| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 230, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.043478260869565216, | |
| "grad_norm": 175.07723361554483, | |
| "learning_rate": 8.695652173913044e-07, | |
| "loss": 1.9054, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.21739130434782608, | |
| "grad_norm": 62.962929820729734, | |
| "learning_rate": 4.347826086956522e-06, | |
| "loss": 1.4576, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.43478260869565216, | |
| "grad_norm": 8.648281883873826, | |
| "learning_rate": 8.695652173913044e-06, | |
| "loss": 0.6132, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.6521739130434783, | |
| "grad_norm": 16.239431113601718, | |
| "learning_rate": 1.3043478260869566e-05, | |
| "loss": 0.3115, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.8695652173913043, | |
| "grad_norm": 3.450808311744288, | |
| "learning_rate": 1.739130434782609e-05, | |
| "loss": 0.203, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 0.10863249748945236, | |
| "eval_runtime": 67.8349, | |
| "eval_samples_per_second": 21.169, | |
| "eval_steps_per_second": 0.663, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 1.0869565217391304, | |
| "grad_norm": 34.3749757818615, | |
| "learning_rate": 1.9995393663024054e-05, | |
| "loss": 0.1339, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 1.3043478260869565, | |
| "grad_norm": 2.3841429135265555, | |
| "learning_rate": 1.9943621095573588e-05, | |
| "loss": 0.1249, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 1.5217391304347827, | |
| "grad_norm": 1.2759644316371075, | |
| "learning_rate": 1.9834617016337424e-05, | |
| "loss": 0.088, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 1.7391304347826086, | |
| "grad_norm": 34.92728113964099, | |
| "learning_rate": 1.9669008809262064e-05, | |
| "loss": 0.0829, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 1.9565217391304348, | |
| "grad_norm": 2.0028289380489945, | |
| "learning_rate": 1.944774964904754e-05, | |
| "loss": 0.0942, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 0.07020141929388046, | |
| "eval_runtime": 67.1251, | |
| "eval_samples_per_second": 21.393, | |
| "eval_steps_per_second": 0.67, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 2.1739130434782608, | |
| "grad_norm": 1.3438276529586066, | |
| "learning_rate": 1.917211301505453e-05, | |
| "loss": 0.0674, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 2.391304347826087, | |
| "grad_norm": 1142.7739612838905, | |
| "learning_rate": 1.8843685361665724e-05, | |
| "loss": 0.0668, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 2.608695652173913, | |
| "grad_norm": 2.660687233406025, | |
| "learning_rate": 1.8464356987288012e-05, | |
| "loss": 0.065, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 2.8260869565217392, | |
| "grad_norm": 6.723586538182285, | |
| "learning_rate": 1.8036311154549783e-05, | |
| "loss": 0.064, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 0.057872917503118515, | |
| "eval_runtime": 67.6569, | |
| "eval_samples_per_second": 21.225, | |
| "eval_steps_per_second": 0.665, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 3.0434782608695654, | |
| "grad_norm": 0.7827050845032839, | |
| "learning_rate": 1.7562011524313187e-05, | |
| "loss": 0.0604, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 3.260869565217391, | |
| "grad_norm": 1.2639368123822572, | |
| "learning_rate": 1.7044187975826126e-05, | |
| "loss": 0.0755, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 3.4782608695652173, | |
| "grad_norm": 0.9047997047917037, | |
| "learning_rate": 1.648582089462756e-05, | |
| "loss": 0.0569, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 3.6956521739130435, | |
| "grad_norm": 0.5428178712369748, | |
| "learning_rate": 1.589012401863864e-05, | |
| "loss": 0.0577, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 3.9130434782608696, | |
| "grad_norm": 0.6450527453899985, | |
| "learning_rate": 1.526052594117071e-05, | |
| "loss": 0.055, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 0.05387277901172638, | |
| "eval_runtime": 67.296, | |
| "eval_samples_per_second": 21.339, | |
| "eval_steps_per_second": 0.669, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 4.130434782608695, | |
| "grad_norm": 0.6167289429330038, | |
| "learning_rate": 1.4600650377311523e-05, | |
| "loss": 0.0549, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 4.3478260869565215, | |
| "grad_norm": 0.5274268025844417, | |
| "learning_rate": 1.3914295307268396e-05, | |
| "loss": 0.0535, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 4.565217391304348, | |
| "grad_norm": 4.645394192055986, | |
| "learning_rate": 1.3205411116710973e-05, | |
| "loss": 0.0558, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 4.782608695652174, | |
| "grad_norm": 0.44016634132559596, | |
| "learning_rate": 1.2478077859929e-05, | |
| "loss": 0.0557, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.37871781752696215, | |
| "learning_rate": 1.1736481776669307e-05, | |
| "loss": 0.0533, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_loss": 0.05291759595274925, | |
| "eval_runtime": 67.622, | |
| "eval_samples_per_second": 21.236, | |
| "eval_steps_per_second": 0.665, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 5.217391304347826, | |
| "grad_norm": 0.4605215085179153, | |
| "learning_rate": 1.0984891197811686e-05, | |
| "loss": 0.0526, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 5.434782608695652, | |
| "grad_norm": 0.5874492251340868, | |
| "learning_rate": 1.0227631978561057e-05, | |
| "loss": 0.0531, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 5.6521739130434785, | |
| "grad_norm": 0.29510087532653284, | |
| "learning_rate": 9.469062600552509e-06, | |
| "loss": 0.0525, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 5.869565217391305, | |
| "grad_norm": 0.344906161076209, | |
| "learning_rate": 8.71354908617169e-06, | |
| "loss": 0.0525, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 0.0515441857278347, | |
| "eval_runtime": 67.5553, | |
| "eval_samples_per_second": 21.257, | |
| "eval_steps_per_second": 0.666, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 6.086956521739131, | |
| "grad_norm": 0.36354029610859134, | |
| "learning_rate": 7.965439869473664e-06, | |
| "loss": 0.0514, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 6.304347826086957, | |
| "grad_norm": 0.26428382375759113, | |
| "learning_rate": 7.2290407683331154e-06, | |
| "loss": 0.0515, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 6.521739130434782, | |
| "grad_norm": 0.22822648732081055, | |
| "learning_rate": 6.508590201876317e-06, | |
| "loss": 0.0512, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 6.739130434782608, | |
| "grad_norm": 0.25828737593751994, | |
| "learning_rate": 5.8082347958333625e-06, | |
| "loss": 0.0513, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 6.956521739130435, | |
| "grad_norm": 0.2285799063924952, | |
| "learning_rate": 5.132005516216512e-06, | |
| "loss": 0.0519, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_loss": 0.05051277205348015, | |
| "eval_runtime": 67.5123, | |
| "eval_samples_per_second": 21.27, | |
| "eval_steps_per_second": 0.667, | |
| "step": 161 | |
| }, | |
| { | |
| "epoch": 7.173913043478261, | |
| "grad_norm": 0.24475626828923558, | |
| "learning_rate": 4.483794468689728e-06, | |
| "loss": 0.0508, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 7.391304347826087, | |
| "grad_norm": 0.2520072015362384, | |
| "learning_rate": 3.867332497162836e-06, | |
| "loss": 0.0504, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 7.608695652173913, | |
| "grad_norm": 0.27382520169651525, | |
| "learning_rate": 3.2861677105440335e-06, | |
| "loss": 0.0504, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 7.826086956521739, | |
| "grad_norm": 0.2716287514584957, | |
| "learning_rate": 2.7436450612420098e-06, | |
| "loss": 0.0505, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 0.04957514628767967, | |
| "eval_runtime": 67.276, | |
| "eval_samples_per_second": 21.345, | |
| "eval_steps_per_second": 0.669, | |
| "step": 184 | |
| }, | |
| { | |
| "epoch": 8.043478260869565, | |
| "grad_norm": 0.21787711573464694, | |
| "learning_rate": 2.2428870929558012e-06, | |
| "loss": 0.0496, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 8.26086956521739, | |
| "grad_norm": 0.21445383159071862, | |
| "learning_rate": 1.7867759685603115e-06, | |
| "loss": 0.0494, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 8.478260869565217, | |
| "grad_norm": 0.24405861737397047, | |
| "learning_rate": 1.3779368815278648e-06, | |
| "loss": 0.0491, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 8.695652173913043, | |
| "grad_norm": 0.2697463930100617, | |
| "learning_rate": 1.01872294636304e-06, | |
| "loss": 0.0495, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 8.91304347826087, | |
| "grad_norm": 0.21112878071345867, | |
| "learning_rate": 7.1120165501533e-07, | |
| "loss": 0.0494, | |
| "step": 205 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_loss": 0.048730239272117615, | |
| "eval_runtime": 66.913, | |
| "eval_samples_per_second": 21.461, | |
| "eval_steps_per_second": 0.673, | |
| "step": 207 | |
| }, | |
| { | |
| "epoch": 9.130434782608695, | |
| "grad_norm": 0.19010353057716406, | |
| "learning_rate": 4.5714297722121105e-07, | |
| "loss": 0.049, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 9.347826086956522, | |
| "grad_norm": 0.17948234694281104, | |
| "learning_rate": 2.5800917326521013e-07, | |
| "loss": 0.0491, | |
| "step": 215 | |
| }, | |
| { | |
| "epoch": 9.565217391304348, | |
| "grad_norm": 0.19343675188279383, | |
| "learning_rate": 1.1494637779369766e-07, | |
| "loss": 0.0489, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 9.782608695652174, | |
| "grad_norm": 0.21072621537133224, | |
| "learning_rate": 2.8778003121607834e-08, | |
| "loss": 0.0483, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 0.18309652115048786, | |
| "learning_rate": 0.0, | |
| "loss": 0.0484, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 0.0484623983502388, | |
| "eval_runtime": 67.5159, | |
| "eval_samples_per_second": 21.269, | |
| "eval_steps_per_second": 0.667, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 230, | |
| "total_flos": 48157320806400.0, | |
| "train_loss": 0.1130801611620447, | |
| "train_runtime": 3471.4646, | |
| "train_samples_per_second": 4.137, | |
| "train_steps_per_second": 0.066 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 230, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 100, | |
| "total_flos": 48157320806400.0, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |