| { | |
| "best_metric": 20.77837333954789, | |
| "best_model_checkpoint": "./the-final-whisper/checkpoint-500", | |
| "epoch": 0.8237232289950577, | |
| "eval_steps": 125, | |
| "global_step": 500, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.04118616144975288, | |
| "grad_norm": 52.4793701171875, | |
| "learning_rate": 4.2000000000000006e-07, | |
| "loss": 2.7503, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.08237232289950576, | |
| "grad_norm": 16.525333404541016, | |
| "learning_rate": 9.200000000000001e-07, | |
| "loss": 2.0994, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.12355848434925865, | |
| "grad_norm": 14.271560668945312, | |
| "learning_rate": 1.42e-06, | |
| "loss": 1.3213, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.16474464579901152, | |
| "grad_norm": 8.628070831298828, | |
| "learning_rate": 1.9200000000000003e-06, | |
| "loss": 0.7342, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.20593080724876442, | |
| "grad_norm": 8.307537078857422, | |
| "learning_rate": 2.42e-06, | |
| "loss": 0.5674, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.20593080724876442, | |
| "eval_loss": 0.6090311408042908, | |
| "eval_runtime": 2805.8489, | |
| "eval_samples_per_second": 1.442, | |
| "eval_steps_per_second": 0.18, | |
| "eval_wer": 45.15497553017945, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.2471169686985173, | |
| "grad_norm": 6.402714729309082, | |
| "learning_rate": 2.92e-06, | |
| "loss": 0.4853, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.2883031301482702, | |
| "grad_norm": 7.138896942138672, | |
| "learning_rate": 3.4200000000000007e-06, | |
| "loss": 0.424, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 0.32948929159802304, | |
| "grad_norm": 6.78637170791626, | |
| "learning_rate": 3.920000000000001e-06, | |
| "loss": 0.3443, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.37067545304777594, | |
| "grad_norm": 6.042959690093994, | |
| "learning_rate": 4.42e-06, | |
| "loss": 0.2769, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 0.41186161449752884, | |
| "grad_norm": 2.8671882152557373, | |
| "learning_rate": 4.92e-06, | |
| "loss": 0.1545, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.41186161449752884, | |
| "eval_loss": 0.18588024377822876, | |
| "eval_runtime": 2810.3027, | |
| "eval_samples_per_second": 1.439, | |
| "eval_steps_per_second": 0.18, | |
| "eval_wer": 30.12351433232347, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.45304777594728174, | |
| "grad_norm": 4.210362911224365, | |
| "learning_rate": 5.420000000000001e-06, | |
| "loss": 0.094, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 0.4942339373970346, | |
| "grad_norm": 3.7346065044403076, | |
| "learning_rate": 5.92e-06, | |
| "loss": 0.0858, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.5354200988467874, | |
| "grad_norm": 2.7393386363983154, | |
| "learning_rate": 6.42e-06, | |
| "loss": 0.0745, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 0.5766062602965404, | |
| "grad_norm": 2.8941423892974854, | |
| "learning_rate": 6.92e-06, | |
| "loss": 0.0631, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.6177924217462932, | |
| "grad_norm": 3.1176912784576416, | |
| "learning_rate": 7.420000000000001e-06, | |
| "loss": 0.0658, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 0.6177924217462932, | |
| "eval_loss": 0.1334521770477295, | |
| "eval_runtime": 2821.0386, | |
| "eval_samples_per_second": 1.434, | |
| "eval_steps_per_second": 0.179, | |
| "eval_wer": 22.91773479375437, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 0.6589785831960461, | |
| "grad_norm": 3.2934091091156006, | |
| "learning_rate": 7.92e-06, | |
| "loss": 0.0552, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.700164744645799, | |
| "grad_norm": 3.6386263370513916, | |
| "learning_rate": 8.42e-06, | |
| "loss": 0.0529, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 0.7413509060955519, | |
| "grad_norm": 1.1384811401367188, | |
| "learning_rate": 8.920000000000001e-06, | |
| "loss": 0.0418, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.7825370675453048, | |
| "grad_norm": 4.086984634399414, | |
| "learning_rate": 9.42e-06, | |
| "loss": 0.0493, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 0.8237232289950577, | |
| "grad_norm": 2.331613540649414, | |
| "learning_rate": 9.920000000000002e-06, | |
| "loss": 0.0378, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.8237232289950577, | |
| "eval_loss": 0.12090734392404556, | |
| "eval_runtime": 2820.3109, | |
| "eval_samples_per_second": 1.434, | |
| "eval_steps_per_second": 0.179, | |
| "eval_wer": 20.77837333954789, | |
| "step": 500 | |
| } | |
| ], | |
| "logging_steps": 25, | |
| "max_steps": 1000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 2, | |
| "save_steps": 250, | |
| "total_flos": 2.30868320256e+18, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |