{ "best_metric": 22.162243780906724, "best_model_checkpoint": "Models/whisper-small-sv-extra-data\\checkpoint-2000", "epoch": 2.6260504201680672, "eval_steps": 500, "global_step": 2500, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.03, "learning_rate": 4.2000000000000006e-07, "loss": 3.4232, "step": 25 }, { "epoch": 0.05, "learning_rate": 9.200000000000001e-07, "loss": 2.9342, "step": 50 }, { "epoch": 0.08, "learning_rate": 1.42e-06, "loss": 2.1531, "step": 75 }, { "epoch": 0.11, "learning_rate": 1.9200000000000003e-06, "loss": 1.3958, "step": 100 }, { "epoch": 0.13, "learning_rate": 2.42e-06, "loss": 1.0906, "step": 125 }, { "epoch": 0.16, "learning_rate": 2.92e-06, "loss": 0.9704, "step": 150 }, { "epoch": 0.18, "learning_rate": 3.4200000000000007e-06, "loss": 0.8795, "step": 175 }, { "epoch": 0.21, "learning_rate": 3.920000000000001e-06, "loss": 0.7978, "step": 200 }, { "epoch": 0.24, "learning_rate": 4.42e-06, "loss": 0.7003, "step": 225 }, { "epoch": 0.26, "learning_rate": 4.92e-06, "loss": 0.6263, "step": 250 }, { "epoch": 0.29, "learning_rate": 5.420000000000001e-06, "loss": 0.5262, "step": 275 }, { "epoch": 0.32, "learning_rate": 5.92e-06, "loss": 0.3673, "step": 300 }, { "epoch": 0.34, "learning_rate": 6.42e-06, "loss": 0.3793, "step": 325 }, { "epoch": 0.37, "learning_rate": 6.92e-06, "loss": 0.3302, "step": 350 }, { "epoch": 0.39, "learning_rate": 7.420000000000001e-06, "loss": 0.3261, "step": 375 }, { "epoch": 0.42, "learning_rate": 7.92e-06, "loss": 0.3556, "step": 400 }, { "epoch": 0.45, "learning_rate": 8.42e-06, "loss": 0.3235, "step": 425 }, { "epoch": 0.47, "learning_rate": 8.920000000000001e-06, "loss": 0.3257, "step": 450 }, { "epoch": 0.5, "learning_rate": 9.42e-06, "loss": 0.2913, "step": 475 }, { "epoch": 0.53, "learning_rate": 9.920000000000002e-06, "loss": 0.3189, "step": 500 }, { "epoch": 0.53, "eval_loss": 0.3610133230686188, "eval_runtime": 1451.6344, "eval_samples_per_second": 4.05, "eval_steps_per_second": 0.254, "eval_wer": 25.202899097806803, "step": 500 }, { "epoch": 0.55, "learning_rate": 1e-05, "loss": 0.3253, "step": 525 }, { "epoch": 0.58, "learning_rate": 1e-05, "loss": 0.3088, "step": 550 }, { "epoch": 0.6, "learning_rate": 1e-05, "loss": 0.2972, "step": 575 }, { "epoch": 0.63, "learning_rate": 1e-05, "loss": 0.3002, "step": 600 }, { "epoch": 0.66, "learning_rate": 1e-05, "loss": 0.2983, "step": 625 }, { "epoch": 0.68, "learning_rate": 1e-05, "loss": 0.3121, "step": 650 }, { "epoch": 0.71, "learning_rate": 1e-05, "loss": 0.2909, "step": 675 }, { "epoch": 0.74, "learning_rate": 1e-05, "loss": 0.3386, "step": 700 }, { "epoch": 0.76, "learning_rate": 1e-05, "loss": 0.2835, "step": 725 }, { "epoch": 0.79, "learning_rate": 1e-05, "loss": 0.294, "step": 750 }, { "epoch": 0.81, "learning_rate": 1e-05, "loss": 0.2854, "step": 775 }, { "epoch": 0.84, "learning_rate": 1e-05, "loss": 0.2808, "step": 800 }, { "epoch": 0.87, "learning_rate": 1e-05, "loss": 0.2929, "step": 825 }, { "epoch": 0.89, "learning_rate": 1e-05, "loss": 0.2894, "step": 850 }, { "epoch": 0.92, "learning_rate": 1e-05, "loss": 0.2773, "step": 875 }, { "epoch": 0.95, "learning_rate": 1e-05, "loss": 0.2896, "step": 900 }, { "epoch": 0.97, "learning_rate": 1e-05, "loss": 0.2758, "step": 925 }, { "epoch": 1.0, "learning_rate": 1e-05, "loss": 0.2582, "step": 950 }, { "epoch": 1.02, "learning_rate": 1e-05, "loss": 0.1561, "step": 975 }, { "epoch": 1.05, "learning_rate": 1e-05, "loss": 0.1451, "step": 1000 }, { "epoch": 1.05, "eval_loss": 0.3336940407752991, "eval_runtime": 1109.0015, "eval_samples_per_second": 5.301, "eval_steps_per_second": 0.332, "eval_wer": 23.649541353667285, "step": 1000 }, { "epoch": 1.08, "learning_rate": 1e-05, "loss": 0.1528, "step": 1025 }, { "epoch": 1.1, "learning_rate": 1e-05, "loss": 0.1421, "step": 1050 }, { "epoch": 1.13, "learning_rate": 1e-05, "loss": 0.1536, "step": 1075 }, { "epoch": 1.16, "learning_rate": 1e-05, "loss": 0.1437, "step": 1100 }, { "epoch": 1.18, "learning_rate": 1e-05, "loss": 0.148, "step": 1125 }, { "epoch": 1.21, "learning_rate": 1e-05, "loss": 0.1419, "step": 1150 }, { "epoch": 1.23, "learning_rate": 1e-05, "loss": 0.1495, "step": 1175 }, { "epoch": 1.26, "learning_rate": 1e-05, "loss": 0.1395, "step": 1200 }, { "epoch": 1.29, "learning_rate": 1e-05, "loss": 0.1481, "step": 1225 }, { "epoch": 1.31, "learning_rate": 1e-05, "loss": 0.137, "step": 1250 }, { "epoch": 1.34, "learning_rate": 1e-05, "loss": 0.1429, "step": 1275 }, { "epoch": 1.37, "learning_rate": 1e-05, "loss": 0.1268, "step": 1300 }, { "epoch": 1.39, "learning_rate": 1e-05, "loss": 0.1368, "step": 1325 }, { "epoch": 1.42, "learning_rate": 1e-05, "loss": 0.148, "step": 1350 }, { "epoch": 1.44, "learning_rate": 1e-05, "loss": 0.1479, "step": 1375 }, { "epoch": 1.47, "learning_rate": 1e-05, "loss": 0.1326, "step": 1400 }, { "epoch": 1.5, "learning_rate": 1e-05, "loss": 0.1428, "step": 1425 }, { "epoch": 1.52, "learning_rate": 1e-05, "loss": 0.164, "step": 1450 }, { "epoch": 1.55, "learning_rate": 1e-05, "loss": 0.1381, "step": 1475 }, { "epoch": 1.58, "learning_rate": 1e-05, "loss": 0.1432, "step": 1500 }, { "epoch": 1.58, "eval_loss": 0.32634562253952026, "eval_runtime": 1113.1016, "eval_samples_per_second": 5.282, "eval_steps_per_second": 0.331, "eval_wer": 22.89079309954324, "step": 1500 }, { "epoch": 1.6, "learning_rate": 1e-05, "loss": 0.1419, "step": 1525 }, { "epoch": 1.63, "learning_rate": 1e-05, "loss": 0.1485, "step": 1550 }, { "epoch": 1.65, "learning_rate": 1e-05, "loss": 0.1456, "step": 1575 }, { "epoch": 1.68, "learning_rate": 1e-05, "loss": 0.1394, "step": 1600 }, { "epoch": 1.71, "learning_rate": 1e-05, "loss": 0.1546, "step": 1625 }, { "epoch": 1.73, "learning_rate": 1e-05, "loss": 0.1498, "step": 1650 }, { "epoch": 1.76, "learning_rate": 1e-05, "loss": 0.1338, "step": 1675 }, { "epoch": 1.79, "learning_rate": 1e-05, "loss": 0.1381, "step": 1700 }, { "epoch": 1.81, "learning_rate": 1e-05, "loss": 0.1461, "step": 1725 }, { "epoch": 1.84, "learning_rate": 1e-05, "loss": 0.1493, "step": 1750 }, { "epoch": 1.86, "learning_rate": 1e-05, "loss": 0.124, "step": 1775 }, { "epoch": 1.89, "learning_rate": 1e-05, "loss": 0.1356, "step": 1800 }, { "epoch": 1.92, "learning_rate": 1e-05, "loss": 0.1452, "step": 1825 }, { "epoch": 1.94, "learning_rate": 1e-05, "loss": 0.1473, "step": 1850 }, { "epoch": 1.97, "learning_rate": 1e-05, "loss": 0.1364, "step": 1875 }, { "epoch": 2.0, "learning_rate": 1e-05, "loss": 0.1471, "step": 1900 }, { "epoch": 2.02, "learning_rate": 1e-05, "loss": 0.071, "step": 1925 }, { "epoch": 2.05, "learning_rate": 1e-05, "loss": 0.057, "step": 1950 }, { "epoch": 2.07, "learning_rate": 1e-05, "loss": 0.0621, "step": 1975 }, { "epoch": 2.1, "learning_rate": 1e-05, "loss": 0.0572, "step": 2000 }, { "epoch": 2.1, "eval_loss": 0.32835841178894043, "eval_runtime": 1103.1902, "eval_samples_per_second": 5.329, "eval_steps_per_second": 0.334, "eval_wer": 22.162243780906724, "step": 2000 }, { "epoch": 2.13, "learning_rate": 1e-05, "loss": 0.0534, "step": 2025 }, { "epoch": 2.15, "learning_rate": 1e-05, "loss": 0.058, "step": 2050 }, { "epoch": 2.18, "learning_rate": 1e-05, "loss": 0.0612, "step": 2075 }, { "epoch": 2.21, "learning_rate": 1e-05, "loss": 0.0486, "step": 2100 }, { "epoch": 2.23, "learning_rate": 1e-05, "loss": 0.0533, "step": 2125 }, { "epoch": 2.26, "learning_rate": 1e-05, "loss": 0.0584, "step": 2150 }, { "epoch": 2.28, "learning_rate": 1e-05, "loss": 0.0622, "step": 2175 }, { "epoch": 2.31, "learning_rate": 1e-05, "loss": 0.0507, "step": 2200 }, { "epoch": 2.34, "learning_rate": 1e-05, "loss": 0.0563, "step": 2225 }, { "epoch": 2.36, "learning_rate": 1e-05, "loss": 0.0532, "step": 2250 }, { "epoch": 2.39, "learning_rate": 1e-05, "loss": 0.057, "step": 2275 }, { "epoch": 2.42, "learning_rate": 1e-05, "loss": 0.0557, "step": 2300 }, { "epoch": 2.44, "learning_rate": 1e-05, "loss": 0.0553, "step": 2325 }, { "epoch": 2.47, "learning_rate": 1e-05, "loss": 0.0532, "step": 2350 }, { "epoch": 2.49, "learning_rate": 1e-05, "loss": 0.0579, "step": 2375 }, { "epoch": 2.52, "learning_rate": 1e-05, "loss": 0.0594, "step": 2400 }, { "epoch": 2.55, "learning_rate": 1e-05, "loss": 0.0594, "step": 2425 }, { "epoch": 2.57, "learning_rate": 1e-05, "loss": 0.063, "step": 2450 }, { "epoch": 2.6, "learning_rate": 1e-05, "loss": 0.0639, "step": 2475 }, { "epoch": 2.63, "learning_rate": 1e-05, "loss": 0.0502, "step": 2500 }, { "epoch": 2.63, "eval_loss": 0.34048977494239807, "eval_runtime": 1106.2301, "eval_samples_per_second": 5.314, "eval_steps_per_second": 0.333, "eval_wer": 22.19999245026613, "step": 2500 } ], "logging_steps": 25, "max_steps": 4000, "num_input_tokens_seen": 0, "num_train_epochs": 5, "save_steps": 500, "total_flos": 1.153879864639488e+19, "trial_name": null, "trial_params": null }