|
{ |
|
"best_metric": 3.3522434244455903, |
|
"best_model_checkpoint": "./whisper-small/checkpoint-2000", |
|
"epoch": 6.472491909385114, |
|
"eval_steps": 10, |
|
"global_step": 2000, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 0.08, |
|
"grad_norm": 81.5032730102539, |
|
"learning_rate": 4.2000000000000006e-07, |
|
"loss": 2.9603, |
|
"step": 25 |
|
}, |
|
{ |
|
"epoch": 0.16, |
|
"grad_norm": 17.11170768737793, |
|
"learning_rate": 9.200000000000001e-07, |
|
"loss": 2.144, |
|
"step": 50 |
|
}, |
|
{ |
|
"epoch": 0.24, |
|
"grad_norm": 11.967795372009277, |
|
"learning_rate": 1.42e-06, |
|
"loss": 1.3446, |
|
"step": 75 |
|
}, |
|
{ |
|
"epoch": 0.32, |
|
"grad_norm": 8.235530853271484, |
|
"learning_rate": 1.9200000000000003e-06, |
|
"loss": 0.7532, |
|
"step": 100 |
|
}, |
|
{ |
|
"epoch": 0.4, |
|
"grad_norm": 6.933337688446045, |
|
"learning_rate": 2.42e-06, |
|
"loss": 0.5681, |
|
"step": 125 |
|
}, |
|
{ |
|
"epoch": 0.49, |
|
"grad_norm": 8.942669868469238, |
|
"learning_rate": 2.92e-06, |
|
"loss": 0.496, |
|
"step": 150 |
|
}, |
|
{ |
|
"epoch": 0.57, |
|
"grad_norm": 11.463906288146973, |
|
"learning_rate": 3.4200000000000007e-06, |
|
"loss": 0.4111, |
|
"step": 175 |
|
}, |
|
{ |
|
"epoch": 0.65, |
|
"grad_norm": 7.532162666320801, |
|
"learning_rate": 3.920000000000001e-06, |
|
"loss": 0.3332, |
|
"step": 200 |
|
}, |
|
{ |
|
"epoch": 0.73, |
|
"grad_norm": 7.5595197677612305, |
|
"learning_rate": 4.42e-06, |
|
"loss": 0.2728, |
|
"step": 225 |
|
}, |
|
{ |
|
"epoch": 0.81, |
|
"grad_norm": 4.813773155212402, |
|
"learning_rate": 4.92e-06, |
|
"loss": 0.1087, |
|
"step": 250 |
|
}, |
|
{ |
|
"epoch": 0.89, |
|
"grad_norm": 3.7373056411743164, |
|
"learning_rate": 5.420000000000001e-06, |
|
"loss": 0.0653, |
|
"step": 275 |
|
}, |
|
{ |
|
"epoch": 0.97, |
|
"grad_norm": 4.375819683074951, |
|
"learning_rate": 5.92e-06, |
|
"loss": 0.059, |
|
"step": 300 |
|
}, |
|
{ |
|
"epoch": 1.05, |
|
"grad_norm": 1.661054015159607, |
|
"learning_rate": 6.42e-06, |
|
"loss": 0.0381, |
|
"step": 325 |
|
}, |
|
{ |
|
"epoch": 1.13, |
|
"grad_norm": 2.9067795276641846, |
|
"learning_rate": 6.92e-06, |
|
"loss": 0.0327, |
|
"step": 350 |
|
}, |
|
{ |
|
"epoch": 1.21, |
|
"grad_norm": 4.853701114654541, |
|
"learning_rate": 7.420000000000001e-06, |
|
"loss": 0.0295, |
|
"step": 375 |
|
}, |
|
{ |
|
"epoch": 1.29, |
|
"grad_norm": 1.736449122428894, |
|
"learning_rate": 7.92e-06, |
|
"loss": 0.0311, |
|
"step": 400 |
|
}, |
|
{ |
|
"epoch": 1.38, |
|
"grad_norm": 3.658984899520874, |
|
"learning_rate": 8.42e-06, |
|
"loss": 0.0296, |
|
"step": 425 |
|
}, |
|
{ |
|
"epoch": 1.46, |
|
"grad_norm": 2.040172576904297, |
|
"learning_rate": 8.920000000000001e-06, |
|
"loss": 0.0221, |
|
"step": 450 |
|
}, |
|
{ |
|
"epoch": 1.54, |
|
"grad_norm": 2.4981894493103027, |
|
"learning_rate": 9.42e-06, |
|
"loss": 0.0251, |
|
"step": 475 |
|
}, |
|
{ |
|
"epoch": 1.62, |
|
"grad_norm": 2.639390230178833, |
|
"learning_rate": 9.920000000000002e-06, |
|
"loss": 0.0261, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 1.7, |
|
"grad_norm": 2.526102066040039, |
|
"learning_rate": 9.940000000000001e-06, |
|
"loss": 0.0253, |
|
"step": 525 |
|
}, |
|
{ |
|
"epoch": 1.78, |
|
"grad_norm": 1.3722740411758423, |
|
"learning_rate": 9.86857142857143e-06, |
|
"loss": 0.0203, |
|
"step": 550 |
|
}, |
|
{ |
|
"epoch": 1.86, |
|
"grad_norm": 2.3701043128967285, |
|
"learning_rate": 9.797142857142858e-06, |
|
"loss": 0.0225, |
|
"step": 575 |
|
}, |
|
{ |
|
"epoch": 1.94, |
|
"grad_norm": 2.8724324703216553, |
|
"learning_rate": 9.725714285714287e-06, |
|
"loss": 0.0201, |
|
"step": 600 |
|
}, |
|
{ |
|
"epoch": 2.02, |
|
"grad_norm": 1.0449714660644531, |
|
"learning_rate": 9.654285714285716e-06, |
|
"loss": 0.0149, |
|
"step": 625 |
|
}, |
|
{ |
|
"epoch": 2.1, |
|
"grad_norm": 5.038928031921387, |
|
"learning_rate": 9.582857142857143e-06, |
|
"loss": 0.0104, |
|
"step": 650 |
|
}, |
|
{ |
|
"epoch": 2.18, |
|
"grad_norm": 1.6107451915740967, |
|
"learning_rate": 9.511428571428572e-06, |
|
"loss": 0.0086, |
|
"step": 675 |
|
}, |
|
{ |
|
"epoch": 2.27, |
|
"grad_norm": 1.3224350214004517, |
|
"learning_rate": 9.440000000000001e-06, |
|
"loss": 0.006, |
|
"step": 700 |
|
}, |
|
{ |
|
"epoch": 2.35, |
|
"grad_norm": 0.14248408377170563, |
|
"learning_rate": 9.368571428571428e-06, |
|
"loss": 0.0059, |
|
"step": 725 |
|
}, |
|
{ |
|
"epoch": 2.43, |
|
"grad_norm": 2.3801496028900146, |
|
"learning_rate": 9.297142857142857e-06, |
|
"loss": 0.0056, |
|
"step": 750 |
|
}, |
|
{ |
|
"epoch": 2.51, |
|
"grad_norm": 1.156813621520996, |
|
"learning_rate": 9.225714285714286e-06, |
|
"loss": 0.0067, |
|
"step": 775 |
|
}, |
|
{ |
|
"epoch": 2.59, |
|
"grad_norm": 0.9941128492355347, |
|
"learning_rate": 9.154285714285715e-06, |
|
"loss": 0.0076, |
|
"step": 800 |
|
}, |
|
{ |
|
"epoch": 2.67, |
|
"grad_norm": 1.4023518562316895, |
|
"learning_rate": 9.082857142857143e-06, |
|
"loss": 0.006, |
|
"step": 825 |
|
}, |
|
{ |
|
"epoch": 2.75, |
|
"grad_norm": 1.1814390420913696, |
|
"learning_rate": 9.011428571428572e-06, |
|
"loss": 0.0053, |
|
"step": 850 |
|
}, |
|
{ |
|
"epoch": 2.83, |
|
"grad_norm": 1.8819189071655273, |
|
"learning_rate": 8.94e-06, |
|
"loss": 0.008, |
|
"step": 875 |
|
}, |
|
{ |
|
"epoch": 2.91, |
|
"grad_norm": 0.9754526019096375, |
|
"learning_rate": 8.86857142857143e-06, |
|
"loss": 0.0047, |
|
"step": 900 |
|
}, |
|
{ |
|
"epoch": 2.99, |
|
"grad_norm": 0.40376096963882446, |
|
"learning_rate": 8.797142857142857e-06, |
|
"loss": 0.0062, |
|
"step": 925 |
|
}, |
|
{ |
|
"epoch": 3.07, |
|
"grad_norm": 0.07103849202394485, |
|
"learning_rate": 8.725714285714286e-06, |
|
"loss": 0.0028, |
|
"step": 950 |
|
}, |
|
{ |
|
"epoch": 3.16, |
|
"grad_norm": 0.5621446967124939, |
|
"learning_rate": 8.654285714285715e-06, |
|
"loss": 0.0025, |
|
"step": 975 |
|
}, |
|
{ |
|
"epoch": 3.24, |
|
"grad_norm": 0.506003201007843, |
|
"learning_rate": 8.582857142857144e-06, |
|
"loss": 0.0026, |
|
"step": 1000 |
|
}, |
|
{ |
|
"epoch": 3.24, |
|
"eval_loss": 0.020510252565145493, |
|
"eval_runtime": 791.6891, |
|
"eval_samples_per_second": 1.557, |
|
"eval_steps_per_second": 0.196, |
|
"eval_wer": 4.486848891181021, |
|
"step": 1000 |
|
}, |
|
{ |
|
"epoch": 3.32, |
|
"grad_norm": 0.723780632019043, |
|
"learning_rate": 8.511428571428571e-06, |
|
"loss": 0.003, |
|
"step": 1025 |
|
}, |
|
{ |
|
"epoch": 3.4, |
|
"grad_norm": 0.02721921168267727, |
|
"learning_rate": 8.44e-06, |
|
"loss": 0.0028, |
|
"step": 1050 |
|
}, |
|
{ |
|
"epoch": 3.48, |
|
"grad_norm": 1.0563775300979614, |
|
"learning_rate": 8.36857142857143e-06, |
|
"loss": 0.002, |
|
"step": 1075 |
|
}, |
|
{ |
|
"epoch": 3.56, |
|
"grad_norm": 1.54115629196167, |
|
"learning_rate": 8.297142857142859e-06, |
|
"loss": 0.0027, |
|
"step": 1100 |
|
}, |
|
{ |
|
"epoch": 3.64, |
|
"grad_norm": 0.017975807189941406, |
|
"learning_rate": 8.225714285714288e-06, |
|
"loss": 0.0016, |
|
"step": 1125 |
|
}, |
|
{ |
|
"epoch": 3.72, |
|
"grad_norm": 0.15264229476451874, |
|
"learning_rate": 8.154285714285715e-06, |
|
"loss": 0.0034, |
|
"step": 1150 |
|
}, |
|
{ |
|
"epoch": 3.8, |
|
"grad_norm": 0.655139148235321, |
|
"learning_rate": 8.082857142857144e-06, |
|
"loss": 0.0025, |
|
"step": 1175 |
|
}, |
|
{ |
|
"epoch": 3.88, |
|
"grad_norm": 0.9618725776672363, |
|
"learning_rate": 8.011428571428573e-06, |
|
"loss": 0.0017, |
|
"step": 1200 |
|
}, |
|
{ |
|
"epoch": 3.96, |
|
"grad_norm": 3.206550121307373, |
|
"learning_rate": 7.94e-06, |
|
"loss": 0.0031, |
|
"step": 1225 |
|
}, |
|
{ |
|
"epoch": 4.05, |
|
"grad_norm": 0.41130420565605164, |
|
"learning_rate": 7.86857142857143e-06, |
|
"loss": 0.0012, |
|
"step": 1250 |
|
}, |
|
{ |
|
"epoch": 4.13, |
|
"grad_norm": 0.0929947942495346, |
|
"learning_rate": 7.797142857142858e-06, |
|
"loss": 0.0012, |
|
"step": 1275 |
|
}, |
|
{ |
|
"epoch": 4.21, |
|
"grad_norm": 0.050773363560438156, |
|
"learning_rate": 7.725714285714286e-06, |
|
"loss": 0.0005, |
|
"step": 1300 |
|
}, |
|
{ |
|
"epoch": 4.29, |
|
"grad_norm": 0.0827890932559967, |
|
"learning_rate": 7.654285714285715e-06, |
|
"loss": 0.0005, |
|
"step": 1325 |
|
}, |
|
{ |
|
"epoch": 4.37, |
|
"grad_norm": 0.02210867777466774, |
|
"learning_rate": 7.5828571428571444e-06, |
|
"loss": 0.0007, |
|
"step": 1350 |
|
}, |
|
{ |
|
"epoch": 4.45, |
|
"grad_norm": 0.014925210736691952, |
|
"learning_rate": 7.511428571428572e-06, |
|
"loss": 0.0012, |
|
"step": 1375 |
|
}, |
|
{ |
|
"epoch": 4.53, |
|
"grad_norm": 0.027211977168917656, |
|
"learning_rate": 7.440000000000001e-06, |
|
"loss": 0.0009, |
|
"step": 1400 |
|
}, |
|
{ |
|
"epoch": 4.61, |
|
"grad_norm": 1.4061143398284912, |
|
"learning_rate": 7.36857142857143e-06, |
|
"loss": 0.0008, |
|
"step": 1425 |
|
}, |
|
{ |
|
"epoch": 4.69, |
|
"grad_norm": 0.1997688263654709, |
|
"learning_rate": 7.297142857142858e-06, |
|
"loss": 0.0013, |
|
"step": 1450 |
|
}, |
|
{ |
|
"epoch": 4.77, |
|
"grad_norm": 0.011363201774656773, |
|
"learning_rate": 7.225714285714286e-06, |
|
"loss": 0.0005, |
|
"step": 1475 |
|
}, |
|
{ |
|
"epoch": 4.85, |
|
"grad_norm": 0.020546328276395798, |
|
"learning_rate": 7.154285714285715e-06, |
|
"loss": 0.0004, |
|
"step": 1500 |
|
}, |
|
{ |
|
"epoch": 4.94, |
|
"grad_norm": 0.026502510532736778, |
|
"learning_rate": 7.082857142857143e-06, |
|
"loss": 0.0007, |
|
"step": 1525 |
|
}, |
|
{ |
|
"epoch": 5.02, |
|
"grad_norm": 0.1575097292661667, |
|
"learning_rate": 7.011428571428572e-06, |
|
"loss": 0.0009, |
|
"step": 1550 |
|
}, |
|
{ |
|
"epoch": 5.1, |
|
"grad_norm": 0.89250248670578, |
|
"learning_rate": 6.9400000000000005e-06, |
|
"loss": 0.001, |
|
"step": 1575 |
|
}, |
|
{ |
|
"epoch": 5.18, |
|
"grad_norm": 0.5113543272018433, |
|
"learning_rate": 6.868571428571429e-06, |
|
"loss": 0.0003, |
|
"step": 1600 |
|
}, |
|
{ |
|
"epoch": 5.26, |
|
"grad_norm": 0.05453397333621979, |
|
"learning_rate": 6.797142857142858e-06, |
|
"loss": 0.0003, |
|
"step": 1625 |
|
}, |
|
{ |
|
"epoch": 5.34, |
|
"grad_norm": 0.01590396836400032, |
|
"learning_rate": 6.725714285714287e-06, |
|
"loss": 0.0002, |
|
"step": 1650 |
|
}, |
|
{ |
|
"epoch": 5.42, |
|
"grad_norm": 0.00851709395647049, |
|
"learning_rate": 6.654285714285716e-06, |
|
"loss": 0.0005, |
|
"step": 1675 |
|
}, |
|
{ |
|
"epoch": 5.5, |
|
"grad_norm": 0.014185091480612755, |
|
"learning_rate": 6.582857142857143e-06, |
|
"loss": 0.0004, |
|
"step": 1700 |
|
}, |
|
{ |
|
"epoch": 5.58, |
|
"grad_norm": 0.017834780737757683, |
|
"learning_rate": 6.511428571428572e-06, |
|
"loss": 0.0003, |
|
"step": 1725 |
|
}, |
|
{ |
|
"epoch": 5.66, |
|
"grad_norm": 0.017684083431959152, |
|
"learning_rate": 6.440000000000001e-06, |
|
"loss": 0.0003, |
|
"step": 1750 |
|
}, |
|
{ |
|
"epoch": 5.74, |
|
"grad_norm": 0.009825997054576874, |
|
"learning_rate": 6.368571428571429e-06, |
|
"loss": 0.0001, |
|
"step": 1775 |
|
}, |
|
{ |
|
"epoch": 5.83, |
|
"grad_norm": 0.08621440827846527, |
|
"learning_rate": 6.297142857142857e-06, |
|
"loss": 0.0004, |
|
"step": 1800 |
|
}, |
|
{ |
|
"epoch": 5.91, |
|
"grad_norm": 0.019367387518286705, |
|
"learning_rate": 6.225714285714286e-06, |
|
"loss": 0.0007, |
|
"step": 1825 |
|
}, |
|
{ |
|
"epoch": 5.99, |
|
"grad_norm": 0.04589274898171425, |
|
"learning_rate": 6.1542857142857145e-06, |
|
"loss": 0.0003, |
|
"step": 1850 |
|
}, |
|
{ |
|
"epoch": 6.07, |
|
"grad_norm": 0.010185680352151394, |
|
"learning_rate": 6.0828571428571435e-06, |
|
"loss": 0.0002, |
|
"step": 1875 |
|
}, |
|
{ |
|
"epoch": 6.15, |
|
"grad_norm": 0.004478767979890108, |
|
"learning_rate": 6.011428571428572e-06, |
|
"loss": 0.0002, |
|
"step": 1900 |
|
}, |
|
{ |
|
"epoch": 6.23, |
|
"grad_norm": 0.03858327120542526, |
|
"learning_rate": 5.94e-06, |
|
"loss": 0.0001, |
|
"step": 1925 |
|
}, |
|
{ |
|
"epoch": 6.31, |
|
"grad_norm": 0.008098349906504154, |
|
"learning_rate": 5.868571428571429e-06, |
|
"loss": 0.0002, |
|
"step": 1950 |
|
}, |
|
{ |
|
"epoch": 6.39, |
|
"grad_norm": 0.0035556757356971502, |
|
"learning_rate": 5.797142857142858e-06, |
|
"loss": 0.0001, |
|
"step": 1975 |
|
}, |
|
{ |
|
"epoch": 6.47, |
|
"grad_norm": 0.00917895045131445, |
|
"learning_rate": 5.725714285714287e-06, |
|
"loss": 0.0003, |
|
"step": 2000 |
|
}, |
|
{ |
|
"epoch": 6.47, |
|
"eval_loss": 0.01798020303249359, |
|
"eval_runtime": 807.9427, |
|
"eval_samples_per_second": 1.526, |
|
"eval_steps_per_second": 0.192, |
|
"eval_wer": 3.3522434244455903, |
|
"step": 2000 |
|
} |
|
], |
|
"logging_steps": 25, |
|
"max_steps": 4000, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 13, |
|
"save_steps": 10, |
|
"total_flos": 9.21395466141696e+18, |
|
"train_batch_size": 16, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|