| { |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 15.0, |
| "eval_steps": 500, |
| "global_step": 615, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.24390243902439024, |
| "grad_norm": 399.5959777832031, |
| "learning_rate": 1.2195121951219514e-06, |
| "loss": 20.1185, |
| "step": 10 |
| }, |
| { |
| "epoch": 0.4878048780487805, |
| "grad_norm": 488.2876892089844, |
| "learning_rate": 2.4390243902439027e-06, |
| "loss": 17.8526, |
| "step": 20 |
| }, |
| { |
| "epoch": 0.7317073170731707, |
| "grad_norm": 234.39358520507812, |
| "learning_rate": 3.6585365853658537e-06, |
| "loss": 14.6449, |
| "step": 30 |
| }, |
| { |
| "epoch": 0.975609756097561, |
| "grad_norm": 132.1689910888672, |
| "learning_rate": 4.8780487804878055e-06, |
| "loss": 12.2963, |
| "step": 40 |
| }, |
| { |
| "epoch": 1.2195121951219512, |
| "grad_norm": 108.00102996826172, |
| "learning_rate": 6.0975609756097564e-06, |
| "loss": 10.2374, |
| "step": 50 |
| }, |
| { |
| "epoch": 1.4634146341463414, |
| "grad_norm": 72.84204864501953, |
| "learning_rate": 7.317073170731707e-06, |
| "loss": 8.1356, |
| "step": 60 |
| }, |
| { |
| "epoch": 1.7073170731707317, |
| "grad_norm": 59.829524993896484, |
| "learning_rate": 8.536585365853658e-06, |
| "loss": 6.7279, |
| "step": 70 |
| }, |
| { |
| "epoch": 1.951219512195122, |
| "grad_norm": 42.48904800415039, |
| "learning_rate": 9.756097560975611e-06, |
| "loss": 6.0069, |
| "step": 80 |
| }, |
| { |
| "epoch": 2.1951219512195124, |
| "grad_norm": 37.831233978271484, |
| "learning_rate": 9.89159891598916e-06, |
| "loss": 5.2742, |
| "step": 90 |
| }, |
| { |
| "epoch": 2.4390243902439024, |
| "grad_norm": 27.763004302978516, |
| "learning_rate": 9.756097560975611e-06, |
| "loss": 4.8932, |
| "step": 100 |
| }, |
| { |
| "epoch": 2.682926829268293, |
| "grad_norm": 18.03118896484375, |
| "learning_rate": 9.620596205962061e-06, |
| "loss": 4.6851, |
| "step": 110 |
| }, |
| { |
| "epoch": 2.926829268292683, |
| "grad_norm": 18.090776443481445, |
| "learning_rate": 9.485094850948512e-06, |
| "loss": 4.5649, |
| "step": 120 |
| }, |
| { |
| "epoch": 3.1707317073170733, |
| "grad_norm": 15.51219654083252, |
| "learning_rate": 9.34959349593496e-06, |
| "loss": 4.492, |
| "step": 130 |
| }, |
| { |
| "epoch": 3.4146341463414633, |
| "grad_norm": 14.675006866455078, |
| "learning_rate": 9.21409214092141e-06, |
| "loss": 4.406, |
| "step": 140 |
| }, |
| { |
| "epoch": 3.658536585365854, |
| "grad_norm": 14.471592903137207, |
| "learning_rate": 9.07859078590786e-06, |
| "loss": 4.4155, |
| "step": 150 |
| }, |
| { |
| "epoch": 3.902439024390244, |
| "grad_norm": 13.054672241210938, |
| "learning_rate": 8.94308943089431e-06, |
| "loss": 4.3375, |
| "step": 160 |
| }, |
| { |
| "epoch": 4.146341463414634, |
| "grad_norm": 12.959569931030273, |
| "learning_rate": 8.80758807588076e-06, |
| "loss": 4.266, |
| "step": 170 |
| }, |
| { |
| "epoch": 4.390243902439025, |
| "grad_norm": 13.852252960205078, |
| "learning_rate": 8.67208672086721e-06, |
| "loss": 4.2173, |
| "step": 180 |
| }, |
| { |
| "epoch": 4.634146341463414, |
| "grad_norm": 13.482157707214355, |
| "learning_rate": 8.536585365853658e-06, |
| "loss": 4.2434, |
| "step": 190 |
| }, |
| { |
| "epoch": 4.878048780487805, |
| "grad_norm": 12.567044258117676, |
| "learning_rate": 8.401084010840109e-06, |
| "loss": 4.194, |
| "step": 200 |
| }, |
| { |
| "epoch": 5.121951219512195, |
| "grad_norm": 14.015665054321289, |
| "learning_rate": 8.265582655826559e-06, |
| "loss": 4.0342, |
| "step": 210 |
| }, |
| { |
| "epoch": 5.365853658536586, |
| "grad_norm": 16.776611328125, |
| "learning_rate": 8.130081300813009e-06, |
| "loss": 4.2782, |
| "step": 220 |
| }, |
| { |
| "epoch": 5.609756097560975, |
| "grad_norm": 12.224383354187012, |
| "learning_rate": 7.99457994579946e-06, |
| "loss": 4.1003, |
| "step": 230 |
| }, |
| { |
| "epoch": 5.853658536585366, |
| "grad_norm": 13.62357234954834, |
| "learning_rate": 7.859078590785908e-06, |
| "loss": 4.0503, |
| "step": 240 |
| }, |
| { |
| "epoch": 6.097560975609756, |
| "grad_norm": 13.351895332336426, |
| "learning_rate": 7.723577235772358e-06, |
| "loss": 3.9835, |
| "step": 250 |
| }, |
| { |
| "epoch": 6.341463414634147, |
| "grad_norm": 14.620662689208984, |
| "learning_rate": 7.5880758807588085e-06, |
| "loss": 4.0102, |
| "step": 260 |
| }, |
| { |
| "epoch": 6.585365853658536, |
| "grad_norm": 15.941452026367188, |
| "learning_rate": 7.452574525745257e-06, |
| "loss": 3.9992, |
| "step": 270 |
| }, |
| { |
| "epoch": 6.829268292682927, |
| "grad_norm": 11.67154598236084, |
| "learning_rate": 7.317073170731707e-06, |
| "loss": 4.015, |
| "step": 280 |
| }, |
| { |
| "epoch": 7.073170731707317, |
| "grad_norm": 12.082201957702637, |
| "learning_rate": 7.181571815718158e-06, |
| "loss": 3.9566, |
| "step": 290 |
| }, |
| { |
| "epoch": 7.317073170731708, |
| "grad_norm": 14.346577644348145, |
| "learning_rate": 7.046070460704607e-06, |
| "loss": 3.9484, |
| "step": 300 |
| }, |
| { |
| "epoch": 7.560975609756097, |
| "grad_norm": 13.629508018493652, |
| "learning_rate": 6.910569105691057e-06, |
| "loss": 3.8996, |
| "step": 310 |
| }, |
| { |
| "epoch": 7.804878048780488, |
| "grad_norm": 12.285475730895996, |
| "learning_rate": 6.775067750677508e-06, |
| "loss": 3.8117, |
| "step": 320 |
| }, |
| { |
| "epoch": 8.048780487804878, |
| "grad_norm": 13.470380783081055, |
| "learning_rate": 6.639566395663957e-06, |
| "loss": 3.7927, |
| "step": 330 |
| }, |
| { |
| "epoch": 8.292682926829269, |
| "grad_norm": 12.478575706481934, |
| "learning_rate": 6.504065040650407e-06, |
| "loss": 3.851, |
| "step": 340 |
| }, |
| { |
| "epoch": 8.536585365853659, |
| "grad_norm": 13.219705581665039, |
| "learning_rate": 6.368563685636857e-06, |
| "loss": 3.9406, |
| "step": 350 |
| }, |
| { |
| "epoch": 8.78048780487805, |
| "grad_norm": 12.26967716217041, |
| "learning_rate": 6.233062330623306e-06, |
| "loss": 3.8733, |
| "step": 360 |
| }, |
| { |
| "epoch": 9.024390243902438, |
| "grad_norm": 14.403388023376465, |
| "learning_rate": 6.0975609756097564e-06, |
| "loss": 3.6919, |
| "step": 370 |
| }, |
| { |
| "epoch": 9.268292682926829, |
| "grad_norm": 12.453432083129883, |
| "learning_rate": 5.962059620596207e-06, |
| "loss": 3.7806, |
| "step": 380 |
| }, |
| { |
| "epoch": 9.512195121951219, |
| "grad_norm": 12.452446937561035, |
| "learning_rate": 5.826558265582656e-06, |
| "loss": 3.6793, |
| "step": 390 |
| }, |
| { |
| "epoch": 9.75609756097561, |
| "grad_norm": 12.078032493591309, |
| "learning_rate": 5.691056910569106e-06, |
| "loss": 3.9065, |
| "step": 400 |
| }, |
| { |
| "epoch": 10.0, |
| "grad_norm": 15.963644027709961, |
| "learning_rate": 5.555555555555557e-06, |
| "loss": 3.7498, |
| "step": 410 |
| }, |
| { |
| "epoch": 10.24390243902439, |
| "grad_norm": 12.582329750061035, |
| "learning_rate": 5.420054200542005e-06, |
| "loss": 3.85, |
| "step": 420 |
| }, |
| { |
| "epoch": 10.487804878048781, |
| "grad_norm": 12.392144203186035, |
| "learning_rate": 5.2845528455284555e-06, |
| "loss": 3.8093, |
| "step": 430 |
| }, |
| { |
| "epoch": 10.731707317073171, |
| "grad_norm": 13.39164924621582, |
| "learning_rate": 5.149051490514906e-06, |
| "loss": 3.829, |
| "step": 440 |
| }, |
| { |
| "epoch": 10.975609756097562, |
| "grad_norm": 11.60807991027832, |
| "learning_rate": 5.013550135501355e-06, |
| "loss": 3.8262, |
| "step": 450 |
| }, |
| { |
| "epoch": 11.21951219512195, |
| "grad_norm": 11.956347465515137, |
| "learning_rate": 4.8780487804878055e-06, |
| "loss": 3.6282, |
| "step": 460 |
| }, |
| { |
| "epoch": 11.463414634146341, |
| "grad_norm": 12.04183292388916, |
| "learning_rate": 4.742547425474256e-06, |
| "loss": 3.7594, |
| "step": 470 |
| }, |
| { |
| "epoch": 11.707317073170731, |
| "grad_norm": 11.708353042602539, |
| "learning_rate": 4.607046070460705e-06, |
| "loss": 3.6854, |
| "step": 480 |
| }, |
| { |
| "epoch": 11.951219512195122, |
| "grad_norm": 12.383793830871582, |
| "learning_rate": 4.471544715447155e-06, |
| "loss": 3.8401, |
| "step": 490 |
| }, |
| { |
| "epoch": 12.195121951219512, |
| "grad_norm": 13.459123611450195, |
| "learning_rate": 4.336043360433605e-06, |
| "loss": 3.6459, |
| "step": 500 |
| }, |
| { |
| "epoch": 12.439024390243903, |
| "grad_norm": 12.854645729064941, |
| "learning_rate": 4.200542005420054e-06, |
| "loss": 3.5457, |
| "step": 510 |
| }, |
| { |
| "epoch": 12.682926829268293, |
| "grad_norm": 12.653203964233398, |
| "learning_rate": 4.0650406504065046e-06, |
| "loss": 3.8569, |
| "step": 520 |
| }, |
| { |
| "epoch": 12.926829268292684, |
| "grad_norm": 11.430753707885742, |
| "learning_rate": 3.929539295392954e-06, |
| "loss": 3.7373, |
| "step": 530 |
| }, |
| { |
| "epoch": 13.170731707317072, |
| "grad_norm": 12.394974708557129, |
| "learning_rate": 3.7940379403794043e-06, |
| "loss": 3.5982, |
| "step": 540 |
| }, |
| { |
| "epoch": 13.414634146341463, |
| "grad_norm": 12.253921508789062, |
| "learning_rate": 3.6585365853658537e-06, |
| "loss": 3.7153, |
| "step": 550 |
| }, |
| { |
| "epoch": 13.658536585365853, |
| "grad_norm": 12.761003494262695, |
| "learning_rate": 3.5230352303523035e-06, |
| "loss": 3.5524, |
| "step": 560 |
| }, |
| { |
| "epoch": 13.902439024390244, |
| "grad_norm": 15.800996780395508, |
| "learning_rate": 3.387533875338754e-06, |
| "loss": 3.8582, |
| "step": 570 |
| }, |
| { |
| "epoch": 14.146341463414634, |
| "grad_norm": 14.060684204101562, |
| "learning_rate": 3.2520325203252037e-06, |
| "loss": 3.5614, |
| "step": 580 |
| }, |
| { |
| "epoch": 14.390243902439025, |
| "grad_norm": 12.352811813354492, |
| "learning_rate": 3.116531165311653e-06, |
| "loss": 3.5614, |
| "step": 590 |
| }, |
| { |
| "epoch": 14.634146341463415, |
| "grad_norm": 12.797225952148438, |
| "learning_rate": 2.9810298102981034e-06, |
| "loss": 3.6201, |
| "step": 600 |
| }, |
| { |
| "epoch": 14.878048780487806, |
| "grad_norm": 13.079516410827637, |
| "learning_rate": 2.845528455284553e-06, |
| "loss": 3.8288, |
| "step": 610 |
| } |
| ], |
| "logging_steps": 10, |
| "max_steps": 820, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 20, |
| "save_steps": 500, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": false |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 0.0, |
| "train_batch_size": 64, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
