| { | |
| "best_metric": 0.42670491337776184, | |
| "best_model_checkpoint": "/scratch/skscla001/speech/results/mms-1b-lozgen-combined-model/checkpoint-2900", | |
| "epoch": 13.008130081300813, | |
| "eval_steps": 100, | |
| "global_step": 3200, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.4065040650406504, | |
| "grad_norm": 5.229663372039795, | |
| "learning_rate": 0.00029099999999999997, | |
| "loss": 6.5686, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.4065040650406504, | |
| "eval_loss": 3.0826728343963623, | |
| "eval_runtime": 40.6972, | |
| "eval_samples_per_second": 13.588, | |
| "eval_steps_per_second": 3.415, | |
| "eval_wer": 0.970125974805039, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.8130081300813008, | |
| "grad_norm": 2.938697576522827, | |
| "learning_rate": 0.0002960027472527472, | |
| "loss": 2.6223, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.8130081300813008, | |
| "eval_loss": 2.23785662651062, | |
| "eval_runtime": 40.5049, | |
| "eval_samples_per_second": 13.653, | |
| "eval_steps_per_second": 3.432, | |
| "eval_wer": 0.9112177564487103, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.2195121951219512, | |
| "grad_norm": 2.3466386795043945, | |
| "learning_rate": 0.0002918818681318681, | |
| "loss": 1.4386, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.2195121951219512, | |
| "eval_loss": 0.6909840106964111, | |
| "eval_runtime": 40.4202, | |
| "eval_samples_per_second": 13.681, | |
| "eval_steps_per_second": 3.439, | |
| "eval_wer": 0.7809238152369526, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.6260162601626016, | |
| "grad_norm": 9.536104202270508, | |
| "learning_rate": 0.000287760989010989, | |
| "loss": 0.8073, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.6260162601626016, | |
| "eval_loss": 0.5903483629226685, | |
| "eval_runtime": 40.294, | |
| "eval_samples_per_second": 13.724, | |
| "eval_steps_per_second": 3.45, | |
| "eval_wer": 0.5698860227954409, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.032520325203252, | |
| "grad_norm": 1.1359059810638428, | |
| "learning_rate": 0.00028364010989010986, | |
| "loss": 0.651, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.032520325203252, | |
| "eval_loss": 0.5555368065834045, | |
| "eval_runtime": 40.6146, | |
| "eval_samples_per_second": 13.616, | |
| "eval_steps_per_second": 3.422, | |
| "eval_wer": 0.5036592681463707, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.4390243902439024, | |
| "grad_norm": 1.2806068658828735, | |
| "learning_rate": 0.00027951923076923076, | |
| "loss": 0.655, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 2.4390243902439024, | |
| "eval_loss": 0.529751718044281, | |
| "eval_runtime": 40.5683, | |
| "eval_samples_per_second": 13.631, | |
| "eval_steps_per_second": 3.426, | |
| "eval_wer": 0.4818236352729454, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 2.845528455284553, | |
| "grad_norm": 0.6596723198890686, | |
| "learning_rate": 0.0002753983516483516, | |
| "loss": 0.6579, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.845528455284553, | |
| "eval_loss": 0.5297980308532715, | |
| "eval_runtime": 40.2452, | |
| "eval_samples_per_second": 13.741, | |
| "eval_steps_per_second": 3.454, | |
| "eval_wer": 0.4603479304139172, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 3.252032520325203, | |
| "grad_norm": 1.53997802734375, | |
| "learning_rate": 0.0002712774725274725, | |
| "loss": 0.5699, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 3.252032520325203, | |
| "eval_loss": 0.5159797072410583, | |
| "eval_runtime": 40.3095, | |
| "eval_samples_per_second": 13.719, | |
| "eval_steps_per_second": 3.448, | |
| "eval_wer": 0.4284343131373725, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 3.658536585365854, | |
| "grad_norm": 5.62721061706543, | |
| "learning_rate": 0.00026715659340659334, | |
| "loss": 0.6104, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 3.658536585365854, | |
| "eval_loss": 0.5069619417190552, | |
| "eval_runtime": 40.7288, | |
| "eval_samples_per_second": 13.578, | |
| "eval_steps_per_second": 3.413, | |
| "eval_wer": 0.4320335932813437, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 4.065040650406504, | |
| "grad_norm": 0.9546043276786804, | |
| "learning_rate": 0.00026303571428571424, | |
| "loss": 0.604, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.065040650406504, | |
| "eval_loss": 0.49775609374046326, | |
| "eval_runtime": 40.5226, | |
| "eval_samples_per_second": 13.647, | |
| "eval_steps_per_second": 3.43, | |
| "eval_wer": 0.4098380323935213, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.471544715447155, | |
| "grad_norm": 1.088158130645752, | |
| "learning_rate": 0.00025895604395604396, | |
| "loss": 0.5681, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 4.471544715447155, | |
| "eval_loss": 0.49754026532173157, | |
| "eval_runtime": 40.4172, | |
| "eval_samples_per_second": 13.682, | |
| "eval_steps_per_second": 3.439, | |
| "eval_wer": 0.40719856028794243, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 4.878048780487805, | |
| "grad_norm": 1.1959774494171143, | |
| "learning_rate": 0.0002548351648351648, | |
| "loss": 0.5493, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 4.878048780487805, | |
| "eval_loss": 0.4878113865852356, | |
| "eval_runtime": 40.4064, | |
| "eval_samples_per_second": 13.686, | |
| "eval_steps_per_second": 3.44, | |
| "eval_wer": 0.40383923215356926, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 5.284552845528455, | |
| "grad_norm": 0.7603219747543335, | |
| "learning_rate": 0.0002507142857142857, | |
| "loss": 0.581, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 5.284552845528455, | |
| "eval_loss": 0.48260873556137085, | |
| "eval_runtime": 40.7585, | |
| "eval_samples_per_second": 13.568, | |
| "eval_steps_per_second": 3.41, | |
| "eval_wer": 0.39652069586082783, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 5.691056910569106, | |
| "grad_norm": 1.554537296295166, | |
| "learning_rate": 0.00024659340659340654, | |
| "loss": 0.5746, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 5.691056910569106, | |
| "eval_loss": 0.47932884097099304, | |
| "eval_runtime": 41.051, | |
| "eval_samples_per_second": 13.471, | |
| "eval_steps_per_second": 3.386, | |
| "eval_wer": 0.4242351529694061, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 6.097560975609756, | |
| "grad_norm": 1.0625553131103516, | |
| "learning_rate": 0.00024247252747252747, | |
| "loss": 0.5238, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 6.097560975609756, | |
| "eval_loss": 0.472389817237854, | |
| "eval_runtime": 40.6042, | |
| "eval_samples_per_second": 13.619, | |
| "eval_steps_per_second": 3.423, | |
| "eval_wer": 0.3833233353329334, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 6.504065040650406, | |
| "grad_norm": 1.2422752380371094, | |
| "learning_rate": 0.0002383516483516483, | |
| "loss": 0.5204, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 6.504065040650406, | |
| "eval_loss": 0.48664653301239014, | |
| "eval_runtime": 40.5474, | |
| "eval_samples_per_second": 13.638, | |
| "eval_steps_per_second": 3.428, | |
| "eval_wer": 0.3864427114577085, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 6.9105691056910565, | |
| "grad_norm": 1.324337363243103, | |
| "learning_rate": 0.000234271978021978, | |
| "loss": 0.5563, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 6.9105691056910565, | |
| "eval_loss": 0.46718040108680725, | |
| "eval_runtime": 40.8357, | |
| "eval_samples_per_second": 13.542, | |
| "eval_steps_per_second": 3.404, | |
| "eval_wer": 0.38392321535692864, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 7.317073170731708, | |
| "grad_norm": 1.2208648920059204, | |
| "learning_rate": 0.0002301510989010989, | |
| "loss": 0.5121, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 7.317073170731708, | |
| "eval_loss": 0.466389536857605, | |
| "eval_runtime": 40.8903, | |
| "eval_samples_per_second": 13.524, | |
| "eval_steps_per_second": 3.399, | |
| "eval_wer": 0.3719256148770246, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 7.723577235772358, | |
| "grad_norm": 1.342207908630371, | |
| "learning_rate": 0.00022603021978021977, | |
| "loss": 0.4774, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 7.723577235772358, | |
| "eval_loss": 0.4625222682952881, | |
| "eval_runtime": 40.6308, | |
| "eval_samples_per_second": 13.61, | |
| "eval_steps_per_second": 3.421, | |
| "eval_wer": 0.36520695860827834, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 8.130081300813009, | |
| "grad_norm": 1.1112520694732666, | |
| "learning_rate": 0.00022190934065934064, | |
| "loss": 0.5356, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 8.130081300813009, | |
| "eval_loss": 0.4720795750617981, | |
| "eval_runtime": 40.7009, | |
| "eval_samples_per_second": 13.587, | |
| "eval_steps_per_second": 3.415, | |
| "eval_wer": 0.3692861427714457, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 8.536585365853659, | |
| "grad_norm": 1.3869178295135498, | |
| "learning_rate": 0.0002177884615384615, | |
| "loss": 0.4385, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 8.536585365853659, | |
| "eval_loss": 0.4559965133666992, | |
| "eval_runtime": 40.8418, | |
| "eval_samples_per_second": 13.54, | |
| "eval_steps_per_second": 3.403, | |
| "eval_wer": 0.3695260947810438, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 8.94308943089431, | |
| "grad_norm": 1.1107573509216309, | |
| "learning_rate": 0.0002136675824175824, | |
| "loss": 0.5561, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 8.94308943089431, | |
| "eval_loss": 0.4452870488166809, | |
| "eval_runtime": 40.9093, | |
| "eval_samples_per_second": 13.518, | |
| "eval_steps_per_second": 3.398, | |
| "eval_wer": 0.3594481103779244, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 9.34959349593496, | |
| "grad_norm": 1.0222277641296387, | |
| "learning_rate": 0.00020954670329670325, | |
| "loss": 0.414, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 9.34959349593496, | |
| "eval_loss": 0.44891127943992615, | |
| "eval_runtime": 40.5136, | |
| "eval_samples_per_second": 13.65, | |
| "eval_steps_per_second": 3.431, | |
| "eval_wer": 0.3546490701859628, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 9.75609756097561, | |
| "grad_norm": 1.8858979940414429, | |
| "learning_rate": 0.00020542582417582415, | |
| "loss": 0.4763, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 9.75609756097561, | |
| "eval_loss": 0.4524897038936615, | |
| "eval_runtime": 40.6457, | |
| "eval_samples_per_second": 13.605, | |
| "eval_steps_per_second": 3.42, | |
| "eval_wer": 0.352129574085183, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 10.16260162601626, | |
| "grad_norm": 1.759278655052185, | |
| "learning_rate": 0.00020130494505494502, | |
| "loss": 0.5317, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 10.16260162601626, | |
| "eval_loss": 0.44243165850639343, | |
| "eval_runtime": 40.8637, | |
| "eval_samples_per_second": 13.533, | |
| "eval_steps_per_second": 3.402, | |
| "eval_wer": 0.3557288542291542, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 10.56910569105691, | |
| "grad_norm": 1.8020973205566406, | |
| "learning_rate": 0.00019718406593406592, | |
| "loss": 0.4939, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 10.56910569105691, | |
| "eval_loss": 0.43977972865104675, | |
| "eval_runtime": 40.8341, | |
| "eval_samples_per_second": 13.543, | |
| "eval_steps_per_second": 3.404, | |
| "eval_wer": 0.3502099580083983, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 10.975609756097562, | |
| "grad_norm": 1.453413724899292, | |
| "learning_rate": 0.00019306318681318682, | |
| "loss": 0.4456, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 10.975609756097562, | |
| "eval_loss": 0.4414619505405426, | |
| "eval_runtime": 40.6722, | |
| "eval_samples_per_second": 13.597, | |
| "eval_steps_per_second": 3.418, | |
| "eval_wer": 0.34673065386922614, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 11.382113821138212, | |
| "grad_norm": 1.151430606842041, | |
| "learning_rate": 0.00018898351648351645, | |
| "loss": 0.4583, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 11.382113821138212, | |
| "eval_loss": 0.4501504600048065, | |
| "eval_runtime": 40.6207, | |
| "eval_samples_per_second": 13.614, | |
| "eval_steps_per_second": 3.422, | |
| "eval_wer": 0.34457108578284346, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 11.788617886178862, | |
| "grad_norm": 4.184691429138184, | |
| "learning_rate": 0.00018486263736263735, | |
| "loss": 0.4573, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 11.788617886178862, | |
| "eval_loss": 0.42670491337776184, | |
| "eval_runtime": 40.8415, | |
| "eval_samples_per_second": 13.54, | |
| "eval_steps_per_second": 3.403, | |
| "eval_wer": 0.340251949610078, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 12.195121951219512, | |
| "grad_norm": 16.549129486083984, | |
| "learning_rate": 0.00018074175824175822, | |
| "loss": 0.398, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 12.195121951219512, | |
| "eval_loss": 0.4304564595222473, | |
| "eval_runtime": 40.8815, | |
| "eval_samples_per_second": 13.527, | |
| "eval_steps_per_second": 3.4, | |
| "eval_wer": 0.3406118776244751, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 12.601626016260163, | |
| "grad_norm": 3.6663782596588135, | |
| "learning_rate": 0.00017662087912087912, | |
| "loss": 0.472, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 12.601626016260163, | |
| "eval_loss": 0.42684391140937805, | |
| "eval_runtime": 40.8829, | |
| "eval_samples_per_second": 13.526, | |
| "eval_steps_per_second": 3.4, | |
| "eval_wer": 0.3319736052789442, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 13.008130081300813, | |
| "grad_norm": 0.5571214556694031, | |
| "learning_rate": 0.00017249999999999996, | |
| "loss": 0.3993, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 13.008130081300813, | |
| "eval_loss": 0.42878594994544983, | |
| "eval_runtime": 40.7789, | |
| "eval_samples_per_second": 13.561, | |
| "eval_steps_per_second": 3.409, | |
| "eval_wer": 0.32969406118776245, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 13.008130081300813, | |
| "step": 3200, | |
| "total_flos": 9.286133731448259e+18, | |
| "train_loss": 0.8163932430744171, | |
| "train_runtime": 3897.3821, | |
| "train_samples_per_second": 7.559, | |
| "train_steps_per_second": 1.894 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 7380, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 30, | |
| "save_steps": 400, | |
| "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": 9.286133731448259e+18, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |