| { | |
| "best_metric": 24.56800162634682, | |
| "best_model_checkpoint": "./checkpoint-1000", | |
| "epoch": 166.004, | |
| "eval_steps": 500, | |
| "global_step": 2500, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 4.6e-07, | |
| "loss": 1.2564, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 9.6e-07, | |
| "loss": 1.0031, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 4.01, | |
| "learning_rate": 9.906122448979592e-07, | |
| "loss": 0.5375, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "learning_rate": 9.80408163265306e-07, | |
| "loss": 0.3396, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "learning_rate": 9.70204081632653e-07, | |
| "loss": 0.2754, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 9.01, | |
| "learning_rate": 9.6e-07, | |
| "loss": 0.2336, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "learning_rate": 9.497959183673469e-07, | |
| "loss": 0.197, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "learning_rate": 9.395918367346939e-07, | |
| "loss": 0.1695, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 14.01, | |
| "learning_rate": 9.293877551020408e-07, | |
| "loss": 0.1478, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "learning_rate": 9.191836734693877e-07, | |
| "loss": 0.1242, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "learning_rate": 9.089795918367347e-07, | |
| "loss": 0.1056, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 19.01, | |
| "learning_rate": 8.987755102040815e-07, | |
| "loss": 0.0889, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 21.0, | |
| "learning_rate": 8.885714285714285e-07, | |
| "loss": 0.0733, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "learning_rate": 8.783673469387755e-07, | |
| "loss": 0.0592, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 24.01, | |
| "learning_rate": 8.681632653061224e-07, | |
| "loss": 0.0502, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 26.0, | |
| "learning_rate": 8.579591836734693e-07, | |
| "loss": 0.0383, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 28.0, | |
| "learning_rate": 8.477551020408162e-07, | |
| "loss": 0.0313, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 29.01, | |
| "learning_rate": 8.375510204081633e-07, | |
| "loss": 0.0244, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 31.0, | |
| "learning_rate": 8.273469387755102e-07, | |
| "loss": 0.0199, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 33.0, | |
| "learning_rate": 8.171428571428571e-07, | |
| "loss": 0.0174, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 33.0, | |
| "eval_loss": 0.6206670999526978, | |
| "eval_runtime": 252.4938, | |
| "eval_samples_per_second": 5.252, | |
| "eval_steps_per_second": 0.166, | |
| "eval_wer": 24.66964830250051, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 34.01, | |
| "learning_rate": 8.06938775510204e-07, | |
| "loss": 0.015, | |
| "step": 525 | |
| }, | |
| { | |
| "epoch": 36.0, | |
| "learning_rate": 7.96734693877551e-07, | |
| "loss": 0.0134, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 38.0, | |
| "learning_rate": 7.86530612244898e-07, | |
| "loss": 0.012, | |
| "step": 575 | |
| }, | |
| { | |
| "epoch": 39.01, | |
| "learning_rate": 7.763265306122448e-07, | |
| "loss": 0.011, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 41.0, | |
| "learning_rate": 7.661224489795918e-07, | |
| "loss": 0.01, | |
| "step": 625 | |
| }, | |
| { | |
| "epoch": 43.0, | |
| "learning_rate": 7.559183673469387e-07, | |
| "loss": 0.0092, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 44.01, | |
| "learning_rate": 7.457142857142856e-07, | |
| "loss": 0.0087, | |
| "step": 675 | |
| }, | |
| { | |
| "epoch": 46.0, | |
| "learning_rate": 7.355102040816326e-07, | |
| "loss": 0.0081, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 48.0, | |
| "learning_rate": 7.253061224489796e-07, | |
| "loss": 0.0075, | |
| "step": 725 | |
| }, | |
| { | |
| "epoch": 49.01, | |
| "learning_rate": 7.151020408163266e-07, | |
| "loss": 0.0072, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 51.0, | |
| "learning_rate": 7.048979591836734e-07, | |
| "loss": 0.0068, | |
| "step": 775 | |
| }, | |
| { | |
| "epoch": 53.0, | |
| "learning_rate": 6.946938775510204e-07, | |
| "loss": 0.0064, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 54.01, | |
| "learning_rate": 6.844897959183673e-07, | |
| "loss": 0.0061, | |
| "step": 825 | |
| }, | |
| { | |
| "epoch": 56.0, | |
| "learning_rate": 6.742857142857143e-07, | |
| "loss": 0.0058, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 58.0, | |
| "learning_rate": 6.640816326530611e-07, | |
| "loss": 0.0055, | |
| "step": 875 | |
| }, | |
| { | |
| "epoch": 59.01, | |
| "learning_rate": 6.538775510204081e-07, | |
| "loss": 0.0054, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 61.0, | |
| "learning_rate": 6.43673469387755e-07, | |
| "loss": 0.0051, | |
| "step": 925 | |
| }, | |
| { | |
| "epoch": 63.0, | |
| "learning_rate": 6.334693877551021e-07, | |
| "loss": 0.0049, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 64.01, | |
| "learning_rate": 6.23265306122449e-07, | |
| "loss": 0.0047, | |
| "step": 975 | |
| }, | |
| { | |
| "epoch": 66.0, | |
| "learning_rate": 6.130612244897959e-07, | |
| "loss": 0.0045, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 66.0, | |
| "eval_loss": 0.6704666614532471, | |
| "eval_runtime": 250.1153, | |
| "eval_samples_per_second": 5.302, | |
| "eval_steps_per_second": 0.168, | |
| "eval_wer": 24.56800162634682, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 68.0, | |
| "learning_rate": 6.028571428571429e-07, | |
| "loss": 0.0044, | |
| "step": 1025 | |
| }, | |
| { | |
| "epoch": 69.01, | |
| "learning_rate": 5.926530612244898e-07, | |
| "loss": 0.0042, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 71.0, | |
| "learning_rate": 5.824489795918367e-07, | |
| "loss": 0.0041, | |
| "step": 1075 | |
| }, | |
| { | |
| "epoch": 73.0, | |
| "learning_rate": 5.722448979591836e-07, | |
| "loss": 0.004, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 74.01, | |
| "learning_rate": 5.620408163265306e-07, | |
| "loss": 0.0038, | |
| "step": 1125 | |
| }, | |
| { | |
| "epoch": 76.0, | |
| "learning_rate": 5.518367346938774e-07, | |
| "loss": 0.0037, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 78.0, | |
| "learning_rate": 5.416326530612245e-07, | |
| "loss": 0.0036, | |
| "step": 1175 | |
| }, | |
| { | |
| "epoch": 79.01, | |
| "learning_rate": 5.314285714285715e-07, | |
| "loss": 0.0035, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 81.0, | |
| "learning_rate": 5.212244897959184e-07, | |
| "loss": 0.0034, | |
| "step": 1225 | |
| }, | |
| { | |
| "epoch": 83.0, | |
| "learning_rate": 5.110204081632653e-07, | |
| "loss": 0.0033, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 84.01, | |
| "learning_rate": 5.008163265306122e-07, | |
| "loss": 0.0033, | |
| "step": 1275 | |
| }, | |
| { | |
| "epoch": 86.0, | |
| "learning_rate": 4.906122448979592e-07, | |
| "loss": 0.0031, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 88.0, | |
| "learning_rate": 4.804081632653061e-07, | |
| "loss": 0.0031, | |
| "step": 1325 | |
| }, | |
| { | |
| "epoch": 89.01, | |
| "learning_rate": 4.7020408163265303e-07, | |
| "loss": 0.0031, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 91.0, | |
| "learning_rate": 4.6e-07, | |
| "loss": 0.003, | |
| "step": 1375 | |
| }, | |
| { | |
| "epoch": 93.0, | |
| "learning_rate": 4.4979591836734695e-07, | |
| "loss": 0.0029, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 94.01, | |
| "learning_rate": 4.3959183673469385e-07, | |
| "loss": 0.0029, | |
| "step": 1425 | |
| }, | |
| { | |
| "epoch": 96.0, | |
| "learning_rate": 4.293877551020408e-07, | |
| "loss": 0.0028, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 98.0, | |
| "learning_rate": 4.191836734693877e-07, | |
| "loss": 0.0027, | |
| "step": 1475 | |
| }, | |
| { | |
| "epoch": 99.01, | |
| "learning_rate": 4.089795918367347e-07, | |
| "loss": 0.0027, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 99.01, | |
| "eval_loss": 0.6944558024406433, | |
| "eval_runtime": 263.7584, | |
| "eval_samples_per_second": 5.027, | |
| "eval_steps_per_second": 0.159, | |
| "eval_wer": 25.27952835942265, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 101.0, | |
| "learning_rate": 3.987755102040816e-07, | |
| "loss": 0.0026, | |
| "step": 1525 | |
| }, | |
| { | |
| "epoch": 103.0, | |
| "learning_rate": 3.8857142857142853e-07, | |
| "loss": 0.0026, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 104.01, | |
| "learning_rate": 3.783673469387755e-07, | |
| "loss": 0.0025, | |
| "step": 1575 | |
| }, | |
| { | |
| "epoch": 106.0, | |
| "learning_rate": 3.681632653061224e-07, | |
| "loss": 0.0025, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 108.0, | |
| "learning_rate": 3.579591836734694e-07, | |
| "loss": 0.0024, | |
| "step": 1625 | |
| }, | |
| { | |
| "epoch": 109.01, | |
| "learning_rate": 3.477551020408163e-07, | |
| "loss": 0.0024, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 111.0, | |
| "learning_rate": 3.3755102040816326e-07, | |
| "loss": 0.0023, | |
| "step": 1675 | |
| }, | |
| { | |
| "epoch": 113.0, | |
| "learning_rate": 3.2734693877551016e-07, | |
| "loss": 0.0023, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 114.01, | |
| "learning_rate": 3.171428571428571e-07, | |
| "loss": 0.0023, | |
| "step": 1725 | |
| }, | |
| { | |
| "epoch": 116.0, | |
| "learning_rate": 3.0693877551020407e-07, | |
| "loss": 0.0023, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 118.0, | |
| "learning_rate": 2.9673469387755103e-07, | |
| "loss": 0.0022, | |
| "step": 1775 | |
| }, | |
| { | |
| "epoch": 119.01, | |
| "learning_rate": 2.8653061224489793e-07, | |
| "loss": 0.0022, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 121.0, | |
| "learning_rate": 2.763265306122449e-07, | |
| "loss": 0.0022, | |
| "step": 1825 | |
| }, | |
| { | |
| "epoch": 123.0, | |
| "learning_rate": 2.6612244897959185e-07, | |
| "loss": 0.0021, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 124.01, | |
| "learning_rate": 2.559183673469388e-07, | |
| "loss": 0.0021, | |
| "step": 1875 | |
| }, | |
| { | |
| "epoch": 126.0, | |
| "learning_rate": 2.457142857142857e-07, | |
| "loss": 0.0021, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 128.0, | |
| "learning_rate": 2.3551020408163264e-07, | |
| "loss": 0.0021, | |
| "step": 1925 | |
| }, | |
| { | |
| "epoch": 129.01, | |
| "learning_rate": 2.253061224489796e-07, | |
| "loss": 0.002, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 131.0, | |
| "learning_rate": 2.1510204081632652e-07, | |
| "loss": 0.002, | |
| "step": 1975 | |
| }, | |
| { | |
| "epoch": 133.0, | |
| "learning_rate": 2.0489795918367345e-07, | |
| "loss": 0.002, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 133.0, | |
| "eval_loss": 0.7078843116760254, | |
| "eval_runtime": 279.582, | |
| "eval_samples_per_second": 4.743, | |
| "eval_steps_per_second": 0.15, | |
| "eval_wer": 26.478959138036185, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 134.01, | |
| "learning_rate": 1.946938775510204e-07, | |
| "loss": 0.002, | |
| "step": 2025 | |
| }, | |
| { | |
| "epoch": 136.0, | |
| "learning_rate": 1.8448979591836734e-07, | |
| "loss": 0.002, | |
| "step": 2050 | |
| }, | |
| { | |
| "epoch": 138.0, | |
| "learning_rate": 1.742857142857143e-07, | |
| "loss": 0.0019, | |
| "step": 2075 | |
| }, | |
| { | |
| "epoch": 139.01, | |
| "learning_rate": 1.6408163265306123e-07, | |
| "loss": 0.0019, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 141.0, | |
| "learning_rate": 1.5387755102040818e-07, | |
| "loss": 0.0019, | |
| "step": 2125 | |
| }, | |
| { | |
| "epoch": 143.0, | |
| "learning_rate": 1.4367346938775509e-07, | |
| "loss": 0.0019, | |
| "step": 2150 | |
| }, | |
| { | |
| "epoch": 144.01, | |
| "learning_rate": 1.3346938775510202e-07, | |
| "loss": 0.0019, | |
| "step": 2175 | |
| }, | |
| { | |
| "epoch": 146.0, | |
| "learning_rate": 1.2326530612244897e-07, | |
| "loss": 0.0019, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 148.0, | |
| "learning_rate": 1.130612244897959e-07, | |
| "loss": 0.0019, | |
| "step": 2225 | |
| }, | |
| { | |
| "epoch": 149.01, | |
| "learning_rate": 1.0285714285714285e-07, | |
| "loss": 0.0018, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 151.0, | |
| "learning_rate": 9.265306122448979e-08, | |
| "loss": 0.0018, | |
| "step": 2275 | |
| }, | |
| { | |
| "epoch": 153.0, | |
| "learning_rate": 8.244897959183673e-08, | |
| "loss": 0.0018, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 154.01, | |
| "learning_rate": 7.224489795918368e-08, | |
| "loss": 0.0018, | |
| "step": 2325 | |
| }, | |
| { | |
| "epoch": 156.0, | |
| "learning_rate": 6.20408163265306e-08, | |
| "loss": 0.0018, | |
| "step": 2350 | |
| }, | |
| { | |
| "epoch": 158.0, | |
| "learning_rate": 5.183673469387754e-08, | |
| "loss": 0.0018, | |
| "step": 2375 | |
| }, | |
| { | |
| "epoch": 159.01, | |
| "learning_rate": 4.1632653061224486e-08, | |
| "loss": 0.0018, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 161.0, | |
| "learning_rate": 3.142857142857143e-08, | |
| "loss": 0.0018, | |
| "step": 2425 | |
| }, | |
| { | |
| "epoch": 163.0, | |
| "learning_rate": 2.122448979591837e-08, | |
| "loss": 0.0018, | |
| "step": 2450 | |
| }, | |
| { | |
| "epoch": 164.01, | |
| "learning_rate": 1.1020408163265306e-08, | |
| "loss": 0.0018, | |
| "step": 2475 | |
| }, | |
| { | |
| "epoch": 166.0, | |
| "learning_rate": 8.163265306122449e-10, | |
| "loss": 0.0018, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 166.0, | |
| "eval_loss": 0.7126978635787964, | |
| "eval_runtime": 279.4326, | |
| "eval_samples_per_second": 4.745, | |
| "eval_steps_per_second": 0.15, | |
| "eval_wer": 26.397641797113238, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 166.0, | |
| "step": 2500, | |
| "total_flos": 4.579042263957504e+19, | |
| "train_loss": 0.05099657071977854, | |
| "train_runtime": 22043.5536, | |
| "train_samples_per_second": 7.258, | |
| "train_steps_per_second": 0.113 | |
| } | |
| ], | |
| "logging_steps": 25, | |
| "max_steps": 2500, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 9223372036854775807, | |
| "save_steps": 500, | |
| "total_flos": 4.579042263957504e+19, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |