| { | |
| "best_metric": 23.09402795425667, | |
| "best_model_checkpoint": "whisper4/checkpoint-130", | |
| "epoch": 8.333333333333334, | |
| "eval_steps": 10, | |
| "global_step": 300, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.1388888888888889, | |
| "grad_norm": 45.38771057128906, | |
| "learning_rate": 1.0000000000000002e-06, | |
| "loss": 3.9755, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.2777777777777778, | |
| "grad_norm": 42.970829010009766, | |
| "learning_rate": 2.0000000000000003e-06, | |
| "loss": 3.8231, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.2777777777777778, | |
| "eval_loss": 3.7088065147399902, | |
| "eval_runtime": 261.4715, | |
| "eval_samples_per_second": 1.912, | |
| "eval_steps_per_second": 0.241, | |
| "eval_wer": 76.93773824650572, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.4166666666666667, | |
| "grad_norm": 42.34159469604492, | |
| "learning_rate": 3e-06, | |
| "loss": 3.5982, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.5555555555555556, | |
| "grad_norm": 41.4112548828125, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 3.1925, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.5555555555555556, | |
| "eval_loss": 2.9438674449920654, | |
| "eval_runtime": 254.9124, | |
| "eval_samples_per_second": 1.961, | |
| "eval_steps_per_second": 0.247, | |
| "eval_wer": 65.56543837357052, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.6944444444444444, | |
| "grad_norm": 39.323768615722656, | |
| "learning_rate": 5e-06, | |
| "loss": 2.7453, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.8333333333333334, | |
| "grad_norm": 30.337932586669922, | |
| "learning_rate": 6e-06, | |
| "loss": 2.1383, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.8333333333333334, | |
| "eval_loss": 1.722086787223816, | |
| "eval_runtime": 256.1025, | |
| "eval_samples_per_second": 1.952, | |
| "eval_steps_per_second": 0.246, | |
| "eval_wer": 61.53113087674714, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.9722222222222222, | |
| "grad_norm": 18.886234283447266, | |
| "learning_rate": 7.000000000000001e-06, | |
| "loss": 1.4936, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 1.1111111111111112, | |
| "grad_norm": 10.311893463134766, | |
| "learning_rate": 8.000000000000001e-06, | |
| "loss": 1.0671, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 1.1111111111111112, | |
| "eval_loss": 0.8320145606994629, | |
| "eval_runtime": 254.169, | |
| "eval_samples_per_second": 1.967, | |
| "eval_steps_per_second": 0.248, | |
| "eval_wer": 50.69885641677255, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "grad_norm": 6.484711647033691, | |
| "learning_rate": 9e-06, | |
| "loss": 0.7616, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 1.3888888888888888, | |
| "grad_norm": 5.488741874694824, | |
| "learning_rate": 1e-05, | |
| "loss": 0.6947, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 1.3888888888888888, | |
| "eval_loss": 0.6587409377098083, | |
| "eval_runtime": 256.6295, | |
| "eval_samples_per_second": 1.948, | |
| "eval_steps_per_second": 0.245, | |
| "eval_wer": 41.01016518424396, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 1.5277777777777777, | |
| "grad_norm": 5.178290843963623, | |
| "learning_rate": 1.1000000000000001e-05, | |
| "loss": 0.6099, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 1.6666666666666665, | |
| "grad_norm": 5.149004936218262, | |
| "learning_rate": 1.2e-05, | |
| "loss": 0.6263, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.6666666666666665, | |
| "eval_loss": 0.5873834490776062, | |
| "eval_runtime": 254.3491, | |
| "eval_samples_per_second": 1.966, | |
| "eval_steps_per_second": 0.248, | |
| "eval_wer": 29.796696315120712, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.8055555555555556, | |
| "grad_norm": 5.192800521850586, | |
| "learning_rate": 1.3000000000000001e-05, | |
| "loss": 0.5654, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 1.9444444444444444, | |
| "grad_norm": 4.883419036865234, | |
| "learning_rate": 1.4000000000000001e-05, | |
| "loss": 0.5827, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.9444444444444444, | |
| "eval_loss": 0.5402054190635681, | |
| "eval_runtime": 256.4853, | |
| "eval_samples_per_second": 1.949, | |
| "eval_steps_per_second": 0.246, | |
| "eval_wer": 27.38246505717916, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 2.0833333333333335, | |
| "grad_norm": 4.878260612487793, | |
| "learning_rate": 1.5e-05, | |
| "loss": 0.5057, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 2.2222222222222223, | |
| "grad_norm": 4.442017555236816, | |
| "learning_rate": 1.6000000000000003e-05, | |
| "loss": 0.4222, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 2.2222222222222223, | |
| "eval_loss": 0.5154020190238953, | |
| "eval_runtime": 256.4153, | |
| "eval_samples_per_second": 1.95, | |
| "eval_steps_per_second": 0.246, | |
| "eval_wer": 32.05209656925032, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 2.361111111111111, | |
| "grad_norm": 4.793032646179199, | |
| "learning_rate": 1.7000000000000003e-05, | |
| "loss": 0.3807, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 4.618255138397217, | |
| "learning_rate": 1.8e-05, | |
| "loss": 0.4065, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "eval_loss": 0.49971044063568115, | |
| "eval_runtime": 254.4642, | |
| "eval_samples_per_second": 1.965, | |
| "eval_steps_per_second": 0.248, | |
| "eval_wer": 25.698856416772554, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 2.638888888888889, | |
| "grad_norm": 4.196300983428955, | |
| "learning_rate": 1.9e-05, | |
| "loss": 0.3807, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 2.7777777777777777, | |
| "grad_norm": 3.9986796379089355, | |
| "learning_rate": 2e-05, | |
| "loss": 0.3959, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 2.7777777777777777, | |
| "eval_loss": 0.48037058115005493, | |
| "eval_runtime": 254.5231, | |
| "eval_samples_per_second": 1.964, | |
| "eval_steps_per_second": 0.248, | |
| "eval_wer": 23.824650571791615, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 2.9166666666666665, | |
| "grad_norm": 4.1782402992248535, | |
| "learning_rate": 2.1e-05, | |
| "loss": 0.3847, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 3.0555555555555554, | |
| "grad_norm": 3.125694751739502, | |
| "learning_rate": 2.2000000000000003e-05, | |
| "loss": 0.3081, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 3.0555555555555554, | |
| "eval_loss": 0.46701571345329285, | |
| "eval_runtime": 254.757, | |
| "eval_samples_per_second": 1.963, | |
| "eval_steps_per_second": 0.247, | |
| "eval_wer": 24.841168996188056, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 3.1944444444444446, | |
| "grad_norm": 3.2956347465515137, | |
| "learning_rate": 2.3000000000000003e-05, | |
| "loss": 0.2701, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 3.3333333333333335, | |
| "grad_norm": 3.4233288764953613, | |
| "learning_rate": 2.4e-05, | |
| "loss": 0.2497, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 3.3333333333333335, | |
| "eval_loss": 0.4687294661998749, | |
| "eval_runtime": 255.0619, | |
| "eval_samples_per_second": 1.96, | |
| "eval_steps_per_second": 0.247, | |
| "eval_wer": 23.284625158831005, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 3.4722222222222223, | |
| "grad_norm": 3.383148193359375, | |
| "learning_rate": 2.5e-05, | |
| "loss": 0.268, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 3.611111111111111, | |
| "grad_norm": 3.4522266387939453, | |
| "learning_rate": 2.6000000000000002e-05, | |
| "loss": 0.2535, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 3.611111111111111, | |
| "eval_loss": 0.4594000279903412, | |
| "eval_runtime": 254.9771, | |
| "eval_samples_per_second": 1.961, | |
| "eval_steps_per_second": 0.247, | |
| "eval_wer": 23.09402795425667, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 3.75, | |
| "grad_norm": 3.3171420097351074, | |
| "learning_rate": 2.7000000000000002e-05, | |
| "loss": 0.2174, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 3.888888888888889, | |
| "grad_norm": 3.3734307289123535, | |
| "learning_rate": 2.8000000000000003e-05, | |
| "loss": 0.2428, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 3.888888888888889, | |
| "eval_loss": 0.45448967814445496, | |
| "eval_runtime": 252.6636, | |
| "eval_samples_per_second": 1.979, | |
| "eval_steps_per_second": 0.249, | |
| "eval_wer": 23.506988564167724, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 4.027777777777778, | |
| "grad_norm": 3.152697801589966, | |
| "learning_rate": 2.9e-05, | |
| "loss": 0.2136, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 4.166666666666667, | |
| "grad_norm": 3.5505619049072266, | |
| "learning_rate": 3e-05, | |
| "loss": 0.1627, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 4.166666666666667, | |
| "eval_loss": 0.46508893370628357, | |
| "eval_runtime": 253.2857, | |
| "eval_samples_per_second": 1.974, | |
| "eval_steps_per_second": 0.249, | |
| "eval_wer": 24.49174078780178, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 4.305555555555555, | |
| "grad_norm": 2.473661184310913, | |
| "learning_rate": 3.1e-05, | |
| "loss": 0.1302, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 4.444444444444445, | |
| "grad_norm": 2.767871379852295, | |
| "learning_rate": 3.2000000000000005e-05, | |
| "loss": 0.1224, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 4.444444444444445, | |
| "eval_loss": 0.46860620379447937, | |
| "eval_runtime": 252.3676, | |
| "eval_samples_per_second": 1.981, | |
| "eval_steps_per_second": 0.25, | |
| "eval_wer": 23.69758576874206, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 4.583333333333333, | |
| "grad_norm": 2.4746224880218506, | |
| "learning_rate": 3.3e-05, | |
| "loss": 0.1386, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 4.722222222222222, | |
| "grad_norm": 2.802751302719116, | |
| "learning_rate": 3.4000000000000007e-05, | |
| "loss": 0.1326, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 4.722222222222222, | |
| "eval_loss": 0.46528080105781555, | |
| "eval_runtime": 252.901, | |
| "eval_samples_per_second": 1.977, | |
| "eval_steps_per_second": 0.249, | |
| "eval_wer": 23.69758576874206, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 4.861111111111111, | |
| "grad_norm": 3.2861883640289307, | |
| "learning_rate": 3.5e-05, | |
| "loss": 0.1416, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 7.430758953094482, | |
| "learning_rate": 3.6e-05, | |
| "loss": 0.1334, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_loss": 0.474071204662323, | |
| "eval_runtime": 253.2666, | |
| "eval_samples_per_second": 1.974, | |
| "eval_steps_per_second": 0.249, | |
| "eval_wer": 24.74587039390089, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 5.138888888888889, | |
| "grad_norm": 2.6109585762023926, | |
| "learning_rate": 3.7e-05, | |
| "loss": 0.0807, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 5.277777777777778, | |
| "grad_norm": 1.921268343925476, | |
| "learning_rate": 3.8e-05, | |
| "loss": 0.0659, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 5.277777777777778, | |
| "eval_loss": 0.4791569113731384, | |
| "eval_runtime": 252.5034, | |
| "eval_samples_per_second": 1.98, | |
| "eval_steps_per_second": 0.25, | |
| "eval_wer": 24.68233799237611, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 5.416666666666667, | |
| "grad_norm": 1.5123528242111206, | |
| "learning_rate": 3.9000000000000006e-05, | |
| "loss": 0.0548, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 5.555555555555555, | |
| "grad_norm": 1.7394624948501587, | |
| "learning_rate": 4e-05, | |
| "loss": 0.0639, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 5.555555555555555, | |
| "eval_loss": 0.4760441780090332, | |
| "eval_runtime": 254.7055, | |
| "eval_samples_per_second": 1.963, | |
| "eval_steps_per_second": 0.247, | |
| "eval_wer": 33.38627700127065, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 5.694444444444445, | |
| "grad_norm": 2.006833553314209, | |
| "learning_rate": 4.1e-05, | |
| "loss": 0.0718, | |
| "step": 205 | |
| }, | |
| { | |
| "epoch": 5.833333333333333, | |
| "grad_norm": 1.9101831912994385, | |
| "learning_rate": 4.2e-05, | |
| "loss": 0.0667, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 5.833333333333333, | |
| "eval_loss": 0.48197290301322937, | |
| "eval_runtime": 252.8934, | |
| "eval_samples_per_second": 1.977, | |
| "eval_steps_per_second": 0.249, | |
| "eval_wer": 25.47649301143583, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 5.972222222222222, | |
| "grad_norm": 2.15120267868042, | |
| "learning_rate": 4.3e-05, | |
| "loss": 0.0833, | |
| "step": 215 | |
| }, | |
| { | |
| "epoch": 6.111111111111111, | |
| "grad_norm": 1.230398178100586, | |
| "learning_rate": 4.4000000000000006e-05, | |
| "loss": 0.042, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 6.111111111111111, | |
| "eval_loss": 0.4932812750339508, | |
| "eval_runtime": 254.0189, | |
| "eval_samples_per_second": 1.968, | |
| "eval_steps_per_second": 0.248, | |
| "eval_wer": 29.415501905972047, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 6.25, | |
| "grad_norm": 1.1129037141799927, | |
| "learning_rate": 4.5e-05, | |
| "loss": 0.0285, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 6.388888888888889, | |
| "grad_norm": 1.4872666597366333, | |
| "learning_rate": 4.600000000000001e-05, | |
| "loss": 0.0325, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 6.388888888888889, | |
| "eval_loss": 0.5065749883651733, | |
| "eval_runtime": 253.3018, | |
| "eval_samples_per_second": 1.974, | |
| "eval_steps_per_second": 0.249, | |
| "eval_wer": 29.987293519695047, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 6.527777777777778, | |
| "grad_norm": 1.527269959449768, | |
| "learning_rate": 4.7e-05, | |
| "loss": 0.0322, | |
| "step": 235 | |
| }, | |
| { | |
| "epoch": 6.666666666666667, | |
| "grad_norm": 1.6628929376602173, | |
| "learning_rate": 4.8e-05, | |
| "loss": 0.0333, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 6.666666666666667, | |
| "eval_loss": 0.5125746130943298, | |
| "eval_runtime": 252.7719, | |
| "eval_samples_per_second": 1.978, | |
| "eval_steps_per_second": 0.249, | |
| "eval_wer": 26.08005082592122, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 6.805555555555555, | |
| "grad_norm": 2.2532899379730225, | |
| "learning_rate": 4.9e-05, | |
| "loss": 0.0314, | |
| "step": 245 | |
| }, | |
| { | |
| "epoch": 6.944444444444445, | |
| "grad_norm": 3.415356397628784, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0333, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 6.944444444444445, | |
| "eval_loss": 0.5072929859161377, | |
| "eval_runtime": 252.6089, | |
| "eval_samples_per_second": 1.979, | |
| "eval_steps_per_second": 0.249, | |
| "eval_wer": 24.618805590851334, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 7.083333333333333, | |
| "grad_norm": 0.6166062951087952, | |
| "learning_rate": 5.1000000000000006e-05, | |
| "loss": 0.0251, | |
| "step": 255 | |
| }, | |
| { | |
| "epoch": 7.222222222222222, | |
| "grad_norm": 1.0601465702056885, | |
| "learning_rate": 5.2000000000000004e-05, | |
| "loss": 0.0187, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 7.222222222222222, | |
| "eval_loss": 0.5128570795059204, | |
| "eval_runtime": 253.8562, | |
| "eval_samples_per_second": 1.97, | |
| "eval_steps_per_second": 0.248, | |
| "eval_wer": 27.350698856416773, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 7.361111111111111, | |
| "grad_norm": 1.31247878074646, | |
| "learning_rate": 5.300000000000001e-05, | |
| "loss": 0.0219, | |
| "step": 265 | |
| }, | |
| { | |
| "epoch": 7.5, | |
| "grad_norm": 0.7731389999389648, | |
| "learning_rate": 5.4000000000000005e-05, | |
| "loss": 0.0214, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 7.5, | |
| "eval_loss": 0.5208793878555298, | |
| "eval_runtime": 254.7138, | |
| "eval_samples_per_second": 1.963, | |
| "eval_steps_per_second": 0.247, | |
| "eval_wer": 28.208386277001267, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 7.638888888888889, | |
| "grad_norm": 2.0568604469299316, | |
| "learning_rate": 5.500000000000001e-05, | |
| "loss": 0.0248, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 7.777777777777778, | |
| "grad_norm": 1.7497611045837402, | |
| "learning_rate": 5.6000000000000006e-05, | |
| "loss": 0.0187, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 7.777777777777778, | |
| "eval_loss": 0.5212948322296143, | |
| "eval_runtime": 254.6913, | |
| "eval_samples_per_second": 1.963, | |
| "eval_steps_per_second": 0.247, | |
| "eval_wer": 29.320203303684877, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 7.916666666666667, | |
| "grad_norm": 2.408604383468628, | |
| "learning_rate": 5.6999999999999996e-05, | |
| "loss": 0.0303, | |
| "step": 285 | |
| }, | |
| { | |
| "epoch": 8.055555555555555, | |
| "grad_norm": 1.3855714797973633, | |
| "learning_rate": 5.8e-05, | |
| "loss": 0.0312, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 8.055555555555555, | |
| "eval_loss": 0.5274094939231873, | |
| "eval_runtime": 253.5749, | |
| "eval_samples_per_second": 1.972, | |
| "eval_steps_per_second": 0.248, | |
| "eval_wer": 34.6569250317662, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 8.194444444444445, | |
| "grad_norm": 1.87900972366333, | |
| "learning_rate": 5.9e-05, | |
| "loss": 0.0153, | |
| "step": 295 | |
| }, | |
| { | |
| "epoch": 8.333333333333334, | |
| "grad_norm": 1.318336009979248, | |
| "learning_rate": 6e-05, | |
| "loss": 0.0172, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 8.333333333333334, | |
| "eval_loss": 0.5408744812011719, | |
| "eval_runtime": 252.2148, | |
| "eval_samples_per_second": 1.982, | |
| "eval_steps_per_second": 0.25, | |
| "eval_wer": 28.27191867852605, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 8.333333333333334, | |
| "step": 300, | |
| "total_flos": 9.2409447186432e+17, | |
| "train_loss": 0.5446730978041887, | |
| "train_runtime": 8175.9362, | |
| "train_samples_per_second": 4.697, | |
| "train_steps_per_second": 0.037 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 300, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 9, | |
| "save_steps": 10, | |
| "total_flos": 9.2409447186432e+17, | |
| "train_batch_size": 128, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |