning_rate": 0.0004893842887473461, | |
| "loss": 2.3137, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 1.5987565517425537, | |
| "learning_rate": 0.00047876857749469217, | |
| "loss": 2.0989, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "grad_norm": 2.519667625427246, | |
| "learning_rate": 0.0004681528662420382, | |
| "loss": 2.1526, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 1.9053800106048584, | |
| "learning_rate": 0.00045753715498938433, | |
| "loss": 2.1042, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 2.938572406768799, | |
| "learning_rate": 0.0004469214437367304, | |
| "loss": 2.0567, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "eval_accuracy": 0.4042857142857143, | |
| "eval_f1_macro": 0.12210295202531032, | |
| "eval_f1_micro": 0.4042857142857143, | |
| "eval_loss": 1.961769700050354, | |
| "eval_runtime": 1.4405, | |
| "eval_samples_per_second": 971.865, | |
| "eval_steps_per_second": 30.544, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "grad_norm": 2.4352169036865234, | |
| "learning_rate": 0.00043630573248407644, | |
| "loss": 1.8701, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "grad_norm": 3.570573568344116, | |
| "learning_rate": 0.00042569002123142254, | |
| "loss": 1.8183, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 2.2673840522766113, | |
| "learning_rate": 0.0004150743099787686, | |
| "loss": 1.7879, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "grad_norm": 2.69811749458313, | |
| "learning_rate": 0.0004044585987261147, | |
| "loss": 1.73, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 2.467646837234497, | |
| "learning_rate": 0.00039384288747346076, | |
| "loss": 1.6729, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "eval_accuracy": 0.4542857142857143, | |
| "eval_f1_macro": 0.16563356681008953, | |
| "eval_f1_micro": 0.4542857142857143, | |
| "eval_loss": 1.7861312627792358, | |
| "eval_runtime": 1.441, | |
| "eval_samples_per_second": 971.567, | |
| "eval_steps_per_second": 30.535, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 3.9276022911071777, | |
| "learning_rate": 0.0003832271762208068, | |
| "loss": 1.691, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "grad_norm": 3.583529472351074, | |
| "learning_rate": 0.00037261146496815286, | |
| "loss": 1.7405, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "grad_norm": 3.401313304901123, | |
| "learning_rate": 0.0003619957537154989, | |
| "loss": 1.6642, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "grad_norm": 3.179844856262207, | |
| "learning_rate": 0.000351380042462845, | |
| "loss": 1.5905, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "grad_norm": 2.6475136280059814, | |
| "learning_rate": 0.0003407643312101911, | |
| "loss": 1.7754, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "eval_accuracy": 0.4664285714285714, | |
| "eval_f1_macro": 0.229693430245784, | |
| "eval_f1_micro": 0.4664285714285714, | |
| "eval_loss": 1.7046456336975098, | |
| "eval_runtime": 1.3928, | |
| "eval_samples_per_second": 1005.172, | |
| "eval_steps_per_second": 31.591, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.02, | |
| "grad_norm": 2.5741515159606934, | |
| "learning_rate": 0.00033014861995753713, | |
| "loss": 1.6254, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.08, | |
| "grad_norm": 3.005293369293213, | |
| "learning_rate": 0.00031953290870488323, | |
| "loss": 1.5145, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "grad_norm": 3.4725496768951416, | |
| "learning_rate": 0.0003089171974522293, | |
| "loss": 1.5766, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "grad_norm": 2.8154027462005615, | |
| "learning_rate": 0.00029830148619957534, | |
| "loss": 1.599, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 1.27, | |
| "grad_norm": 3.205636978149414, | |
| "learning_rate": 0.00028768577494692145, | |
| "loss": 1.5495, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.27, | |
| "eval_accuracy": 0.49357142857142855, | |
| "eval_f1_macro": 0.27123774899491926, | |
| "eval_f1_micro": 0.49357142857142855, | |
| "eval_loss": 1.6304272413253784, | |
| "eval_runtime": 1.4448, | |
| "eval_samples_per_second": 968.997, | |
| "eval_steps_per_second": 30.454, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.34, | |
| "grad_norm": 2.952043056488037, | |
| "learning_rate": 0.0002770700636942675, | |
| "loss": 1.54, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "grad_norm": 3.9415132999420166, | |
| "learning_rate": 0.0002664543524416136, | |
| "loss": 1.3851, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 1.46, | |
| "grad_norm": 3.267624855041504, | |
| "learning_rate": 0.00025583864118895966, | |
| "loss": 1.4578, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 1.53, | |
| "grad_norm": 2.674995183944702, | |
| "learning_rate": 0.0002452229299363057, | |
| "loss": 1.4493, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 1.59, | |
| "grad_norm": 3.1213557720184326, | |
| "learning_rate": 0.00023460721868365182, | |
| "loss": 1.4455, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 1.59, | |
| "eval_accuracy": 0.5114285714285715, | |
| "eval_f1_macro": 0.306348400600396, | |
| "eval_f1_micro": 0.5114285714285715, | |
| "eval_loss": 1.628359079360962, | |
| "eval_runtime": 1.4497, | |
| "eval_samples_per_second": 965.691, | |
| "eval_steps_per_second": 30.35, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 1.66, | |
| "grad_norm": 2.985950231552124, | |
| "learning_rate": 0.0002239915074309979, | |
| "loss": 1.5599, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 1.72, | |
| "grad_norm": 3.0701353549957275, | |
| "learning_rate": 0.00021337579617834395, | |
| "loss": 1.4849, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 1.78, | |
| "grad_norm": 2.7773752212524414, | |
| "learning_rate": 0.00020276008492569003, | |
| "loss": 1.3985, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "grad_norm": 3.100576877593994, | |
| "learning_rate": 0.0001921443736730361, | |
| "loss": 1.4363, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 1.91, | |
| "grad_norm": 2.965590476989746, | |
| "learning_rate": 0.00018152866242038217, | |
| "loss": 1.3035, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.91, | |
| "eval_accuracy": 0.5164285714285715, | |
| "eval_f1_macro": 0.31299962948284965, | |
| "eval_f1_micro": 0.5164285714285715, | |
| "eval_loss": 1.633267879486084, | |
| "eval_runtime": 1.4513, | |
| "eval_samples_per_second": 964.647, | |
| "eval_steps_per_second": 30.317, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "grad_norm": 3.2483999729156494, | |
| "learning_rate": 0.00017091295116772822, | |
| "loss": 1.3943, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 2.04, | |
| "grad_norm": 3.3734776973724365, | |
| "learning_rate": 0.0001602972399150743, | |
| "loss": 1.3575, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "grad_norm": 2.8578903675079346, | |
| "learning_rate": 0.00014968152866242038, | |
| "loss": 1.3466, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 2.17, | |
| "grad_norm": 3.1349565982818604, | |
| "learning_rate": 0.00013906581740976646, | |
| "loss": 1.393, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 2.23, | |
| "grad_norm": 3.431678533554077, | |
| "learning_rate": 0.00012845010615711254, | |
| "loss": 1.3057, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.23, | |
| "eval_accuracy": 0.5285714285714286, | |
| "eval_f1_macro": 0.33411386403759064, | |
| "eval_f1_micro": 0.5285714285714286, | |
| "eval_loss": 1.5789132118225098, | |
| "eval_runtime": 1.4525, | |
| "eval_samples_per_second": 963.834, | |
| "eval_steps_per_second": 30.292, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.29, | |
| "grad_norm": 2.3110225200653076, | |
| "learning_rate": 0.0001178343949044586, | |
| "loss": 1.2896, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 2.36, | |
| "grad_norm": 2.8537588119506836, | |
| "learning_rate": 0.00010721868365180467, | |
| "loss": 1.2275, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "grad_norm": 3.885326623916626, | |
| "learning_rate": 9.660297239915075e-05, | |
| "loss": 1.3663, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 2.48, | |
| "grad_norm": 3.13773250579834, | |
| "learning_rate": 8.598726114649682e-05, | |
| "loss": 1.3194, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "grad_norm": 3.769230604171753, | |
| "learning_rate": 7.537154989384288e-05, | |
| "loss": 1.2992, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "eval_accuracy": 0.5357142857142857, | |
| "eval_f1_macro": 0.3397937397852514, | |
| "eval_f1_micro": 0.5357142857142857, | |
| "eval_loss": 1.5673164129257202, | |
| "eval_runtime": 1.463, | |
| "eval_samples_per_second": 956.921, | |
| "eval_steps_per_second": 30.075, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.61, | |
| "grad_norm": 3.0843989849090576, | |
| "learning_rate": 6.475583864118896e-05, | |
| "loss": 1.3421, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 2.68, | |
| "grad_norm": 3.6903655529022217, | |
| "learning_rate": 5.414012738853504e-05, | |
| "loss": 1.304, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 2.74, | |
| "grad_norm": 3.312966823577881, | |
| "learning_rate": 4.35244161358811e-05, | |
| "loss": 1.2926, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "grad_norm": 3.0309150218963623, | |
| "learning_rate": 3.2908704883227177e-05, | |
| "loss": 1.1642, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 2.87, | |
| "grad_norm": 3.3688299655914307, | |
| "learning_rate": 2.2292993630573246e-05, | |
| "loss": 1.1637, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 2.87, | |
| "eval_accuracy": 0.5364285714285715, | |
| "eval_f1_macro": 0.34155184537672817, | |
| "eval_f1_micro": 0.5364285714285715, | |
| "eval_loss": 1.5666028261184692, | |
| "eval_runtime": 1.467, | |
| "eval_samples_per_second": 954.321, | |
| "eval_steps_per_second": 29.993, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 2.93, | |
| "grad_norm": 3.3567051887512207, | |
| "learning_rate": 1.167728237791932e-05, | |
| "loss": 1.2181, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 2.99, | |
| "grad_norm": 3.1731135845184326, | |
| "learning_rate": 1.0615711252653927e-06, | |
| "loss": 1.2584, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 471, | |
| "total_flos": 513144687755264.0, | |
| "train_loss": 1.5406995027181702, | |
| "train_runtime": 66.3357, | |
| "train_samples_per_second": 226.123, | |
| "train_steps_per_second": 7.1 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 471, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 50, | |
| "total_flos": 513144687755264.0, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |