| { | |
| "best_metric": 0.2767798602581024, | |
| "best_model_checkpoint": "/scratch/skscla001/results/mms-zeroshot-300m-genbed-combined-model/checkpoint-5000", | |
| "epoch": 7.6923076923076925, | |
| "eval_steps": 200, | |
| "global_step": 5600, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.27472527472527475, | |
| "eval_loss": 2.203052520751953, | |
| "eval_runtime": 115.6497, | |
| "eval_samples_per_second": 16.766, | |
| "eval_steps_per_second": 2.101, | |
| "eval_wer": 1.0, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.5494505494505495, | |
| "eval_loss": 0.4235461354255676, | |
| "eval_runtime": 114.7031, | |
| "eval_samples_per_second": 16.905, | |
| "eval_steps_per_second": 2.119, | |
| "eval_wer": 0.6042162162162162, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.6868131868131868, | |
| "grad_norm": 1.5520302057266235, | |
| "learning_rate": 0.00029452161913523457, | |
| "loss": 2.7468, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.8241758241758241, | |
| "eval_loss": 0.3791254162788391, | |
| "eval_runtime": 114.91, | |
| "eval_samples_per_second": 16.874, | |
| "eval_steps_per_second": 2.115, | |
| "eval_wer": 0.5721621621621622, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.098901098901099, | |
| "eval_loss": 0.35884976387023926, | |
| "eval_runtime": 115.1243, | |
| "eval_samples_per_second": 16.843, | |
| "eval_steps_per_second": 2.111, | |
| "eval_wer": 0.5631891891891891, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.3736263736263736, | |
| "grad_norm": 0.5800592303276062, | |
| "learning_rate": 0.00028762189512419504, | |
| "loss": 0.5446, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.3736263736263736, | |
| "eval_loss": 0.34893664717674255, | |
| "eval_runtime": 115.1834, | |
| "eval_samples_per_second": 16.834, | |
| "eval_steps_per_second": 2.11, | |
| "eval_wer": 0.5446486486486487, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.6483516483516483, | |
| "eval_loss": 0.34060972929000854, | |
| "eval_runtime": 116.725, | |
| "eval_samples_per_second": 16.612, | |
| "eval_steps_per_second": 2.082, | |
| "eval_wer": 0.5457837837837838, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.9230769230769231, | |
| "eval_loss": 0.33255019783973694, | |
| "eval_runtime": 114.9472, | |
| "eval_samples_per_second": 16.869, | |
| "eval_steps_per_second": 2.114, | |
| "eval_wer": 0.5182702702702703, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 2.0604395604395602, | |
| "grad_norm": 0.7607436776161194, | |
| "learning_rate": 0.00028072217111315545, | |
| "loss": 0.4932, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.197802197802198, | |
| "eval_loss": 0.3234807848930359, | |
| "eval_runtime": 116.1043, | |
| "eval_samples_per_second": 16.701, | |
| "eval_steps_per_second": 2.093, | |
| "eval_wer": 0.5237297297297298, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 2.4725274725274726, | |
| "eval_loss": 0.3202188313007355, | |
| "eval_runtime": 114.8249, | |
| "eval_samples_per_second": 16.887, | |
| "eval_steps_per_second": 2.116, | |
| "eval_wer": 0.5057837837837837, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 2.7472527472527473, | |
| "grad_norm": 0.899978756904602, | |
| "learning_rate": 0.0002738224471021159, | |
| "loss": 0.4644, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.7472527472527473, | |
| "eval_loss": 0.32043078541755676, | |
| "eval_runtime": 114.5219, | |
| "eval_samples_per_second": 16.931, | |
| "eval_steps_per_second": 2.122, | |
| "eval_wer": 0.504, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 3.021978021978022, | |
| "eval_loss": 0.31440961360931396, | |
| "eval_runtime": 115.5011, | |
| "eval_samples_per_second": 16.788, | |
| "eval_steps_per_second": 2.104, | |
| "eval_wer": 0.490972972972973, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 3.2967032967032965, | |
| "eval_loss": 0.30763551592826843, | |
| "eval_runtime": 115.3262, | |
| "eval_samples_per_second": 16.813, | |
| "eval_steps_per_second": 2.107, | |
| "eval_wer": 0.49572972972972973, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 3.4340659340659343, | |
| "grad_norm": 0.6096507906913757, | |
| "learning_rate": 0.00026692272309107633, | |
| "loss": 0.4505, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 3.571428571428571, | |
| "eval_loss": 0.3038763701915741, | |
| "eval_runtime": 114.948, | |
| "eval_samples_per_second": 16.868, | |
| "eval_steps_per_second": 2.114, | |
| "eval_wer": 0.4750810810810811, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "eval_loss": 0.3026913106441498, | |
| "eval_runtime": 115.6051, | |
| "eval_samples_per_second": 16.773, | |
| "eval_steps_per_second": 2.102, | |
| "eval_wer": 0.4785405405405405, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 4.1208791208791204, | |
| "grad_norm": 0.8457227945327759, | |
| "learning_rate": 0.00026002299908003674, | |
| "loss": 0.433, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 4.1208791208791204, | |
| "eval_loss": 0.3013380765914917, | |
| "eval_runtime": 116.3739, | |
| "eval_samples_per_second": 16.662, | |
| "eval_steps_per_second": 2.088, | |
| "eval_wer": 0.46816216216216217, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 4.395604395604396, | |
| "eval_loss": 0.2960599362850189, | |
| "eval_runtime": 115.9661, | |
| "eval_samples_per_second": 16.72, | |
| "eval_steps_per_second": 2.095, | |
| "eval_wer": 0.4765405405405405, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 4.670329670329671, | |
| "eval_loss": 0.29141488671302795, | |
| "eval_runtime": 115.6468, | |
| "eval_samples_per_second": 16.767, | |
| "eval_steps_per_second": 2.101, | |
| "eval_wer": 0.468, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 4.8076923076923075, | |
| "grad_norm": 0.6512793302536011, | |
| "learning_rate": 0.0002531232750689972, | |
| "loss": 0.4203, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 4.945054945054945, | |
| "eval_loss": 0.29088592529296875, | |
| "eval_runtime": 116.2657, | |
| "eval_samples_per_second": 16.677, | |
| "eval_steps_per_second": 2.09, | |
| "eval_wer": 0.46562162162162163, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 5.21978021978022, | |
| "eval_loss": 0.29455265402793884, | |
| "eval_runtime": 116.5267, | |
| "eval_samples_per_second": 16.64, | |
| "eval_steps_per_second": 2.085, | |
| "eval_wer": 0.4508108108108108, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 5.4945054945054945, | |
| "grad_norm": 0.45442482829093933, | |
| "learning_rate": 0.0002462235510579577, | |
| "loss": 0.4042, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 5.4945054945054945, | |
| "eval_loss": 0.2901473343372345, | |
| "eval_runtime": 116.1982, | |
| "eval_samples_per_second": 16.687, | |
| "eval_steps_per_second": 2.091, | |
| "eval_wer": 0.44335135135135134, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 5.769230769230769, | |
| "eval_loss": 0.2899898886680603, | |
| "eval_runtime": 115.6122, | |
| "eval_samples_per_second": 16.772, | |
| "eval_steps_per_second": 2.102, | |
| "eval_wer": 0.44632432432432434, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 6.043956043956044, | |
| "eval_loss": 0.2839508354663849, | |
| "eval_runtime": 115.0777, | |
| "eval_samples_per_second": 16.849, | |
| "eval_steps_per_second": 2.112, | |
| "eval_wer": 0.4448648648648649, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 6.181318681318682, | |
| "grad_norm": 0.6338995695114136, | |
| "learning_rate": 0.0002393238270469181, | |
| "loss": 0.3962, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 6.318681318681318, | |
| "eval_loss": 0.2820078432559967, | |
| "eval_runtime": 114.3206, | |
| "eval_samples_per_second": 16.961, | |
| "eval_steps_per_second": 2.126, | |
| "eval_wer": 0.44475675675675674, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 6.593406593406593, | |
| "eval_loss": 0.27714216709136963, | |
| "eval_runtime": 115.2116, | |
| "eval_samples_per_second": 16.83, | |
| "eval_steps_per_second": 2.109, | |
| "eval_wer": 0.4325945945945946, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 6.868131868131869, | |
| "grad_norm": 0.5752519965171814, | |
| "learning_rate": 0.00023242410303587856, | |
| "loss": 0.3881, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 6.868131868131869, | |
| "eval_loss": 0.2767798602581024, | |
| "eval_runtime": 114.4373, | |
| "eval_samples_per_second": 16.944, | |
| "eval_steps_per_second": 2.123, | |
| "eval_wer": 0.4294054054054054, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 7.142857142857143, | |
| "eval_loss": 0.27870669960975647, | |
| "eval_runtime": 117.3195, | |
| "eval_samples_per_second": 16.528, | |
| "eval_steps_per_second": 2.071, | |
| "eval_wer": 0.43854054054054054, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 7.417582417582418, | |
| "eval_loss": 0.2795184254646301, | |
| "eval_runtime": 116.3581, | |
| "eval_samples_per_second": 16.664, | |
| "eval_steps_per_second": 2.088, | |
| "eval_wer": 0.4309189189189189, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 7.554945054945055, | |
| "grad_norm": 0.30598416924476624, | |
| "learning_rate": 0.000225524379024839, | |
| "loss": 0.3784, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 7.6923076923076925, | |
| "eval_loss": 0.2802658975124359, | |
| "eval_runtime": 115.2901, | |
| "eval_samples_per_second": 16.818, | |
| "eval_steps_per_second": 2.108, | |
| "eval_wer": 0.43567567567567567, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 7.6923076923076925, | |
| "step": 5600, | |
| "total_flos": 8.779281950497536e+18, | |
| "train_loss": 0.6424136958803449, | |
| "train_runtime": 7260.5039, | |
| "train_samples_per_second": 24.036, | |
| "train_steps_per_second": 3.008 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 21840, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 30, | |
| "save_steps": 200, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 3, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 3 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 8.779281950497536e+18, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |