{ "best_metric": 0.6752873563218391, "best_model_checkpoint": "./whisper-base-dora-finetune-final\\checkpoint-50", "epoch": 7.6923076923076925, "eval_steps": 5, "global_step": 100, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.38461538461538464, "grad_norm": 2.0491838455200195, "learning_rate": 9.615384615384617e-05, "loss": 3.7474, "step": 5 }, { "epoch": 0.38461538461538464, "eval_loss": 3.6077585220336914, "eval_runtime": 42.0765, "eval_samples_per_second": 1.236, "eval_steps_per_second": 0.618, "eval_wer": 0.8017241379310345, "step": 5 }, { "epoch": 0.7692307692307693, "grad_norm": 1.6141822338104248, "learning_rate": 9.230769230769232e-05, "loss": 3.6035, "step": 10 }, { "epoch": 0.7692307692307693, "eval_loss": 3.5107133388519287, "eval_runtime": 45.2422, "eval_samples_per_second": 1.149, "eval_steps_per_second": 0.575, "eval_wer": 0.7902298850574713, "step": 10 }, { "epoch": 1.1538461538461537, "grad_norm": 2.250202178955078, "learning_rate": 8.846153846153847e-05, "loss": 3.5629, "step": 15 }, { "epoch": 1.1538461538461537, "eval_loss": 3.403240442276001, "eval_runtime": 49.2399, "eval_samples_per_second": 1.056, "eval_steps_per_second": 0.528, "eval_wer": 0.7787356321839081, "step": 15 }, { "epoch": 1.5384615384615383, "grad_norm": 2.4907565116882324, "learning_rate": 8.461538461538461e-05, "loss": 3.2546, "step": 20 }, { "epoch": 1.5384615384615383, "eval_loss": 3.289149045944214, "eval_runtime": 52.6686, "eval_samples_per_second": 0.987, "eval_steps_per_second": 0.494, "eval_wer": 0.7758620689655172, "step": 20 }, { "epoch": 1.9230769230769231, "grad_norm": 2.132004737854004, "learning_rate": 8.076923076923078e-05, "loss": 3.32, "step": 25 }, { "epoch": 1.9230769230769231, "eval_loss": 3.167633056640625, "eval_runtime": 76.6327, "eval_samples_per_second": 0.679, "eval_steps_per_second": 0.339, "eval_wer": 0.75, "step": 25 }, { "epoch": 2.3076923076923075, "grad_norm": 2.6776862144470215, "learning_rate": 7.692307692307693e-05, "loss": 3.1347, "step": 30 }, { "epoch": 2.3076923076923075, "eval_loss": 3.04315447807312, "eval_runtime": 57.0686, "eval_samples_per_second": 0.911, "eval_steps_per_second": 0.456, "eval_wer": 0.7442528735632183, "step": 30 }, { "epoch": 2.6923076923076925, "grad_norm": 2.3373188972473145, "learning_rate": 7.307692307692307e-05, "loss": 3.1053, "step": 35 }, { "epoch": 2.6923076923076925, "eval_loss": 2.9155848026275635, "eval_runtime": 61.0227, "eval_samples_per_second": 0.852, "eval_steps_per_second": 0.426, "eval_wer": 0.7413793103448276, "step": 35 }, { "epoch": 3.076923076923077, "grad_norm": 2.704690456390381, "learning_rate": 6.923076923076924e-05, "loss": 2.909, "step": 40 }, { "epoch": 3.076923076923077, "eval_loss": 2.783993721008301, "eval_runtime": 64.1288, "eval_samples_per_second": 0.811, "eval_steps_per_second": 0.405, "eval_wer": 0.7097701149425287, "step": 40 }, { "epoch": 3.4615384615384617, "grad_norm": 3.074899196624756, "learning_rate": 6.538461538461539e-05, "loss": 2.7333, "step": 45 }, { "epoch": 3.4615384615384617, "eval_loss": 2.6482086181640625, "eval_runtime": 53.0884, "eval_samples_per_second": 0.979, "eval_steps_per_second": 0.49, "eval_wer": 0.6839080459770115, "step": 45 }, { "epoch": 3.8461538461538463, "grad_norm": 3.329392671585083, "learning_rate": 6.153846153846155e-05, "loss": 2.533, "step": 50 }, { "epoch": 3.8461538461538463, "eval_loss": 2.503553628921509, "eval_runtime": 61.5253, "eval_samples_per_second": 0.845, "eval_steps_per_second": 0.423, "eval_wer": 0.6752873563218391, "step": 50 }, { "epoch": 4.230769230769231, "grad_norm": 3.4477381706237793, "learning_rate": 5.769230769230769e-05, "loss": 2.5396, "step": 55 }, { "epoch": 4.230769230769231, "eval_loss": 2.3452019691467285, "eval_runtime": 53.9276, "eval_samples_per_second": 0.964, "eval_steps_per_second": 0.482, "eval_wer": 0.6551724137931034, "step": 55 }, { "epoch": 4.615384615384615, "grad_norm": 3.381711721420288, "learning_rate": 5.384615384615385e-05, "loss": 2.157, "step": 60 }, { "epoch": 4.615384615384615, "eval_loss": 2.1744039058685303, "eval_runtime": 56.4473, "eval_samples_per_second": 0.921, "eval_steps_per_second": 0.461, "eval_wer": 0.6264367816091954, "step": 60 }, { "epoch": 5.0, "grad_norm": 4.3120527267456055, "learning_rate": 5e-05, "loss": 2.1894, "step": 65 }, { "epoch": 5.0, "eval_loss": 1.9950294494628906, "eval_runtime": 54.651, "eval_samples_per_second": 0.951, "eval_steps_per_second": 0.476, "eval_wer": 0.6063218390804598, "step": 65 }, { "epoch": 5.384615384615385, "grad_norm": 4.320505142211914, "learning_rate": 4.615384615384616e-05, "loss": 2.055, "step": 70 }, { "epoch": 5.384615384615385, "eval_loss": 1.824130892753601, "eval_runtime": 54.4674, "eval_samples_per_second": 0.955, "eval_steps_per_second": 0.477, "eval_wer": 0.5862068965517241, "step": 70 }, { "epoch": 5.769230769230769, "grad_norm": 3.6472277641296387, "learning_rate": 4.230769230769231e-05, "loss": 1.6534, "step": 75 }, { "epoch": 5.769230769230769, "eval_loss": 1.6903691291809082, "eval_runtime": 54.4877, "eval_samples_per_second": 0.954, "eval_steps_per_second": 0.477, "eval_wer": 0.5833333333333334, "step": 75 }, { "epoch": 6.153846153846154, "grad_norm": 3.6005494594573975, "learning_rate": 3.846153846153846e-05, "loss": 1.6471, "step": 80 }, { "epoch": 6.153846153846154, "eval_loss": 1.5975652933120728, "eval_runtime": 55.1685, "eval_samples_per_second": 0.943, "eval_steps_per_second": 0.471, "eval_wer": 0.5775862068965517, "step": 80 }, { "epoch": 6.538461538461538, "grad_norm": 3.008185625076294, "learning_rate": 3.461538461538462e-05, "loss": 1.4667, "step": 85 }, { "epoch": 6.538461538461538, "eval_loss": 1.5278733968734741, "eval_runtime": 54.8539, "eval_samples_per_second": 0.948, "eval_steps_per_second": 0.474, "eval_wer": 0.5804597701149425, "step": 85 }, { "epoch": 6.923076923076923, "grad_norm": 3.152780055999756, "learning_rate": 3.0769230769230774e-05, "loss": 1.4498, "step": 90 }, { "epoch": 6.923076923076923, "eval_loss": 1.4769728183746338, "eval_runtime": 53.6893, "eval_samples_per_second": 0.969, "eval_steps_per_second": 0.484, "eval_wer": 0.5890804597701149, "step": 90 }, { "epoch": 7.3076923076923075, "grad_norm": 2.687987804412842, "learning_rate": 2.6923076923076923e-05, "loss": 1.4869, "step": 95 }, { "epoch": 7.3076923076923075, "eval_loss": 1.4400862455368042, "eval_runtime": 55.1748, "eval_samples_per_second": 0.942, "eval_steps_per_second": 0.471, "eval_wer": 0.5977011494252874, "step": 95 }, { "epoch": 7.6923076923076925, "grad_norm": 2.5183703899383545, "learning_rate": 2.307692307692308e-05, "loss": 1.4399, "step": 100 }, { "epoch": 7.6923076923076925, "eval_loss": 1.414151906967163, "eval_runtime": 58.6938, "eval_samples_per_second": 0.886, "eval_steps_per_second": 0.443, "eval_wer": 0.5775862068965517, "step": 100 } ], "logging_steps": 5, "max_steps": 130, "num_input_tokens_seen": 0, "num_train_epochs": 10, "save_steps": 50, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": false }, "attributes": {} } }, "total_flos": 1.0399847841792e+17, "train_batch_size": 2, "trial_name": null, "trial_params": null }