| { | |
| "best_metric": 7.411630558722919, | |
| "best_model_checkpoint": "../whisper-NST2-unfreeze-constanti-low-lr/checkpoint-14000", | |
| "epoch": 1.0001, | |
| "global_step": 20002, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01, | |
| "learning_rate": 2.47e-06, | |
| "loss": 1.7243, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 4.970000000000001e-06, | |
| "loss": 0.3597, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 7.4700000000000005e-06, | |
| "loss": 0.2237, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "learning_rate": 9.970000000000001e-06, | |
| "loss": 0.1901, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "eval_loss": 0.30687370896339417, | |
| "eval_runtime": 48.6375, | |
| "eval_samples_per_second": 2.056, | |
| "eval_steps_per_second": 0.144, | |
| "eval_wer": 14.823261117445838, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 1e-05, | |
| "loss": 0.1714, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "learning_rate": 1e-05, | |
| "loss": 0.1581, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "learning_rate": 1e-05, | |
| "loss": 0.1451, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 1e-05, | |
| "loss": 0.1323, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "eval_loss": 0.2687492072582245, | |
| "eval_runtime": 59.4025, | |
| "eval_samples_per_second": 1.683, | |
| "eval_steps_per_second": 0.118, | |
| "eval_wer": 11.2884834663626, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "learning_rate": 1e-05, | |
| "loss": 0.1283, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 1e-05, | |
| "loss": 0.1227, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "learning_rate": 1e-05, | |
| "loss": 0.1172, | |
| "step": 2750 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 1e-05, | |
| "loss": 0.1137, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "eval_loss": 0.26204466819763184, | |
| "eval_runtime": 53.0968, | |
| "eval_samples_per_second": 1.883, | |
| "eval_steps_per_second": 0.132, | |
| "eval_wer": 10.832383124287343, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 1e-05, | |
| "loss": 0.1104, | |
| "step": 3250 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 1e-05, | |
| "loss": 0.1071, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 1e-05, | |
| "loss": 0.1022, | |
| "step": 3750 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 1e-05, | |
| "loss": 0.1022, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "eval_loss": 0.297575443983078, | |
| "eval_runtime": 45.3592, | |
| "eval_samples_per_second": 2.205, | |
| "eval_steps_per_second": 0.154, | |
| "eval_wer": 9.007981755986318, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 1e-05, | |
| "loss": 0.096, | |
| "step": 4250 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0939, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0918, | |
| "step": 4750 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0937, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "eval_loss": 0.25842034816741943, | |
| "eval_runtime": 48.6209, | |
| "eval_samples_per_second": 2.057, | |
| "eval_steps_per_second": 0.144, | |
| "eval_wer": 9.578107183580387, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0895, | |
| "step": 5250 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0874, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0875, | |
| "step": 5750 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0875, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "eval_loss": 0.27036869525909424, | |
| "eval_runtime": 76.4925, | |
| "eval_samples_per_second": 1.307, | |
| "eval_steps_per_second": 0.092, | |
| "eval_wer": 20.296465222348917, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0816, | |
| "step": 6250 | |
| }, | |
| { | |
| "epoch": 1.02, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0748, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0656, | |
| "step": 6750 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0592, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "eval_loss": 0.2750946283340454, | |
| "eval_runtime": 45.0139, | |
| "eval_samples_per_second": 2.222, | |
| "eval_steps_per_second": 0.156, | |
| "eval_wer": 9.007981755986318, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 1.06, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0553, | |
| "step": 7250 | |
| }, | |
| { | |
| "epoch": 1.07, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0542, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0533, | |
| "step": 7750 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0488, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "eval_loss": 0.27778884768486023, | |
| "eval_runtime": 49.8405, | |
| "eval_samples_per_second": 2.006, | |
| "eval_steps_per_second": 0.14, | |
| "eval_wer": 8.665906499429875, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0481, | |
| "step": 8250 | |
| }, | |
| { | |
| "epoch": 1.12, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0477, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 1.14, | |
| "learning_rate": 1e-05, | |
| "loss": 0.047, | |
| "step": 8750 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0475, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "eval_loss": 0.2791878581047058, | |
| "eval_runtime": 48.8356, | |
| "eval_samples_per_second": 2.048, | |
| "eval_steps_per_second": 0.143, | |
| "eval_wer": 9.464082098061574, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.16, | |
| "learning_rate": 1e-05, | |
| "loss": 0.045, | |
| "step": 9250 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0464, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0419, | |
| "step": 9750 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0439, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "eval_loss": 0.2879847586154938, | |
| "eval_runtime": 49.0674, | |
| "eval_samples_per_second": 2.038, | |
| "eval_steps_per_second": 0.143, | |
| "eval_wer": 8.323831242873432, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0413, | |
| "step": 10250 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "learning_rate": 1e-05, | |
| "loss": 0.041, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 1.24, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0413, | |
| "step": 10750 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0425, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "eval_loss": 0.2954462766647339, | |
| "eval_runtime": 46.9453, | |
| "eval_samples_per_second": 2.13, | |
| "eval_steps_per_second": 0.149, | |
| "eval_wer": 8.551881413911062, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 1.26, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0409, | |
| "step": 11250 | |
| }, | |
| { | |
| "epoch": 1.27, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0389, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 1.29, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0403, | |
| "step": 11750 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0416, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "eval_loss": 0.2895849347114563, | |
| "eval_runtime": 72.8965, | |
| "eval_samples_per_second": 1.372, | |
| "eval_steps_per_second": 0.096, | |
| "eval_wer": 20.296465222348917, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 2.01, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0384, | |
| "step": 12250 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0355, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 2.03, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0323, | |
| "step": 12750 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0289, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "eval_loss": 0.29897502064704895, | |
| "eval_runtime": 54.9916, | |
| "eval_samples_per_second": 1.818, | |
| "eval_steps_per_second": 0.127, | |
| "eval_wer": 7.98175598631699, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 2.06, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0266, | |
| "step": 13250 | |
| }, | |
| { | |
| "epoch": 2.07, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0264, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0267, | |
| "step": 13750 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0229, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "eval_loss": 0.3026929497718811, | |
| "eval_runtime": 48.2743, | |
| "eval_samples_per_second": 2.071, | |
| "eval_steps_per_second": 0.145, | |
| "eval_wer": 7.411630558722919, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 2.11, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0228, | |
| "step": 14250 | |
| }, | |
| { | |
| "epoch": 2.12, | |
| "learning_rate": 1e-05, | |
| "loss": 0.024, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0238, | |
| "step": 14750 | |
| }, | |
| { | |
| "epoch": 2.15, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0248, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 2.15, | |
| "eval_loss": 0.29682403802871704, | |
| "eval_runtime": 47.3876, | |
| "eval_samples_per_second": 2.11, | |
| "eval_steps_per_second": 0.148, | |
| "eval_wer": 8.665906499429875, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 2.16, | |
| "learning_rate": 1e-05, | |
| "loss": 0.022, | |
| "step": 15250 | |
| }, | |
| { | |
| "epoch": 2.17, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0241, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 2.18, | |
| "learning_rate": 1e-05, | |
| "loss": 0.021, | |
| "step": 15750 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0225, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "eval_loss": 0.30999651551246643, | |
| "eval_runtime": 56.9907, | |
| "eval_samples_per_second": 1.755, | |
| "eval_steps_per_second": 0.123, | |
| "eval_wer": 8.551881413911062, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 2.21, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0211, | |
| "step": 16250 | |
| }, | |
| { | |
| "epoch": 2.22, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0214, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 2.23, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0215, | |
| "step": 16750 | |
| }, | |
| { | |
| "epoch": 2.25, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0222, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 2.25, | |
| "eval_loss": 0.3132005035877228, | |
| "eval_runtime": 54.3277, | |
| "eval_samples_per_second": 1.841, | |
| "eval_steps_per_second": 0.129, | |
| "eval_wer": 9.35005701254276, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 2.26, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0211, | |
| "step": 17250 | |
| }, | |
| { | |
| "epoch": 2.27, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0199, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 2.28, | |
| "learning_rate": 1e-05, | |
| "loss": 0.021, | |
| "step": 17750 | |
| }, | |
| { | |
| "epoch": 2.3, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0219, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 2.3, | |
| "eval_loss": 0.3229832053184509, | |
| "eval_runtime": 56.2842, | |
| "eval_samples_per_second": 1.777, | |
| "eval_steps_per_second": 0.124, | |
| "eval_wer": 7.639680729760548, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 3.01, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0205, | |
| "step": 18250 | |
| }, | |
| { | |
| "epoch": 3.02, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0194, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 3.03, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0178, | |
| "step": 18750 | |
| }, | |
| { | |
| "epoch": 3.04, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0162, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 3.04, | |
| "eval_loss": 0.33802878856658936, | |
| "eval_runtime": 56.712, | |
| "eval_samples_per_second": 1.763, | |
| "eval_steps_per_second": 0.123, | |
| "eval_wer": 9.806157354618016, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 3.06, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0144, | |
| "step": 19250 | |
| }, | |
| { | |
| "epoch": 3.07, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0151, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 3.08, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0151, | |
| "step": 19750 | |
| }, | |
| { | |
| "epoch": 3.09, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0132, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 3.09, | |
| "eval_loss": 0.35617145895957947, | |
| "eval_runtime": 42.4011, | |
| "eval_samples_per_second": 2.358, | |
| "eval_steps_per_second": 0.165, | |
| "eval_wer": 8.551881413911062, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 3.09, | |
| "step": 20000, | |
| "total_flos": 2.7701254859636736e+20, | |
| "train_loss": 0.0831078713297844, | |
| "train_runtime": 247222.6725, | |
| "train_samples_per_second": 7.766, | |
| "train_steps_per_second": 0.081 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 20001, | |
| "total_flos": 2.7702640069558272e+20, | |
| "train_loss": 1.7381801103855406e-06, | |
| "train_runtime": 64.0295, | |
| "train_samples_per_second": 29986.165, | |
| "train_steps_per_second": 312.356 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 20002, | |
| "total_flos": 2.7704025279479808e+20, | |
| "train_loss": 5.53145265653126e-07, | |
| "train_runtime": 51.8572, | |
| "train_samples_per_second": 37024.777, | |
| "train_steps_per_second": 385.675 | |
| } | |
| ], | |
| "max_steps": 20000, | |
| "num_train_epochs": 9223372036854775807, | |
| "total_flos": 2.7704025279479808e+20, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |