| { | |
| "best_metric": 0.1746993511915207, | |
| "best_model_checkpoint": "./checkpoint-8000", | |
| "epoch": 69.56521739130434, | |
| "global_step": 8000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.87, | |
| "learning_rate": 1.0304e-05, | |
| "loss": 6.0688, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.74, | |
| "learning_rate": 2.0204000000000002e-05, | |
| "loss": 3.1889, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.61, | |
| "learning_rate": 3.0104e-05, | |
| "loss": 2.2092, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 3.48, | |
| "learning_rate": 4.000399999999999e-05, | |
| "loss": 1.5574, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 4.35, | |
| "learning_rate": 4.9903999999999995e-05, | |
| "loss": 1.3719, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 4.35, | |
| "eval_cer": 0.08328164981226813, | |
| "eval_loss": 0.3388712704181671, | |
| "eval_runtime": 199.6026, | |
| "eval_samples_per_second": 29.068, | |
| "eval_steps_per_second": 0.456, | |
| "eval_wer": 0.42359243003445185, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 5.22, | |
| "learning_rate": 5.9804e-05, | |
| "loss": 1.3634, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 6.09, | |
| "learning_rate": 6.970400000000001e-05, | |
| "loss": 1.3345, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 6.96, | |
| "learning_rate": 7.960399999999999e-05, | |
| "loss": 1.2407, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 7.83, | |
| "learning_rate": 8e-05, | |
| "loss": 1.172, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 8.7, | |
| "learning_rate": 8e-05, | |
| "loss": 1.1361, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 8.7, | |
| "eval_cer": 0.06299695098173931, | |
| "eval_loss": 0.23091697692871094, | |
| "eval_runtime": 209.236, | |
| "eval_samples_per_second": 27.729, | |
| "eval_steps_per_second": 0.435, | |
| "eval_wer": 0.31617820567275473, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 9.57, | |
| "learning_rate": 8e-05, | |
| "loss": 1.0819, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 10.43, | |
| "learning_rate": 8e-05, | |
| "loss": 1.0923, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 11.3, | |
| "learning_rate": 8e-05, | |
| "loss": 1.0803, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 12.17, | |
| "learning_rate": 8e-05, | |
| "loss": 1.0719, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 13.04, | |
| "learning_rate": 8e-05, | |
| "loss": 1.0517, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 13.04, | |
| "eval_cer": 0.059732533136255095, | |
| "eval_loss": 0.21664097905158997, | |
| "eval_runtime": 198.1012, | |
| "eval_samples_per_second": 29.288, | |
| "eval_steps_per_second": 0.459, | |
| "eval_wer": 0.3056066827127283, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 13.91, | |
| "learning_rate": 8e-05, | |
| "loss": 1.0523, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 14.78, | |
| "learning_rate": 8e-05, | |
| "loss": 1.0414, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 15.65, | |
| "learning_rate": 8e-05, | |
| "loss": 1.031, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 16.52, | |
| "learning_rate": 8e-05, | |
| "loss": 1.0126, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 17.39, | |
| "learning_rate": 8e-05, | |
| "loss": 1.0118, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 17.39, | |
| "eval_cer": 0.055684506456416864, | |
| "eval_loss": 0.21414823830127716, | |
| "eval_runtime": 204.5017, | |
| "eval_samples_per_second": 28.371, | |
| "eval_steps_per_second": 0.445, | |
| "eval_wer": 0.2783991693803389, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 18.26, | |
| "learning_rate": 8e-05, | |
| "loss": 1.0243, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 19.13, | |
| "learning_rate": 8e-05, | |
| "loss": 1.0084, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "learning_rate": 8e-05, | |
| "loss": 1.0074, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 20.87, | |
| "learning_rate": 8e-05, | |
| "loss": 0.9933, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 21.74, | |
| "learning_rate": 8e-05, | |
| "loss": 0.9922, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 21.74, | |
| "eval_cer": 0.05935372696988491, | |
| "eval_loss": 0.22312845289707184, | |
| "eval_runtime": 198.4087, | |
| "eval_samples_per_second": 29.243, | |
| "eval_steps_per_second": 0.459, | |
| "eval_wer": 0.29411487092359245, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 22.61, | |
| "learning_rate": 8e-05, | |
| "loss": 0.9838, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 23.48, | |
| "learning_rate": 8e-05, | |
| "loss": 1.019, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 24.35, | |
| "learning_rate": 8e-05, | |
| "loss": 0.9897, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 25.22, | |
| "learning_rate": 8e-05, | |
| "loss": 0.9897, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 26.09, | |
| "learning_rate": 8e-05, | |
| "loss": 0.9929, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 26.09, | |
| "eval_cer": 0.05865182142631663, | |
| "eval_loss": 0.21711210906505585, | |
| "eval_runtime": 199.7486, | |
| "eval_samples_per_second": 29.047, | |
| "eval_steps_per_second": 0.456, | |
| "eval_wer": 0.28915946953608, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 26.96, | |
| "learning_rate": 8e-05, | |
| "loss": 0.9971, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 27.83, | |
| "learning_rate": 8e-05, | |
| "loss": 0.9621, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 28.7, | |
| "learning_rate": 8e-05, | |
| "loss": 0.9744, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 29.57, | |
| "learning_rate": 8e-05, | |
| "loss": 0.9587, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 30.43, | |
| "learning_rate": 8e-05, | |
| "loss": 0.9485, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 30.43, | |
| "eval_cer": 0.05992565000538499, | |
| "eval_loss": 0.2236333191394806, | |
| "eval_runtime": 197.9762, | |
| "eval_samples_per_second": 29.307, | |
| "eval_steps_per_second": 0.46, | |
| "eval_wer": 0.29560149133984615, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 31.3, | |
| "learning_rate": 8e-05, | |
| "loss": 0.9468, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 32.17, | |
| "learning_rate": 8e-05, | |
| "loss": 0.9598, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 33.04, | |
| "learning_rate": 8e-05, | |
| "loss": 0.9575, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 33.91, | |
| "learning_rate": 8e-05, | |
| "loss": 0.9491, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 34.78, | |
| "learning_rate": 8e-05, | |
| "loss": 0.9573, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 34.78, | |
| "eval_cer": 0.061589426108657946, | |
| "eval_loss": 0.2313707023859024, | |
| "eval_runtime": 198.2807, | |
| "eval_samples_per_second": 29.262, | |
| "eval_steps_per_second": 0.459, | |
| "eval_wer": 0.304285242342725, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 35.65, | |
| "learning_rate": 7.8176e-05, | |
| "loss": 0.9251, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 36.52, | |
| "learning_rate": 7.627600000000001e-05, | |
| "loss": 0.9299, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 37.39, | |
| "learning_rate": 7.437600000000001e-05, | |
| "loss": 0.928, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 38.26, | |
| "learning_rate": 7.2476e-05, | |
| "loss": 0.9303, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 39.13, | |
| "learning_rate": 7.057600000000001e-05, | |
| "loss": 0.9195, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 39.13, | |
| "eval_cer": 0.058024191601644466, | |
| "eval_loss": 0.21694457530975342, | |
| "eval_runtime": 198.1273, | |
| "eval_samples_per_second": 29.284, | |
| "eval_steps_per_second": 0.459, | |
| "eval_wer": 0.2812308273160602, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 40.0, | |
| "learning_rate": 6.8676e-05, | |
| "loss": 0.9182, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 40.87, | |
| "learning_rate": 6.6776e-05, | |
| "loss": 0.8885, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 41.74, | |
| "learning_rate": 6.487600000000001e-05, | |
| "loss": 0.9037, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 42.61, | |
| "learning_rate": 6.2976e-05, | |
| "loss": 0.8929, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 43.48, | |
| "learning_rate": 6.1076e-05, | |
| "loss": 0.8915, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 43.48, | |
| "eval_cer": 0.05601503340550457, | |
| "eval_loss": 0.21094831824302673, | |
| "eval_runtime": 198.7631, | |
| "eval_samples_per_second": 29.191, | |
| "eval_steps_per_second": 0.458, | |
| "eval_wer": 0.2779508235405163, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 44.35, | |
| "learning_rate": 5.9176000000000004e-05, | |
| "loss": 0.8687, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 45.22, | |
| "learning_rate": 5.7276000000000005e-05, | |
| "loss": 0.8635, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 46.09, | |
| "learning_rate": 5.5376e-05, | |
| "loss": 0.859, | |
| "step": 5300 | |
| }, | |
| { | |
| "epoch": 46.96, | |
| "learning_rate": 5.3476e-05, | |
| "loss": 0.8586, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 47.83, | |
| "learning_rate": 5.157600000000001e-05, | |
| "loss": 0.8449, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 47.83, | |
| "eval_cer": 0.05136537340260782, | |
| "eval_loss": 0.20504631102085114, | |
| "eval_runtime": 198.4879, | |
| "eval_samples_per_second": 29.231, | |
| "eval_steps_per_second": 0.458, | |
| "eval_wer": 0.25338619094813347, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 48.7, | |
| "learning_rate": 4.9676000000000003e-05, | |
| "loss": 0.8345, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 49.57, | |
| "learning_rate": 4.7776e-05, | |
| "loss": 0.8229, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 50.43, | |
| "learning_rate": 4.5876000000000006e-05, | |
| "loss": 0.8203, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 51.3, | |
| "learning_rate": 4.397600000000001e-05, | |
| "loss": 0.8084, | |
| "step": 5900 | |
| }, | |
| { | |
| "epoch": 52.17, | |
| "learning_rate": 4.207600000000001e-05, | |
| "loss": 0.8028, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 52.17, | |
| "eval_cer": 0.04915567076544842, | |
| "eval_loss": 0.2032497674226761, | |
| "eval_runtime": 199.0329, | |
| "eval_samples_per_second": 29.151, | |
| "eval_steps_per_second": 0.457, | |
| "eval_wer": 0.24562272877436406, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 53.04, | |
| "learning_rate": 4.0176e-05, | |
| "loss": 0.7975, | |
| "step": 6100 | |
| }, | |
| { | |
| "epoch": 53.91, | |
| "learning_rate": 3.8295000000000005e-05, | |
| "loss": 0.7942, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 54.78, | |
| "learning_rate": 3.6395e-05, | |
| "loss": 0.7867, | |
| "step": 6300 | |
| }, | |
| { | |
| "epoch": 55.65, | |
| "learning_rate": 3.4495e-05, | |
| "loss": 0.787, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 56.52, | |
| "learning_rate": 3.2595e-05, | |
| "loss": 0.7881, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 56.52, | |
| "eval_cer": 0.04694596812828902, | |
| "eval_loss": 0.18896546959877014, | |
| "eval_runtime": 199.9061, | |
| "eval_samples_per_second": 29.024, | |
| "eval_steps_per_second": 0.455, | |
| "eval_wer": 0.2380008494973807, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 57.39, | |
| "learning_rate": 3.0695e-05, | |
| "loss": 0.7608, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 58.26, | |
| "learning_rate": 2.8795000000000005e-05, | |
| "loss": 0.7542, | |
| "step": 6700 | |
| }, | |
| { | |
| "epoch": 59.13, | |
| "learning_rate": 2.6895e-05, | |
| "loss": 0.755, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 60.0, | |
| "learning_rate": 2.4995000000000004e-05, | |
| "loss": 0.7569, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 60.87, | |
| "learning_rate": 2.3095e-05, | |
| "loss": 0.7423, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 60.87, | |
| "eval_cer": 0.04418662517129838, | |
| "eval_loss": 0.18159770965576172, | |
| "eval_runtime": 198.6364, | |
| "eval_samples_per_second": 29.209, | |
| "eval_steps_per_second": 0.458, | |
| "eval_wer": 0.22452687715323988, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 61.74, | |
| "learning_rate": 2.1195000000000006e-05, | |
| "loss": 0.7409, | |
| "step": 7100 | |
| }, | |
| { | |
| "epoch": 62.61, | |
| "learning_rate": 1.9295e-05, | |
| "loss": 0.7208, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 63.48, | |
| "learning_rate": 1.7395e-05, | |
| "loss": 0.7488, | |
| "step": 7300 | |
| }, | |
| { | |
| "epoch": 64.35, | |
| "learning_rate": 1.5495000000000003e-05, | |
| "loss": 0.7134, | |
| "step": 7400 | |
| }, | |
| { | |
| "epoch": 65.22, | |
| "learning_rate": 1.3594999999999998e-05, | |
| "loss": 0.7248, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 65.22, | |
| "eval_cer": 0.042203463476772125, | |
| "eval_loss": 0.17892056703567505, | |
| "eval_runtime": 197.8292, | |
| "eval_samples_per_second": 29.328, | |
| "eval_steps_per_second": 0.46, | |
| "eval_wer": 0.21650384633536268, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 66.09, | |
| "learning_rate": 1.1695000000000002e-05, | |
| "loss": 0.7076, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 66.96, | |
| "learning_rate": 9.794999999999999e-06, | |
| "loss": 0.7132, | |
| "step": 7700 | |
| }, | |
| { | |
| "epoch": 67.83, | |
| "learning_rate": 7.895000000000003e-06, | |
| "loss": 0.6972, | |
| "step": 7800 | |
| }, | |
| { | |
| "epoch": 68.7, | |
| "learning_rate": 5.994999999999999e-06, | |
| "loss": 0.7019, | |
| "step": 7900 | |
| }, | |
| { | |
| "epoch": 69.57, | |
| "learning_rate": 4.095000000000005e-06, | |
| "loss": 0.6993, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 69.57, | |
| "eval_cer": 0.040792224817745956, | |
| "eval_loss": 0.1746993511915207, | |
| "eval_runtime": 198.0151, | |
| "eval_samples_per_second": 29.301, | |
| "eval_steps_per_second": 0.46, | |
| "eval_wer": 0.21072254471659824, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 69.57, | |
| "step": 8000, | |
| "total_flos": 3.5433976605211066e+20, | |
| "train_loss": 1.039798891067505, | |
| "train_runtime": 49145.562, | |
| "train_samples_per_second": 20.836, | |
| "train_steps_per_second": 0.163 | |
| } | |
| ], | |
| "max_steps": 8000, | |
| "num_train_epochs": 70, | |
| "total_flos": 3.5433976605211066e+20, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |