| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 50.0, | |
| "global_step": 9650, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 3.7125e-06, | |
| "loss": 19.5412, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "learning_rate": 7.4625e-06, | |
| "loss": 11.0922, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "learning_rate": 1.1212499999999998e-05, | |
| "loss": 7.1984, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.07, | |
| "learning_rate": 1.49625e-05, | |
| "loss": 5.9337, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.59, | |
| "learning_rate": 1.8712499999999997e-05, | |
| "loss": 4.9733, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.59, | |
| "eval_loss": 5.06973934173584, | |
| "eval_runtime": 115.8922, | |
| "eval_samples_per_second": 23.237, | |
| "eval_steps_per_second": 2.908, | |
| "eval_wer": 1.0, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 3.11, | |
| "learning_rate": 2.2462499999999997e-05, | |
| "loss": 4.2014, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 3.63, | |
| "learning_rate": 2.6212499999999997e-05, | |
| "loss": 3.704, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 4.15, | |
| "learning_rate": 2.99625e-05, | |
| "loss": 3.5005, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 4.66, | |
| "learning_rate": 3.37125e-05, | |
| "loss": 3.4424, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 5.18, | |
| "learning_rate": 3.7462499999999996e-05, | |
| "loss": 3.3839, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 5.18, | |
| "eval_loss": 3.3517656326293945, | |
| "eval_runtime": 115.48, | |
| "eval_samples_per_second": 23.32, | |
| "eval_steps_per_second": 2.918, | |
| "eval_wer": 1.0, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 5.7, | |
| "learning_rate": 4.12125e-05, | |
| "loss": 3.3036, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 6.22, | |
| "learning_rate": 4.4962499999999995e-05, | |
| "loss": 3.2565, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 6.74, | |
| "learning_rate": 4.871249999999999e-05, | |
| "loss": 2.9583, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 7.25, | |
| "learning_rate": 5.2462499999999994e-05, | |
| "loss": 2.3824, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 7.77, | |
| "learning_rate": 5.62125e-05, | |
| "loss": 2.0596, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 7.77, | |
| "eval_loss": 1.3991833925247192, | |
| "eval_runtime": 116.0149, | |
| "eval_samples_per_second": 23.213, | |
| "eval_steps_per_second": 2.905, | |
| "eval_wer": 0.7869276218611522, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 8.29, | |
| "learning_rate": 5.9962499999999994e-05, | |
| "loss": 1.9024, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 8.81, | |
| "learning_rate": 6.37125e-05, | |
| "loss": 1.7879, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 9.33, | |
| "learning_rate": 6.746249999999999e-05, | |
| "loss": 1.7072, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 9.84, | |
| "learning_rate": 7.121249999999999e-05, | |
| "loss": 1.6689, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 10.36, | |
| "learning_rate": 7.49625e-05, | |
| "loss": 1.6102, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 10.36, | |
| "eval_loss": 1.0711737871170044, | |
| "eval_runtime": 115.3994, | |
| "eval_samples_per_second": 23.336, | |
| "eval_steps_per_second": 2.92, | |
| "eval_wer": 0.6754062038404727, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 10.88, | |
| "learning_rate": 7.40392156862745e-05, | |
| "loss": 1.5881, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 11.4, | |
| "learning_rate": 7.305882352941176e-05, | |
| "loss": 1.5371, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 11.92, | |
| "learning_rate": 7.208823529411764e-05, | |
| "loss": 1.5062, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 12.44, | |
| "learning_rate": 7.11078431372549e-05, | |
| "loss": 1.4704, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 12.95, | |
| "learning_rate": 7.012745098039215e-05, | |
| "loss": 1.4587, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 12.95, | |
| "eval_loss": 0.9280298352241516, | |
| "eval_runtime": 115.6447, | |
| "eval_samples_per_second": 23.287, | |
| "eval_steps_per_second": 2.914, | |
| "eval_wer": 0.6361244460856721, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 13.47, | |
| "learning_rate": 6.91470588235294e-05, | |
| "loss": 1.4201, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 13.99, | |
| "learning_rate": 6.81764705882353e-05, | |
| "loss": 1.4221, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 14.51, | |
| "learning_rate": 6.719607843137255e-05, | |
| "loss": 1.3937, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 15.03, | |
| "learning_rate": 6.62156862745098e-05, | |
| "loss": 1.3984, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 15.54, | |
| "learning_rate": 6.523529411764705e-05, | |
| "loss": 1.3667, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 15.54, | |
| "eval_loss": 0.9280989766120911, | |
| "eval_runtime": 115.1284, | |
| "eval_samples_per_second": 23.391, | |
| "eval_steps_per_second": 2.927, | |
| "eval_wer": 0.6155372968980798, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 16.06, | |
| "learning_rate": 6.425490196078431e-05, | |
| "loss": 1.367, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 16.58, | |
| "learning_rate": 6.327450980392156e-05, | |
| "loss": 1.3535, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 17.1, | |
| "learning_rate": 6.229411764705881e-05, | |
| "loss": 1.3265, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 17.62, | |
| "learning_rate": 6.131372549019608e-05, | |
| "loss": 1.3202, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 18.13, | |
| "learning_rate": 6.033333333333333e-05, | |
| "loss": 1.3042, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 18.13, | |
| "eval_loss": 0.9037219882011414, | |
| "eval_runtime": 114.717, | |
| "eval_samples_per_second": 23.475, | |
| "eval_steps_per_second": 2.938, | |
| "eval_wer": 0.5921344165435746, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 18.65, | |
| "learning_rate": 5.9352941176470584e-05, | |
| "loss": 1.3016, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 19.17, | |
| "learning_rate": 5.837254901960784e-05, | |
| "loss": 1.2862, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 19.69, | |
| "learning_rate": 5.739215686274509e-05, | |
| "loss": 1.2773, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 20.21, | |
| "learning_rate": 5.641176470588235e-05, | |
| "loss": 1.2543, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 20.73, | |
| "learning_rate": 5.54313725490196e-05, | |
| "loss": 1.2544, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 20.73, | |
| "eval_loss": 0.8996412754058838, | |
| "eval_runtime": 115.5118, | |
| "eval_samples_per_second": 23.314, | |
| "eval_steps_per_second": 2.917, | |
| "eval_wer": 0.5824409158050221, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 21.24, | |
| "learning_rate": 5.445098039215686e-05, | |
| "loss": 1.2419, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 21.76, | |
| "learning_rate": 5.347058823529411e-05, | |
| "loss": 1.2347, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 22.28, | |
| "learning_rate": 5.2490196078431365e-05, | |
| "loss": 1.2373, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 22.8, | |
| "learning_rate": 5.150980392156863e-05, | |
| "loss": 1.2337, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 23.32, | |
| "learning_rate": 5.052941176470588e-05, | |
| "loss": 1.2274, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 23.32, | |
| "eval_loss": 0.8933804631233215, | |
| "eval_runtime": 115.4101, | |
| "eval_samples_per_second": 23.334, | |
| "eval_steps_per_second": 2.92, | |
| "eval_wer": 0.5797175036927622, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 23.83, | |
| "learning_rate": 4.9549019607843137e-05, | |
| "loss": 1.2091, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 24.35, | |
| "learning_rate": 4.856862745098039e-05, | |
| "loss": 1.1947, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 24.87, | |
| "learning_rate": 4.759803921568627e-05, | |
| "loss": 1.1901, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 25.39, | |
| "learning_rate": 4.6617647058823525e-05, | |
| "loss": 1.1731, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 25.91, | |
| "learning_rate": 4.5637254901960776e-05, | |
| "loss": 1.1763, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 25.91, | |
| "eval_loss": 0.8642701506614685, | |
| "eval_runtime": 115.2886, | |
| "eval_samples_per_second": 23.359, | |
| "eval_steps_per_second": 2.923, | |
| "eval_wer": 0.5759785819793205, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 26.42, | |
| "learning_rate": 4.465686274509803e-05, | |
| "loss": 1.1661, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 26.94, | |
| "learning_rate": 4.367647058823529e-05, | |
| "loss": 1.1628, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 27.46, | |
| "learning_rate": 4.269607843137254e-05, | |
| "loss": 1.1618, | |
| "step": 5300 | |
| }, | |
| { | |
| "epoch": 27.98, | |
| "learning_rate": 4.1715686274509805e-05, | |
| "loss": 1.1468, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 28.5, | |
| "learning_rate": 4.0735294117647055e-05, | |
| "loss": 1.149, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 28.5, | |
| "eval_loss": 0.8251490592956543, | |
| "eval_runtime": 116.3124, | |
| "eval_samples_per_second": 23.153, | |
| "eval_steps_per_second": 2.897, | |
| "eval_wer": 0.5543759231905465, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 29.02, | |
| "learning_rate": 3.975490196078431e-05, | |
| "loss": 1.1572, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 29.53, | |
| "learning_rate": 3.877450980392157e-05, | |
| "loss": 1.1389, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 30.05, | |
| "learning_rate": 3.779411764705882e-05, | |
| "loss": 1.1337, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 30.57, | |
| "learning_rate": 3.681372549019607e-05, | |
| "loss": 1.1226, | |
| "step": 5900 | |
| }, | |
| { | |
| "epoch": 31.09, | |
| "learning_rate": 3.5833333333333335e-05, | |
| "loss": 1.1207, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 31.09, | |
| "eval_loss": 0.8505932092666626, | |
| "eval_runtime": 115.2841, | |
| "eval_samples_per_second": 23.36, | |
| "eval_steps_per_second": 2.923, | |
| "eval_wer": 0.5527141802067946, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 31.61, | |
| "learning_rate": 3.4852941176470585e-05, | |
| "loss": 1.1243, | |
| "step": 6100 | |
| }, | |
| { | |
| "epoch": 32.12, | |
| "learning_rate": 3.387254901960784e-05, | |
| "loss": 1.1156, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 32.64, | |
| "learning_rate": 3.289215686274509e-05, | |
| "loss": 1.1158, | |
| "step": 6300 | |
| }, | |
| { | |
| "epoch": 33.16, | |
| "learning_rate": 3.191176470588235e-05, | |
| "loss": 1.0918, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 33.68, | |
| "learning_rate": 3.09313725490196e-05, | |
| "loss": 1.091, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 33.68, | |
| "eval_loss": 0.8370497822761536, | |
| "eval_runtime": 115.5394, | |
| "eval_samples_per_second": 23.308, | |
| "eval_steps_per_second": 2.917, | |
| "eval_wer": 0.5365583456425406, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 34.2, | |
| "learning_rate": 2.995098039215686e-05, | |
| "loss": 1.0892, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 34.72, | |
| "learning_rate": 2.8980392156862746e-05, | |
| "loss": 1.0799, | |
| "step": 6700 | |
| }, | |
| { | |
| "epoch": 35.23, | |
| "learning_rate": 2.8e-05, | |
| "loss": 1.0671, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 35.75, | |
| "learning_rate": 2.7019607843137253e-05, | |
| "loss": 1.07, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 36.27, | |
| "learning_rate": 2.6049019607843135e-05, | |
| "loss": 1.0613, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 36.27, | |
| "eval_loss": 0.8345041275024414, | |
| "eval_runtime": 116.5707, | |
| "eval_samples_per_second": 23.102, | |
| "eval_steps_per_second": 2.891, | |
| "eval_wer": 0.5351735598227474, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 36.79, | |
| "learning_rate": 2.506862745098039e-05, | |
| "loss": 1.0797, | |
| "step": 7100 | |
| }, | |
| { | |
| "epoch": 37.31, | |
| "learning_rate": 2.4088235294117646e-05, | |
| "loss": 1.0787, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 37.82, | |
| "learning_rate": 2.31078431372549e-05, | |
| "loss": 1.0706, | |
| "step": 7300 | |
| }, | |
| { | |
| "epoch": 38.34, | |
| "learning_rate": 2.2127450980392153e-05, | |
| "loss": 1.0585, | |
| "step": 7400 | |
| }, | |
| { | |
| "epoch": 38.86, | |
| "learning_rate": 2.114705882352941e-05, | |
| "loss": 1.0495, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 38.86, | |
| "eval_loss": 0.8380374908447266, | |
| "eval_runtime": 115.7848, | |
| "eval_samples_per_second": 23.259, | |
| "eval_steps_per_second": 2.911, | |
| "eval_wer": 0.5321270310192023, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 39.38, | |
| "learning_rate": 2.0166666666666664e-05, | |
| "loss": 1.0456, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 39.9, | |
| "learning_rate": 1.9186274509803922e-05, | |
| "loss": 1.0446, | |
| "step": 7700 | |
| }, | |
| { | |
| "epoch": 40.41, | |
| "learning_rate": 1.8205882352941176e-05, | |
| "loss": 1.0353, | |
| "step": 7800 | |
| }, | |
| { | |
| "epoch": 40.93, | |
| "learning_rate": 1.722549019607843e-05, | |
| "loss": 1.0375, | |
| "step": 7900 | |
| }, | |
| { | |
| "epoch": 41.45, | |
| "learning_rate": 1.6245098039215687e-05, | |
| "loss": 1.0345, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 41.45, | |
| "eval_loss": 0.828546941280365, | |
| "eval_runtime": 115.0908, | |
| "eval_samples_per_second": 23.399, | |
| "eval_steps_per_second": 2.928, | |
| "eval_wer": 0.5269110044313147, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 41.97, | |
| "learning_rate": 1.526470588235294e-05, | |
| "loss": 1.0196, | |
| "step": 8100 | |
| }, | |
| { | |
| "epoch": 42.49, | |
| "learning_rate": 1.4284313725490196e-05, | |
| "loss": 1.0265, | |
| "step": 8200 | |
| }, | |
| { | |
| "epoch": 43.01, | |
| "learning_rate": 1.330392156862745e-05, | |
| "loss": 1.028, | |
| "step": 8300 | |
| }, | |
| { | |
| "epoch": 43.52, | |
| "learning_rate": 1.2323529411764704e-05, | |
| "loss": 1.0281, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 44.04, | |
| "learning_rate": 1.1343137254901961e-05, | |
| "loss": 1.0297, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 44.04, | |
| "eval_loss": 0.7836087346076965, | |
| "eval_runtime": 116.6272, | |
| "eval_samples_per_second": 23.091, | |
| "eval_steps_per_second": 2.89, | |
| "eval_wer": 0.5141248153618907, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 44.56, | |
| "learning_rate": 1.0362745098039215e-05, | |
| "loss": 1.0097, | |
| "step": 8600 | |
| }, | |
| { | |
| "epoch": 45.08, | |
| "learning_rate": 9.382352941176469e-06, | |
| "loss": 1.0046, | |
| "step": 8700 | |
| }, | |
| { | |
| "epoch": 45.6, | |
| "learning_rate": 8.401960784313724e-06, | |
| "loss": 1.0082, | |
| "step": 8800 | |
| }, | |
| { | |
| "epoch": 46.11, | |
| "learning_rate": 7.42156862745098e-06, | |
| "loss": 1.0065, | |
| "step": 8900 | |
| }, | |
| { | |
| "epoch": 46.63, | |
| "learning_rate": 6.4411764705882346e-06, | |
| "loss": 1.027, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 46.63, | |
| "eval_loss": 0.8119935989379883, | |
| "eval_runtime": 115.6592, | |
| "eval_samples_per_second": 23.284, | |
| "eval_steps_per_second": 2.914, | |
| "eval_wer": 0.5179560561299852, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 47.15, | |
| "learning_rate": 5.460784313725489e-06, | |
| "loss": 0.9917, | |
| "step": 9100 | |
| }, | |
| { | |
| "epoch": 47.67, | |
| "learning_rate": 4.480392156862745e-06, | |
| "loss": 0.9925, | |
| "step": 9200 | |
| }, | |
| { | |
| "epoch": 48.19, | |
| "learning_rate": 3.5e-06, | |
| "loss": 0.9924, | |
| "step": 9300 | |
| }, | |
| { | |
| "epoch": 48.7, | |
| "learning_rate": 2.5196078431372547e-06, | |
| "loss": 0.983, | |
| "step": 9400 | |
| }, | |
| { | |
| "epoch": 49.22, | |
| "learning_rate": 1.5392156862745098e-06, | |
| "loss": 0.9876, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 49.22, | |
| "eval_loss": 0.8109092116355896, | |
| "eval_runtime": 115.9344, | |
| "eval_samples_per_second": 23.229, | |
| "eval_steps_per_second": 2.907, | |
| "eval_wer": 0.5187869276218612, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 49.74, | |
| "learning_rate": 5.588235294117647e-07, | |
| "loss": 0.973, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 50.0, | |
| "step": 9650, | |
| "total_flos": 4.0125393308879946e+19, | |
| "train_loss": 1.853262206161578, | |
| "train_runtime": 19140.1846, | |
| "train_samples_per_second": 16.134, | |
| "train_steps_per_second": 0.504 | |
| } | |
| ], | |
| "max_steps": 9650, | |
| "num_train_epochs": 50, | |
| "total_flos": 4.0125393308879946e+19, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |