| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.6775106082036775, | |
| "eval_steps": 100, | |
| "global_step": 2600, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.14144271570014144, | |
| "eval_loss": 3.2194931507110596, | |
| "eval_runtime": 152.4732, | |
| "eval_samples_per_second": 37.095, | |
| "eval_steps_per_second": 4.637, | |
| "eval_wer": 1.0, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.2828854314002829, | |
| "eval_loss": 3.1180293560028076, | |
| "eval_runtime": 149.651, | |
| "eval_samples_per_second": 37.795, | |
| "eval_steps_per_second": 4.724, | |
| "eval_wer": 1.0, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.4243281471004243, | |
| "eval_loss": 1.7235759496688843, | |
| "eval_runtime": 150.6233, | |
| "eval_samples_per_second": 37.551, | |
| "eval_steps_per_second": 4.694, | |
| "eval_wer": 0.9364317696714866, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.5657708628005658, | |
| "eval_loss": 1.157871127128601, | |
| "eval_runtime": 150.7722, | |
| "eval_samples_per_second": 37.514, | |
| "eval_steps_per_second": 4.689, | |
| "eval_wer": 0.76277061834989, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.7072135785007072, | |
| "grad_norm": 2.786440372467041, | |
| "learning_rate": 0.0002964, | |
| "loss": 3.0279, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.7072135785007072, | |
| "eval_loss": 0.9558188319206238, | |
| "eval_runtime": 149.998, | |
| "eval_samples_per_second": 37.707, | |
| "eval_steps_per_second": 4.713, | |
| "eval_wer": 0.6721285166343022, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.8486562942008486, | |
| "eval_loss": 0.8188507556915283, | |
| "eval_runtime": 150.6554, | |
| "eval_samples_per_second": 37.543, | |
| "eval_steps_per_second": 4.693, | |
| "eval_wer": 0.6423905891415641, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.9900990099009901, | |
| "eval_loss": 0.6972773671150208, | |
| "eval_runtime": 152.1964, | |
| "eval_samples_per_second": 37.163, | |
| "eval_steps_per_second": 4.645, | |
| "eval_wer": 0.5163775256375279, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.1315417256011315, | |
| "eval_loss": 0.6183159947395325, | |
| "eval_runtime": 151.4826, | |
| "eval_samples_per_second": 37.338, | |
| "eval_steps_per_second": 4.667, | |
| "eval_wer": 0.4751809471842853, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.272984441301273, | |
| "eval_loss": 0.5935864448547363, | |
| "eval_runtime": 150.8934, | |
| "eval_samples_per_second": 37.483, | |
| "eval_steps_per_second": 4.685, | |
| "eval_wer": 0.47028614530339746, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.4144271570014144, | |
| "grad_norm": 1.0183278322219849, | |
| "learning_rate": 0.0002294285714285714, | |
| "loss": 0.7925, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.4144271570014144, | |
| "eval_loss": 0.5497952103614807, | |
| "eval_runtime": 151.0928, | |
| "eval_samples_per_second": 37.434, | |
| "eval_steps_per_second": 4.679, | |
| "eval_wer": 0.4304055463722296, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.5558698727015559, | |
| "eval_loss": 0.5285686254501343, | |
| "eval_runtime": 151.3525, | |
| "eval_samples_per_second": 37.37, | |
| "eval_steps_per_second": 4.671, | |
| "eval_wer": 0.4148224230071737, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.6973125884016973, | |
| "eval_loss": 0.5129849314689636, | |
| "eval_runtime": 151.9884, | |
| "eval_samples_per_second": 37.213, | |
| "eval_steps_per_second": 4.652, | |
| "eval_wer": 0.39739371860506173, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.8387553041018387, | |
| "eval_loss": 0.48775196075439453, | |
| "eval_runtime": 152.7263, | |
| "eval_samples_per_second": 37.034, | |
| "eval_steps_per_second": 4.629, | |
| "eval_wer": 0.3863844265057534, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.9801980198019802, | |
| "eval_loss": 0.4740111827850342, | |
| "eval_runtime": 152.3303, | |
| "eval_samples_per_second": 37.13, | |
| "eval_steps_per_second": 4.641, | |
| "eval_wer": 0.3733209224695479, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 2.1216407355021216, | |
| "grad_norm": 0.966443657875061, | |
| "learning_rate": 0.0001584285714285714, | |
| "loss": 0.62, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.1216407355021216, | |
| "eval_loss": 0.45776596665382385, | |
| "eval_runtime": 152.151, | |
| "eval_samples_per_second": 37.174, | |
| "eval_steps_per_second": 4.647, | |
| "eval_wer": 0.3593266036494359, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.263083451202263, | |
| "eval_loss": 0.45351675152778625, | |
| "eval_runtime": 151.9696, | |
| "eval_samples_per_second": 37.218, | |
| "eval_steps_per_second": 4.652, | |
| "eval_wer": 0.35000240727961357, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 2.4045261669024045, | |
| "eval_loss": 0.44846034049987793, | |
| "eval_runtime": 152.2898, | |
| "eval_samples_per_second": 37.14, | |
| "eval_steps_per_second": 4.642, | |
| "eval_wer": 0.3496814366644734, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 2.545968882602546, | |
| "eval_loss": 0.4373452067375183, | |
| "eval_runtime": 152.3551, | |
| "eval_samples_per_second": 37.124, | |
| "eval_steps_per_second": 4.64, | |
| "eval_wer": 0.34305339346182856, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 2.6874115983026874, | |
| "eval_loss": 0.4362298250198364, | |
| "eval_runtime": 151.9091, | |
| "eval_samples_per_second": 37.233, | |
| "eval_steps_per_second": 4.654, | |
| "eval_wer": 0.34286081109274447, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 2.828854314002829, | |
| "grad_norm": 0.6778371334075928, | |
| "learning_rate": 8.742857142857142e-05, | |
| "loss": 0.4879, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.828854314002829, | |
| "eval_loss": 0.4235607981681824, | |
| "eval_runtime": 152.5551, | |
| "eval_samples_per_second": 37.075, | |
| "eval_steps_per_second": 4.634, | |
| "eval_wer": 0.33270209112355764, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.9702970297029703, | |
| "eval_loss": 0.4171510338783264, | |
| "eval_runtime": 152.8349, | |
| "eval_samples_per_second": 37.007, | |
| "eval_steps_per_second": 4.626, | |
| "eval_wer": 0.3256728346519876, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 3.1117397454031117, | |
| "eval_loss": 0.4206392467021942, | |
| "eval_runtime": 152.8774, | |
| "eval_samples_per_second": 36.997, | |
| "eval_steps_per_second": 4.625, | |
| "eval_wer": 0.32169279902424935, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 3.253182461103253, | |
| "eval_loss": 0.41660308837890625, | |
| "eval_runtime": 152.074, | |
| "eval_samples_per_second": 37.192, | |
| "eval_steps_per_second": 4.649, | |
| "eval_wer": 0.3198632665179503, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 3.3946251768033946, | |
| "eval_loss": 0.41343268752098083, | |
| "eval_runtime": 152.5935, | |
| "eval_samples_per_second": 37.066, | |
| "eval_steps_per_second": 4.633, | |
| "eval_wer": 0.3172955015968288, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 3.536067892503536, | |
| "grad_norm": 0.7615213990211487, | |
| "learning_rate": 1.6285714285714283e-05, | |
| "loss": 0.4036, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 3.536067892503536, | |
| "eval_loss": 0.41102421283721924, | |
| "eval_runtime": 151.5019, | |
| "eval_samples_per_second": 37.333, | |
| "eval_steps_per_second": 4.667, | |
| "eval_wer": 0.31594742501324, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 3.6775106082036775, | |
| "eval_loss": 0.41045042872428894, | |
| "eval_runtime": 152.2158, | |
| "eval_samples_per_second": 37.158, | |
| "eval_steps_per_second": 4.645, | |
| "eval_wer": 0.3136203880534737, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 3.6775106082036775, | |
| "step": 2600, | |
| "total_flos": 9.93123920141323e+18, | |
| "train_loss": 1.040142478942871, | |
| "train_runtime": 6732.1803, | |
| "train_samples_per_second": 12.359, | |
| "train_steps_per_second": 0.386 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 2600, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 400, | |
| "total_flos": 9.93123920141323e+18, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |