| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 50.0, | |
| "global_step": 11100, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.45, | |
| "learning_rate": 3.6e-06, | |
| "loss": 38.551, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "learning_rate": 7.35e-06, | |
| "loss": 24.0076, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.35, | |
| "learning_rate": 1.1099999999999999e-05, | |
| "loss": 15.8812, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "learning_rate": 1.485e-05, | |
| "loss": 12.8964, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.25, | |
| "learning_rate": 1.8599999999999998e-05, | |
| "loss": 9.9344, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.7, | |
| "learning_rate": 2.2349999999999998e-05, | |
| "loss": 7.1613, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 3.15, | |
| "learning_rate": 2.6099999999999997e-05, | |
| "loss": 5.3772, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 3.6, | |
| "learning_rate": 2.985e-05, | |
| "loss": 4.6404, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 4.05, | |
| "learning_rate": 3.36e-05, | |
| "loss": 4.3831, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 4.5, | |
| "learning_rate": 3.735e-05, | |
| "loss": 4.221, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.5, | |
| "eval_loss": 4.1194539070129395, | |
| "eval_runtime": 239.8438, | |
| "eval_samples_per_second": 18.691, | |
| "eval_steps_per_second": 2.339, | |
| "eval_wer": 2.4023576512455516, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.95, | |
| "learning_rate": 4.11e-05, | |
| "loss": 4.1513, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 5.41, | |
| "learning_rate": 4.484999999999999e-05, | |
| "loss": 4.1055, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 5.86, | |
| "learning_rate": 4.8599999999999995e-05, | |
| "loss": 4.0677, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 6.31, | |
| "learning_rate": 5.234999999999999e-05, | |
| "loss": 4.0164, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 6.76, | |
| "learning_rate": 5.6099999999999995e-05, | |
| "loss": 3.9636, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 7.21, | |
| "learning_rate": 5.985e-05, | |
| "loss": 3.7377, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 7.66, | |
| "learning_rate": 6.359999999999999e-05, | |
| "loss": 3.1744, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 8.11, | |
| "learning_rate": 6.735e-05, | |
| "loss": 2.704, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 8.56, | |
| "learning_rate": 7.11e-05, | |
| "loss": 2.497, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 9.01, | |
| "learning_rate": 7.484999999999999e-05, | |
| "loss": 2.3597, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 9.01, | |
| "eval_loss": 1.102413535118103, | |
| "eval_runtime": 237.4335, | |
| "eval_samples_per_second": 18.881, | |
| "eval_steps_per_second": 2.363, | |
| "eval_wer": 2.761788256227758, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 9.46, | |
| "learning_rate": 7.42087912087912e-05, | |
| "loss": 2.2495, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 9.91, | |
| "learning_rate": 7.338461538461538e-05, | |
| "loss": 2.1767, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 10.36, | |
| "learning_rate": 7.256043956043956e-05, | |
| "loss": 2.1115, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 10.81, | |
| "learning_rate": 7.173626373626374e-05, | |
| "loss": 2.0653, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 11.26, | |
| "learning_rate": 7.09120879120879e-05, | |
| "loss": 2.0077, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 11.71, | |
| "learning_rate": 7.008791208791208e-05, | |
| "loss": 1.9933, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 12.16, | |
| "learning_rate": 6.926373626373626e-05, | |
| "loss": 1.9756, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 12.61, | |
| "learning_rate": 6.843956043956044e-05, | |
| "loss": 1.9246, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 13.06, | |
| "learning_rate": 6.761538461538461e-05, | |
| "loss": 1.9044, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 13.51, | |
| "learning_rate": 6.679120879120879e-05, | |
| "loss": 1.8795, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 13.51, | |
| "eval_loss": 0.749809205532074, | |
| "eval_runtime": 236.7828, | |
| "eval_samples_per_second": 18.933, | |
| "eval_steps_per_second": 2.369, | |
| "eval_wer": 2.588523131672598, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 13.96, | |
| "learning_rate": 6.596703296703297e-05, | |
| "loss": 1.8683, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 14.41, | |
| "learning_rate": 6.514285714285713e-05, | |
| "loss": 1.8367, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 14.86, | |
| "learning_rate": 6.431868131868131e-05, | |
| "loss": 1.8273, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 15.32, | |
| "learning_rate": 6.349450549450548e-05, | |
| "loss": 1.7743, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 15.77, | |
| "learning_rate": 6.267032967032966e-05, | |
| "loss": 1.7832, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 16.22, | |
| "learning_rate": 6.184615384615384e-05, | |
| "loss": 1.7915, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 16.67, | |
| "learning_rate": 6.102197802197801e-05, | |
| "loss": 1.7488, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 17.12, | |
| "learning_rate": 6.019780219780219e-05, | |
| "loss": 1.7178, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 17.57, | |
| "learning_rate": 5.937362637362637e-05, | |
| "loss": 1.6903, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 18.02, | |
| "learning_rate": 5.854945054945054e-05, | |
| "loss": 1.7143, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 18.02, | |
| "eval_loss": 0.6538747549057007, | |
| "eval_runtime": 236.0252, | |
| "eval_samples_per_second": 18.994, | |
| "eval_steps_per_second": 2.377, | |
| "eval_wer": 2.5976423487544484, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 18.47, | |
| "learning_rate": 5.772527472527472e-05, | |
| "loss": 1.7023, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 18.92, | |
| "learning_rate": 5.69010989010989e-05, | |
| "loss": 1.6887, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 19.37, | |
| "learning_rate": 5.607692307692307e-05, | |
| "loss": 1.6596, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 19.82, | |
| "learning_rate": 5.525274725274725e-05, | |
| "loss": 1.6613, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 20.27, | |
| "learning_rate": 5.442857142857143e-05, | |
| "loss": 1.6575, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 20.72, | |
| "learning_rate": 5.36043956043956e-05, | |
| "loss": 1.6486, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 21.17, | |
| "learning_rate": 5.278021978021977e-05, | |
| "loss": 1.6467, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 21.62, | |
| "learning_rate": 5.195604395604395e-05, | |
| "loss": 1.6206, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 22.07, | |
| "learning_rate": 5.1140109890109885e-05, | |
| "loss": 1.6048, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 22.52, | |
| "learning_rate": 5.031593406593406e-05, | |
| "loss": 1.6025, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 22.52, | |
| "eval_loss": 0.598939836025238, | |
| "eval_runtime": 237.9623, | |
| "eval_samples_per_second": 18.839, | |
| "eval_steps_per_second": 2.358, | |
| "eval_wer": 2.6034252669039146, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 22.97, | |
| "learning_rate": 4.949175824175824e-05, | |
| "loss": 1.6151, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 23.42, | |
| "learning_rate": 4.866758241758242e-05, | |
| "loss": 1.5959, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 23.87, | |
| "learning_rate": 4.784340659340658e-05, | |
| "loss": 1.5845, | |
| "step": 5300 | |
| }, | |
| { | |
| "epoch": 24.32, | |
| "learning_rate": 4.701923076923076e-05, | |
| "loss": 1.5662, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 24.77, | |
| "learning_rate": 4.619505494505494e-05, | |
| "loss": 1.5759, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 25.23, | |
| "learning_rate": 4.5370879120879115e-05, | |
| "loss": 1.5928, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 25.68, | |
| "learning_rate": 4.4546703296703295e-05, | |
| "loss": 1.5576, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 26.13, | |
| "learning_rate": 4.3722527472527474e-05, | |
| "loss": 1.5289, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 26.58, | |
| "learning_rate": 4.289835164835164e-05, | |
| "loss": 1.5381, | |
| "step": 5900 | |
| }, | |
| { | |
| "epoch": 27.03, | |
| "learning_rate": 4.207417582417582e-05, | |
| "loss": 1.5403, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 27.03, | |
| "eval_loss": 0.6035020351409912, | |
| "eval_runtime": 237.8256, | |
| "eval_samples_per_second": 18.85, | |
| "eval_steps_per_second": 2.359, | |
| "eval_wer": 2.69461743772242, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 27.48, | |
| "learning_rate": 4.125e-05, | |
| "loss": 1.5314, | |
| "step": 6100 | |
| }, | |
| { | |
| "epoch": 27.93, | |
| "learning_rate": 4.042582417582417e-05, | |
| "loss": 1.5186, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 28.38, | |
| "learning_rate": 3.960164835164835e-05, | |
| "loss": 1.5097, | |
| "step": 6300 | |
| }, | |
| { | |
| "epoch": 28.83, | |
| "learning_rate": 3.8777472527472525e-05, | |
| "loss": 1.52, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 29.28, | |
| "learning_rate": 3.79532967032967e-05, | |
| "loss": 1.51, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 29.73, | |
| "learning_rate": 3.712912087912088e-05, | |
| "loss": 1.5109, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 30.18, | |
| "learning_rate": 3.630494505494505e-05, | |
| "loss": 1.5219, | |
| "step": 6700 | |
| }, | |
| { | |
| "epoch": 30.63, | |
| "learning_rate": 3.548076923076922e-05, | |
| "loss": 1.5006, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 31.08, | |
| "learning_rate": 3.46565934065934e-05, | |
| "loss": 1.475, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 31.53, | |
| "learning_rate": 3.383241758241758e-05, | |
| "loss": 1.4773, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 31.53, | |
| "eval_loss": 0.5646877288818359, | |
| "eval_runtime": 239.6564, | |
| "eval_samples_per_second": 18.706, | |
| "eval_steps_per_second": 2.341, | |
| "eval_wer": 2.555827402135231, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 31.98, | |
| "learning_rate": 3.3008241758241755e-05, | |
| "loss": 1.4921, | |
| "step": 7100 | |
| }, | |
| { | |
| "epoch": 32.43, | |
| "learning_rate": 3.218406593406593e-05, | |
| "loss": 1.4634, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 32.88, | |
| "learning_rate": 3.135989010989011e-05, | |
| "loss": 1.4716, | |
| "step": 7300 | |
| }, | |
| { | |
| "epoch": 33.33, | |
| "learning_rate": 3.054395604395604e-05, | |
| "loss": 1.4313, | |
| "step": 7400 | |
| }, | |
| { | |
| "epoch": 33.78, | |
| "learning_rate": 2.971978021978022e-05, | |
| "loss": 1.4702, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 34.23, | |
| "learning_rate": 2.8895604395604393e-05, | |
| "loss": 1.4679, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 34.68, | |
| "learning_rate": 2.807142857142857e-05, | |
| "loss": 1.4379, | |
| "step": 7700 | |
| }, | |
| { | |
| "epoch": 35.14, | |
| "learning_rate": 2.7247252747252742e-05, | |
| "loss": 1.4211, | |
| "step": 7800 | |
| }, | |
| { | |
| "epoch": 35.59, | |
| "learning_rate": 2.642307692307692e-05, | |
| "loss": 1.4295, | |
| "step": 7900 | |
| }, | |
| { | |
| "epoch": 36.04, | |
| "learning_rate": 2.5598901098901098e-05, | |
| "loss": 1.4228, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 36.04, | |
| "eval_loss": 0.5477184057235718, | |
| "eval_runtime": 235.9806, | |
| "eval_samples_per_second": 18.997, | |
| "eval_steps_per_second": 2.377, | |
| "eval_wer": 2.5676156583629894, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 36.49, | |
| "learning_rate": 2.477472527472527e-05, | |
| "loss": 1.4199, | |
| "step": 8100 | |
| }, | |
| { | |
| "epoch": 36.94, | |
| "learning_rate": 2.395054945054945e-05, | |
| "loss": 1.4317, | |
| "step": 8200 | |
| }, | |
| { | |
| "epoch": 37.39, | |
| "learning_rate": 2.3126373626373626e-05, | |
| "loss": 1.4048, | |
| "step": 8300 | |
| }, | |
| { | |
| "epoch": 37.84, | |
| "learning_rate": 2.23021978021978e-05, | |
| "loss": 1.4232, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 38.29, | |
| "learning_rate": 2.147802197802198e-05, | |
| "loss": 1.4076, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 38.74, | |
| "learning_rate": 2.065384615384615e-05, | |
| "loss": 1.4009, | |
| "step": 8600 | |
| }, | |
| { | |
| "epoch": 39.19, | |
| "learning_rate": 1.9829670329670328e-05, | |
| "loss": 1.4173, | |
| "step": 8700 | |
| }, | |
| { | |
| "epoch": 39.64, | |
| "learning_rate": 1.90054945054945e-05, | |
| "loss": 1.4028, | |
| "step": 8800 | |
| }, | |
| { | |
| "epoch": 40.09, | |
| "learning_rate": 1.818131868131868e-05, | |
| "loss": 1.3694, | |
| "step": 8900 | |
| }, | |
| { | |
| "epoch": 40.54, | |
| "learning_rate": 1.7357142857142856e-05, | |
| "loss": 1.3801, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 40.54, | |
| "eval_loss": 0.5413390398025513, | |
| "eval_runtime": 239.6653, | |
| "eval_samples_per_second": 18.705, | |
| "eval_steps_per_second": 2.341, | |
| "eval_wer": 2.619217081850534, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 40.99, | |
| "learning_rate": 1.6532967032967033e-05, | |
| "loss": 1.4054, | |
| "step": 9100 | |
| }, | |
| { | |
| "epoch": 41.44, | |
| "learning_rate": 1.5708791208791206e-05, | |
| "loss": 1.3823, | |
| "step": 9200 | |
| }, | |
| { | |
| "epoch": 41.89, | |
| "learning_rate": 1.4884615384615383e-05, | |
| "loss": 1.3902, | |
| "step": 9300 | |
| }, | |
| { | |
| "epoch": 42.34, | |
| "learning_rate": 1.4060439560439558e-05, | |
| "loss": 1.3568, | |
| "step": 9400 | |
| }, | |
| { | |
| "epoch": 42.79, | |
| "learning_rate": 1.3236263736263736e-05, | |
| "loss": 1.3766, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 43.24, | |
| "learning_rate": 1.2412087912087912e-05, | |
| "loss": 1.3883, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 43.69, | |
| "learning_rate": 1.1587912087912087e-05, | |
| "loss": 1.3666, | |
| "step": 9700 | |
| }, | |
| { | |
| "epoch": 44.14, | |
| "learning_rate": 1.0763736263736263e-05, | |
| "loss": 1.3718, | |
| "step": 9800 | |
| }, | |
| { | |
| "epoch": 44.59, | |
| "learning_rate": 9.939560439560437e-06, | |
| "loss": 1.367, | |
| "step": 9900 | |
| }, | |
| { | |
| "epoch": 45.05, | |
| "learning_rate": 9.115384615384615e-06, | |
| "loss": 1.3558, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 45.05, | |
| "eval_loss": 0.5343097448348999, | |
| "eval_runtime": 238.241, | |
| "eval_samples_per_second": 18.817, | |
| "eval_steps_per_second": 2.355, | |
| "eval_wer": 2.657473309608541, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 45.5, | |
| "learning_rate": 8.29120879120879e-06, | |
| "loss": 1.3595, | |
| "step": 10100 | |
| }, | |
| { | |
| "epoch": 45.95, | |
| "learning_rate": 7.467032967032966e-06, | |
| "loss": 1.3697, | |
| "step": 10200 | |
| }, | |
| { | |
| "epoch": 46.4, | |
| "learning_rate": 6.6428571428571416e-06, | |
| "loss": 1.3423, | |
| "step": 10300 | |
| }, | |
| { | |
| "epoch": 46.85, | |
| "learning_rate": 5.818681318681319e-06, | |
| "loss": 1.341, | |
| "step": 10400 | |
| }, | |
| { | |
| "epoch": 47.3, | |
| "learning_rate": 4.994505494505494e-06, | |
| "loss": 1.3289, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 47.75, | |
| "learning_rate": 4.17032967032967e-06, | |
| "loss": 1.3462, | |
| "step": 10600 | |
| }, | |
| { | |
| "epoch": 48.2, | |
| "learning_rate": 3.3461538461538457e-06, | |
| "loss": 1.3636, | |
| "step": 10700 | |
| }, | |
| { | |
| "epoch": 48.65, | |
| "learning_rate": 2.521978021978022e-06, | |
| "loss": 1.3366, | |
| "step": 10800 | |
| }, | |
| { | |
| "epoch": 49.1, | |
| "learning_rate": 1.6978021978021975e-06, | |
| "loss": 1.337, | |
| "step": 10900 | |
| }, | |
| { | |
| "epoch": 49.55, | |
| "learning_rate": 8.818681318681318e-07, | |
| "loss": 1.3298, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 49.55, | |
| "eval_loss": 0.5348710417747498, | |
| "eval_runtime": 238.2007, | |
| "eval_samples_per_second": 18.82, | |
| "eval_steps_per_second": 2.355, | |
| "eval_wer": 2.627446619217082, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 50.0, | |
| "learning_rate": 5.769230769230769e-08, | |
| "loss": 1.3663, | |
| "step": 11100 | |
| }, | |
| { | |
| "epoch": 50.0, | |
| "step": 11100, | |
| "total_flos": 8.246726177897223e+19, | |
| "train_loss": 2.741275282851211, | |
| "train_runtime": 31325.7849, | |
| "train_samples_per_second": 16.956, | |
| "train_steps_per_second": 0.354 | |
| } | |
| ], | |
| "max_steps": 11100, | |
| "num_train_epochs": 50, | |
| "total_flos": 8.246726177897223e+19, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |