| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 100.0, | |
| "global_step": 8400, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.19, | |
| "learning_rate": 5.88e-05, | |
| "loss": 6.951, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 2.38, | |
| "learning_rate": 0.0001188, | |
| "loss": 3.1694, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.57, | |
| "learning_rate": 0.00017879999999999998, | |
| "loss": 3.0357, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 4.76, | |
| "learning_rate": 0.0002388, | |
| "loss": 2.8568, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 5.95, | |
| "learning_rate": 0.0002988, | |
| "loss": 1.8805, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 5.95, | |
| "eval_loss": 0.7547174692153931, | |
| "eval_runtime": 47.8349, | |
| "eval_samples_per_second": 25.609, | |
| "eval_steps_per_second": 0.815, | |
| "eval_wer": 0.8438120450033091, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 7.14, | |
| "learning_rate": 0.0002962784810126582, | |
| "loss": 1.5197, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 8.33, | |
| "learning_rate": 0.0002924810126582278, | |
| "loss": 1.3887, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 9.52, | |
| "learning_rate": 0.00028868354430379743, | |
| "loss": 1.3124, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 10.71, | |
| "learning_rate": 0.00028488607594936706, | |
| "loss": 1.2433, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 11.9, | |
| "learning_rate": 0.0002810886075949367, | |
| "loss": 1.2123, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 11.9, | |
| "eval_loss": 0.47322335839271545, | |
| "eval_runtime": 48.1093, | |
| "eval_samples_per_second": 25.463, | |
| "eval_steps_per_second": 0.811, | |
| "eval_wer": 0.6542025148908008, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 13.1, | |
| "learning_rate": 0.0002772911392405063, | |
| "loss": 1.1794, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 14.29, | |
| "learning_rate": 0.0002734936708860759, | |
| "loss": 1.1509, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 15.48, | |
| "learning_rate": 0.0002696962025316455, | |
| "loss": 1.1197, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 16.67, | |
| "learning_rate": 0.00026589873417721515, | |
| "loss": 1.0924, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 17.86, | |
| "learning_rate": 0.0002621012658227848, | |
| "loss": 1.0822, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 17.86, | |
| "eval_loss": 0.402739018201828, | |
| "eval_runtime": 47.5558, | |
| "eval_samples_per_second": 25.759, | |
| "eval_steps_per_second": 0.82, | |
| "eval_wer": 0.5777630708140304, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 19.05, | |
| "learning_rate": 0.0002583037974683544, | |
| "loss": 1.0552, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 20.24, | |
| "learning_rate": 0.000254506329113924, | |
| "loss": 1.0453, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 21.43, | |
| "learning_rate": 0.00025070886075949367, | |
| "loss": 1.0218, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 22.62, | |
| "learning_rate": 0.0002469113924050633, | |
| "loss": 1.024, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 23.81, | |
| "learning_rate": 0.00024311392405063287, | |
| "loss": 0.9938, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 23.81, | |
| "eval_loss": 0.38469693064689636, | |
| "eval_runtime": 47.562, | |
| "eval_samples_per_second": 25.756, | |
| "eval_steps_per_second": 0.82, | |
| "eval_wer": 0.5523935583498787, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "learning_rate": 0.0002393164556962025, | |
| "loss": 0.9897, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 26.19, | |
| "learning_rate": 0.00023551898734177216, | |
| "loss": 0.9857, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 27.38, | |
| "learning_rate": 0.00023172151898734174, | |
| "loss": 0.9537, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 28.57, | |
| "learning_rate": 0.0002279240506329114, | |
| "loss": 0.941, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 29.76, | |
| "learning_rate": 0.000224126582278481, | |
| "loss": 0.9383, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 29.76, | |
| "eval_loss": 0.3845490515232086, | |
| "eval_runtime": 47.7826, | |
| "eval_samples_per_second": 25.637, | |
| "eval_steps_per_second": 0.816, | |
| "eval_wer": 0.5204059121994264, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 30.95, | |
| "learning_rate": 0.00022032911392405062, | |
| "loss": 0.9304, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 32.14, | |
| "learning_rate": 0.00021653164556962025, | |
| "loss": 0.925, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 33.33, | |
| "learning_rate": 0.00021273417721518986, | |
| "loss": 0.9119, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 34.52, | |
| "learning_rate": 0.00020893670886075949, | |
| "loss": 0.8988, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 35.71, | |
| "learning_rate": 0.0002051392405063291, | |
| "loss": 0.8932, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 35.71, | |
| "eval_loss": 0.3832888603210449, | |
| "eval_runtime": 47.4524, | |
| "eval_samples_per_second": 25.815, | |
| "eval_steps_per_second": 0.822, | |
| "eval_wer": 0.529671299360247, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 36.9, | |
| "learning_rate": 0.00020134177215189872, | |
| "loss": 0.8709, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 38.1, | |
| "learning_rate": 0.00019754430379746835, | |
| "loss": 0.8767, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 39.29, | |
| "learning_rate": 0.00019374683544303795, | |
| "loss": 0.8677, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 40.48, | |
| "learning_rate": 0.00018994936708860758, | |
| "loss": 0.8642, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 41.67, | |
| "learning_rate": 0.00018615189873417718, | |
| "loss": 0.8495, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 41.67, | |
| "eval_loss": 0.37586063146591187, | |
| "eval_runtime": 47.6167, | |
| "eval_samples_per_second": 25.726, | |
| "eval_steps_per_second": 0.819, | |
| "eval_wer": 0.5036399735274653, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 42.86, | |
| "learning_rate": 0.0001823544303797468, | |
| "loss": 0.8546, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 44.05, | |
| "learning_rate": 0.00017855696202531644, | |
| "loss": 0.8455, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 45.24, | |
| "learning_rate": 0.00017475949367088604, | |
| "loss": 0.8308, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 46.43, | |
| "learning_rate": 0.0001709620253164557, | |
| "loss": 0.8179, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 47.62, | |
| "learning_rate": 0.00016716455696202527, | |
| "loss": 0.8201, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 47.62, | |
| "eval_loss": 0.36158040165901184, | |
| "eval_runtime": 47.7093, | |
| "eval_samples_per_second": 25.676, | |
| "eval_steps_per_second": 0.817, | |
| "eval_wer": 0.48588131480255903, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 48.81, | |
| "learning_rate": 0.00016340506329113924, | |
| "loss": 0.8052, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 50.0, | |
| "learning_rate": 0.00015960759493670884, | |
| "loss": 0.8005, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 51.19, | |
| "learning_rate": 0.00015581012658227847, | |
| "loss": 0.7943, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 52.38, | |
| "learning_rate": 0.00015201265822784808, | |
| "loss": 0.7811, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 53.57, | |
| "learning_rate": 0.0001482151898734177, | |
| "loss": 0.7794, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 53.57, | |
| "eval_loss": 0.38744866847991943, | |
| "eval_runtime": 47.4049, | |
| "eval_samples_per_second": 25.841, | |
| "eval_steps_per_second": 0.823, | |
| "eval_wer": 0.49382307522611957, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 54.76, | |
| "learning_rate": 0.00014441772151898733, | |
| "loss": 0.7793, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 55.95, | |
| "learning_rate": 0.00014062025316455696, | |
| "loss": 0.7729, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 57.14, | |
| "learning_rate": 0.00013682278481012657, | |
| "loss": 0.76, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 58.33, | |
| "learning_rate": 0.0001330253164556962, | |
| "loss": 0.7593, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 59.52, | |
| "learning_rate": 0.0001292278481012658, | |
| "loss": 0.735, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 59.52, | |
| "eval_loss": 0.374796986579895, | |
| "eval_runtime": 47.4755, | |
| "eval_samples_per_second": 25.803, | |
| "eval_steps_per_second": 0.821, | |
| "eval_wer": 0.47816015883520846, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 60.71, | |
| "learning_rate": 0.00012543037974683543, | |
| "loss": 0.7415, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 61.9, | |
| "learning_rate": 0.00012163291139240506, | |
| "loss": 0.7206, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 63.1, | |
| "learning_rate": 0.00011783544303797467, | |
| "loss": 0.7117, | |
| "step": 5300 | |
| }, | |
| { | |
| "epoch": 64.29, | |
| "learning_rate": 0.00011403797468354429, | |
| "loss": 0.7213, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 65.48, | |
| "learning_rate": 0.0001102405063291139, | |
| "loss": 0.7082, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 65.48, | |
| "eval_loss": 0.3615438938140869, | |
| "eval_runtime": 48.0099, | |
| "eval_samples_per_second": 25.516, | |
| "eval_steps_per_second": 0.812, | |
| "eval_wer": 0.4674608427090227, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 66.67, | |
| "learning_rate": 0.00010644303797468355, | |
| "loss": 0.705, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 67.86, | |
| "learning_rate": 0.00010264556962025316, | |
| "loss": 0.6942, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 69.05, | |
| "learning_rate": 9.884810126582278e-05, | |
| "loss": 0.6894, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 70.24, | |
| "learning_rate": 9.50506329113924e-05, | |
| "loss": 0.6807, | |
| "step": 5900 | |
| }, | |
| { | |
| "epoch": 71.43, | |
| "learning_rate": 9.129113924050632e-05, | |
| "loss": 0.669, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 71.43, | |
| "eval_loss": 0.3796931207180023, | |
| "eval_runtime": 46.7914, | |
| "eval_samples_per_second": 26.18, | |
| "eval_steps_per_second": 0.833, | |
| "eval_wer": 0.4600705934259872, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 72.62, | |
| "learning_rate": 8.749367088607594e-05, | |
| "loss": 0.6803, | |
| "step": 6100 | |
| }, | |
| { | |
| "epoch": 73.81, | |
| "learning_rate": 8.369620253164557e-05, | |
| "loss": 0.6706, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 75.0, | |
| "learning_rate": 7.989873417721518e-05, | |
| "loss": 0.6615, | |
| "step": 6300 | |
| }, | |
| { | |
| "epoch": 76.19, | |
| "learning_rate": 7.61012658227848e-05, | |
| "loss": 0.6541, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 77.38, | |
| "learning_rate": 7.230379746835442e-05, | |
| "loss": 0.6457, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 77.38, | |
| "eval_loss": 0.38117340207099915, | |
| "eval_runtime": 46.7869, | |
| "eval_samples_per_second": 26.183, | |
| "eval_steps_per_second": 0.834, | |
| "eval_wer": 0.4514670196337966, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 78.57, | |
| "learning_rate": 6.850632911392405e-05, | |
| "loss": 0.6379, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 79.76, | |
| "learning_rate": 6.470886075949366e-05, | |
| "loss": 0.6387, | |
| "step": 6700 | |
| }, | |
| { | |
| "epoch": 80.95, | |
| "learning_rate": 6.091139240506329e-05, | |
| "loss": 0.6186, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 82.14, | |
| "learning_rate": 5.711392405063291e-05, | |
| "loss": 0.6095, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 83.33, | |
| "learning_rate": 5.331645569620252e-05, | |
| "loss": 0.6098, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 83.33, | |
| "eval_loss": 0.3660217821598053, | |
| "eval_runtime": 46.6679, | |
| "eval_samples_per_second": 26.249, | |
| "eval_steps_per_second": 0.836, | |
| "eval_wer": 0.4342598720494154, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 84.52, | |
| "learning_rate": 4.9518987341772145e-05, | |
| "loss": 0.6124, | |
| "step": 7100 | |
| }, | |
| { | |
| "epoch": 85.71, | |
| "learning_rate": 4.572151898734177e-05, | |
| "loss": 0.5954, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 86.9, | |
| "learning_rate": 4.192405063291139e-05, | |
| "loss": 0.5993, | |
| "step": 7300 | |
| }, | |
| { | |
| "epoch": 88.1, | |
| "learning_rate": 3.812658227848101e-05, | |
| "loss": 0.5861, | |
| "step": 7400 | |
| }, | |
| { | |
| "epoch": 89.29, | |
| "learning_rate": 3.432911392405063e-05, | |
| "loss": 0.5874, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 89.29, | |
| "eval_loss": 0.3640037775039673, | |
| "eval_runtime": 47.0, | |
| "eval_samples_per_second": 26.064, | |
| "eval_steps_per_second": 0.83, | |
| "eval_wer": 0.4256562982572248, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 90.48, | |
| "learning_rate": 3.053164556962025e-05, | |
| "loss": 0.5818, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 91.67, | |
| "learning_rate": 2.673417721518987e-05, | |
| "loss": 0.5755, | |
| "step": 7700 | |
| }, | |
| { | |
| "epoch": 92.86, | |
| "learning_rate": 2.293670886075949e-05, | |
| "loss": 0.5837, | |
| "step": 7800 | |
| }, | |
| { | |
| "epoch": 94.05, | |
| "learning_rate": 1.9139240506329114e-05, | |
| "loss": 0.5727, | |
| "step": 7900 | |
| }, | |
| { | |
| "epoch": 95.24, | |
| "learning_rate": 1.5341772151898733e-05, | |
| "loss": 0.5627, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 95.24, | |
| "eval_loss": 0.3660779297351837, | |
| "eval_runtime": 46.6417, | |
| "eval_samples_per_second": 26.264, | |
| "eval_steps_per_second": 0.836, | |
| "eval_wer": 0.4238914626075447, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 96.43, | |
| "learning_rate": 1.1582278481012656e-05, | |
| "loss": 0.5602, | |
| "step": 8100 | |
| }, | |
| { | |
| "epoch": 97.62, | |
| "learning_rate": 7.784810126582278e-06, | |
| "loss": 0.5564, | |
| "step": 8200 | |
| }, | |
| { | |
| "epoch": 98.81, | |
| "learning_rate": 3.987341772151899e-06, | |
| "loss": 0.5714, | |
| "step": 8300 | |
| }, | |
| { | |
| "epoch": 100.0, | |
| "learning_rate": 1.8987341772151897e-07, | |
| "loss": 0.5538, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 100.0, | |
| "step": 8400, | |
| "total_flos": 4.26333780649278e+19, | |
| "train_loss": 0.9825477345784506, | |
| "train_runtime": 16451.1257, | |
| "train_samples_per_second": 16.327, | |
| "train_steps_per_second": 0.511 | |
| } | |
| ], | |
| "max_steps": 8400, | |
| "num_train_epochs": 100, | |
| "total_flos": 4.26333780649278e+19, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |