| { | |
| "best_metric": 0.03033074364066124, | |
| "best_model_checkpoint": "runs/deepseek_lora_20240423-133229/checkpoint-1000", | |
| "epoch": 0.31330774653403304, | |
| "eval_steps": 500, | |
| "global_step": 1000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "grad_norm": 1.5461480617523193, | |
| "learning_rate": 4.0000000000000003e-07, | |
| "loss": 1.5239, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 2.797250747680664, | |
| "learning_rate": 8.000000000000001e-07, | |
| "loss": 1.4966, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 4.274460315704346, | |
| "learning_rate": 1.2000000000000002e-06, | |
| "loss": 1.4788, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 1.9981650114059448, | |
| "learning_rate": 1.6000000000000001e-06, | |
| "loss": 1.4001, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "grad_norm": 1.2322547435760498, | |
| "learning_rate": 2.0000000000000003e-06, | |
| "loss": 1.4683, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "grad_norm": 1.250878930091858, | |
| "learning_rate": 2.4000000000000003e-06, | |
| "loss": 1.1961, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "grad_norm": 4.917982578277588, | |
| "learning_rate": 2.8000000000000003e-06, | |
| "loss": 1.4261, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "grad_norm": 3.885340929031372, | |
| "learning_rate": 3.2000000000000003e-06, | |
| "loss": 1.2742, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "grad_norm": 1.170296311378479, | |
| "learning_rate": 3.6000000000000003e-06, | |
| "loss": 1.253, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "grad_norm": 6.992724895477295, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 1.2521, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "grad_norm": 0.9540568590164185, | |
| "learning_rate": 4.4e-06, | |
| "loss": 1.0828, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 0.9779148697853088, | |
| "learning_rate": 4.800000000000001e-06, | |
| "loss": 1.1255, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 1.2271130084991455, | |
| "learning_rate": 5.2e-06, | |
| "loss": 0.9496, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 2.1939122676849365, | |
| "learning_rate": 5.600000000000001e-06, | |
| "loss": 0.953, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 1.5918376445770264, | |
| "learning_rate": 6e-06, | |
| "loss": 0.9348, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 0.772602379322052, | |
| "learning_rate": 6.4000000000000006e-06, | |
| "loss": 0.7762, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 0.8226258158683777, | |
| "learning_rate": 6.800000000000001e-06, | |
| "loss": 0.6798, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "grad_norm": 1.0945135354995728, | |
| "learning_rate": 7.2000000000000005e-06, | |
| "loss": 0.5936, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "grad_norm": 0.6972842812538147, | |
| "learning_rate": 7.600000000000001e-06, | |
| "loss": 0.4161, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "grad_norm": 3.9560859203338623, | |
| "learning_rate": 8.000000000000001e-06, | |
| "loss": 0.3959, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 0.4297197759151459, | |
| "learning_rate": 8.400000000000001e-06, | |
| "loss": 0.3166, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 0.4494841992855072, | |
| "learning_rate": 8.8e-06, | |
| "loss": 0.2056, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 0.6981032490730286, | |
| "learning_rate": 9.200000000000002e-06, | |
| "loss": 0.3024, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 0.5783653259277344, | |
| "learning_rate": 9.600000000000001e-06, | |
| "loss": 0.4392, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 0.33407163619995117, | |
| "learning_rate": 1e-05, | |
| "loss": 0.3708, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 0.3096323311328888, | |
| "learning_rate": 1.04e-05, | |
| "loss": 0.1637, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 1.6187375783920288, | |
| "learning_rate": 1.0800000000000002e-05, | |
| "loss": 0.1078, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "grad_norm": 0.27954748272895813, | |
| "learning_rate": 1.1200000000000001e-05, | |
| "loss": 0.2668, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "grad_norm": 0.4082590937614441, | |
| "learning_rate": 1.16e-05, | |
| "loss": 0.1843, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "grad_norm": 0.2993428707122803, | |
| "learning_rate": 1.2e-05, | |
| "loss": 0.1969, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 3.0335004329681396, | |
| "learning_rate": 1.2400000000000002e-05, | |
| "loss": 0.122, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 0.0972365066409111, | |
| "learning_rate": 1.2800000000000001e-05, | |
| "loss": 0.1354, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 0.26203760504722595, | |
| "learning_rate": 1.3200000000000002e-05, | |
| "loss": 0.17, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "grad_norm": 5.179258346557617, | |
| "learning_rate": 1.3600000000000002e-05, | |
| "loss": 0.1409, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "grad_norm": 0.032190825790166855, | |
| "learning_rate": 1.4e-05, | |
| "loss": 0.076, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "grad_norm": 0.02556140348315239, | |
| "learning_rate": 1.4400000000000001e-05, | |
| "loss": 0.0468, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "grad_norm": 5.667189121246338, | |
| "learning_rate": 1.48e-05, | |
| "loss": 0.1345, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "grad_norm": 0.07244217395782471, | |
| "learning_rate": 1.5200000000000002e-05, | |
| "loss": 0.0411, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "grad_norm": 0.11255475878715515, | |
| "learning_rate": 1.5600000000000003e-05, | |
| "loss": 0.1907, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 1.9602845907211304, | |
| "learning_rate": 1.6000000000000003e-05, | |
| "loss": 0.0717, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 0.013551376760005951, | |
| "learning_rate": 1.64e-05, | |
| "loss": 0.0808, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 0.14421305060386658, | |
| "learning_rate": 1.6800000000000002e-05, | |
| "loss": 0.0604, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 0.01705547235906124, | |
| "learning_rate": 1.72e-05, | |
| "loss": 0.1742, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "grad_norm": 0.021360402926802635, | |
| "learning_rate": 1.76e-05, | |
| "loss": 0.0886, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "grad_norm": 0.04369840770959854, | |
| "learning_rate": 1.8e-05, | |
| "loss": 0.1063, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "grad_norm": 2.2816083431243896, | |
| "learning_rate": 1.8400000000000003e-05, | |
| "loss": 0.1061, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 0.00989863183349371, | |
| "learning_rate": 1.88e-05, | |
| "loss": 0.2317, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 0.014644264243543148, | |
| "learning_rate": 1.9200000000000003e-05, | |
| "loss": 0.1054, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 0.13385985791683197, | |
| "learning_rate": 1.9600000000000002e-05, | |
| "loss": 0.1196, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 2.099637985229492, | |
| "learning_rate": 2e-05, | |
| "loss": 0.1584, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "eval_loss": 0.05546905845403671, | |
| "eval_runtime": 62.1757, | |
| "eval_samples_per_second": 16.083, | |
| "eval_steps_per_second": 16.083, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 0.013259253464639187, | |
| "learning_rate": 1.9955555555555557e-05, | |
| "loss": 0.0373, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 0.1753215193748474, | |
| "learning_rate": 1.9911111111111112e-05, | |
| "loss": 0.0986, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "grad_norm": 0.006998735014349222, | |
| "learning_rate": 1.9866666666666667e-05, | |
| "loss": 0.1372, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "grad_norm": 0.3990759253501892, | |
| "learning_rate": 1.9822222222222226e-05, | |
| "loss": 0.127, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "grad_norm": 1.8248735666275024, | |
| "learning_rate": 1.977777777777778e-05, | |
| "loss": 0.1239, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 0.005832708440721035, | |
| "learning_rate": 1.9733333333333336e-05, | |
| "loss": 0.1181, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 0.10705375671386719, | |
| "learning_rate": 1.968888888888889e-05, | |
| "loss": 0.0972, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 3.7429046630859375, | |
| "learning_rate": 1.9644444444444447e-05, | |
| "loss": 0.1543, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 0.010058793239295483, | |
| "learning_rate": 1.9600000000000002e-05, | |
| "loss": 0.1146, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "grad_norm": 0.012124909088015556, | |
| "learning_rate": 1.9555555555555557e-05, | |
| "loss": 0.1582, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "grad_norm": 0.012700803577899933, | |
| "learning_rate": 1.9511111111111113e-05, | |
| "loss": 0.0711, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "grad_norm": 0.2153177261352539, | |
| "learning_rate": 1.9466666666666668e-05, | |
| "loss": 0.0452, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 0.01289213914424181, | |
| "learning_rate": 1.9422222222222223e-05, | |
| "loss": 0.067, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 0.007021366152912378, | |
| "learning_rate": 1.9377777777777778e-05, | |
| "loss": 0.0776, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 0.005274974275380373, | |
| "learning_rate": 1.9333333333333333e-05, | |
| "loss": 0.2159, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "grad_norm": 1.6235843896865845, | |
| "learning_rate": 1.928888888888889e-05, | |
| "loss": 0.0664, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "grad_norm": 0.009710369631648064, | |
| "learning_rate": 1.9244444444444444e-05, | |
| "loss": 0.1125, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "grad_norm": 0.0918051227927208, | |
| "learning_rate": 1.9200000000000003e-05, | |
| "loss": 0.0755, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 0.012988925911486149, | |
| "learning_rate": 1.9155555555555558e-05, | |
| "loss": 0.095, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 0.09063850343227386, | |
| "learning_rate": 1.9111111111111113e-05, | |
| "loss": 0.0419, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 3.1540229320526123, | |
| "learning_rate": 1.9066666666666668e-05, | |
| "loss": 0.0338, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "grad_norm": 0.009780442342162132, | |
| "learning_rate": 1.9022222222222223e-05, | |
| "loss": 0.0935, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "grad_norm": 0.011248274706304073, | |
| "learning_rate": 1.897777777777778e-05, | |
| "loss": 0.017, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "grad_norm": 0.03275947645306587, | |
| "learning_rate": 1.8933333333333334e-05, | |
| "loss": 0.1128, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "grad_norm": 2.224815607070923, | |
| "learning_rate": 1.888888888888889e-05, | |
| "loss": 0.1213, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 1.9777376651763916, | |
| "learning_rate": 1.8844444444444444e-05, | |
| "loss": 0.1128, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 7.6119561195373535, | |
| "learning_rate": 1.88e-05, | |
| "loss": 0.0557, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 3.7092978954315186, | |
| "learning_rate": 1.8755555555555558e-05, | |
| "loss": 0.1196, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 0.08959495276212692, | |
| "learning_rate": 1.8711111111111113e-05, | |
| "loss": 0.0231, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 2.716322183609009, | |
| "learning_rate": 1.866666666666667e-05, | |
| "loss": 0.1253, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 0.006216592621058226, | |
| "learning_rate": 1.8622222222222224e-05, | |
| "loss": 0.0377, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 0.005503402091562748, | |
| "learning_rate": 1.857777777777778e-05, | |
| "loss": 0.1072, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 0.4769572913646698, | |
| "learning_rate": 1.8533333333333334e-05, | |
| "loss": 0.0888, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 0.034854818135499954, | |
| "learning_rate": 1.848888888888889e-05, | |
| "loss": 0.0424, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "grad_norm": 0.00536635797470808, | |
| "learning_rate": 1.8444444444444448e-05, | |
| "loss": 0.0161, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "grad_norm": 0.002790963975712657, | |
| "learning_rate": 1.8400000000000003e-05, | |
| "loss": 0.025, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "grad_norm": 0.003568426240235567, | |
| "learning_rate": 1.835555555555556e-05, | |
| "loss": 0.0597, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "grad_norm": 0.005008562933653593, | |
| "learning_rate": 1.8311111111111114e-05, | |
| "loss": 0.1817, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "grad_norm": 0.006756258197128773, | |
| "learning_rate": 1.826666666666667e-05, | |
| "loss": 0.0084, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "grad_norm": 2.3732385635375977, | |
| "learning_rate": 1.8222222222222224e-05, | |
| "loss": 0.072, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 0.003554440103471279, | |
| "learning_rate": 1.817777777777778e-05, | |
| "loss": 0.0507, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 0.06785377860069275, | |
| "learning_rate": 1.8133333333333335e-05, | |
| "loss": 0.0428, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 0.009546133689582348, | |
| "learning_rate": 1.808888888888889e-05, | |
| "loss": 0.0179, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 4.595461368560791, | |
| "learning_rate": 1.8044444444444445e-05, | |
| "loss": 0.0506, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 2.310696601867676, | |
| "learning_rate": 1.8e-05, | |
| "loss": 0.0719, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 1.8953649997711182, | |
| "learning_rate": 1.7955555555555556e-05, | |
| "loss": 0.1718, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 1.706213116645813, | |
| "learning_rate": 1.791111111111111e-05, | |
| "loss": 0.0807, | |
| "step": 970 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "grad_norm": 0.06022366136312485, | |
| "learning_rate": 1.7866666666666666e-05, | |
| "loss": 0.0792, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "grad_norm": 0.033781811594963074, | |
| "learning_rate": 1.782222222222222e-05, | |
| "loss": 0.0864, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "grad_norm": 0.014879009686410427, | |
| "learning_rate": 1.7777777777777777e-05, | |
| "loss": 0.0527, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "eval_loss": 0.03033074364066124, | |
| "eval_runtime": 62.1716, | |
| "eval_samples_per_second": 16.085, | |
| "eval_steps_per_second": 16.085, | |
| "step": 1000 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 5000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 2, | |
| "save_steps": 1000, | |
| "total_flos": 1.6102125993984e+16, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |