| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 14.0, | |
| "eval_steps": 500, | |
| "global_step": 10584, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.6613756613756614, | |
| "grad_norm": 4.001740455627441, | |
| "learning_rate": 0.00014879999999999998, | |
| "loss": 5.2541, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.6613756613756614, | |
| "eval_cer": 0.7718660765772963, | |
| "eval_loss": 2.2404534816741943, | |
| "eval_runtime": 101.1138, | |
| "eval_samples_per_second": 29.897, | |
| "eval_steps_per_second": 3.738, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.3227513227513228, | |
| "grad_norm": 2.757023334503174, | |
| "learning_rate": 0.0002988, | |
| "loss": 0.9761, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.3227513227513228, | |
| "eval_cer": 0.15263351698359653, | |
| "eval_loss": 0.45504969358444214, | |
| "eval_runtime": 101.0814, | |
| "eval_samples_per_second": 29.907, | |
| "eval_steps_per_second": 3.74, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.9841269841269842, | |
| "grad_norm": 3.515296697616577, | |
| "learning_rate": 0.000284474123539232, | |
| "loss": 0.6318, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.9841269841269842, | |
| "eval_cer": 0.14255243759249095, | |
| "eval_loss": 0.37101900577545166, | |
| "eval_runtime": 100.1575, | |
| "eval_samples_per_second": 30.182, | |
| "eval_steps_per_second": 3.774, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.6455026455026456, | |
| "grad_norm": 3.0831806659698486, | |
| "learning_rate": 0.00026882303839732884, | |
| "loss": 0.4988, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.6455026455026456, | |
| "eval_cer": 0.13127635774978635, | |
| "eval_loss": 0.33260029554367065, | |
| "eval_runtime": 101.862, | |
| "eval_samples_per_second": 29.677, | |
| "eval_steps_per_second": 3.711, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 3.306878306878307, | |
| "grad_norm": 2.1343257427215576, | |
| "learning_rate": 0.00025317195325542566, | |
| "loss": 0.4457, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 3.306878306878307, | |
| "eval_cer": 0.12462048314146165, | |
| "eval_loss": 0.3164725601673126, | |
| "eval_runtime": 100.3857, | |
| "eval_samples_per_second": 30.114, | |
| "eval_steps_per_second": 3.765, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 3.9682539682539684, | |
| "grad_norm": 2.458252191543579, | |
| "learning_rate": 0.00023752086811352254, | |
| "loss": 0.4001, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 3.9682539682539684, | |
| "eval_cer": 0.11981963830393308, | |
| "eval_loss": 0.29567044973373413, | |
| "eval_runtime": 100.366, | |
| "eval_samples_per_second": 30.12, | |
| "eval_steps_per_second": 3.766, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 4.62962962962963, | |
| "grad_norm": 2.29879093170166, | |
| "learning_rate": 0.00022186978297161937, | |
| "loss": 0.3481, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 4.62962962962963, | |
| "eval_cer": 0.11847873663440629, | |
| "eval_loss": 0.2806848883628845, | |
| "eval_runtime": 100.2612, | |
| "eval_samples_per_second": 30.151, | |
| "eval_steps_per_second": 3.77, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 5.291005291005291, | |
| "grad_norm": 1.4758025407791138, | |
| "learning_rate": 0.0002062186978297162, | |
| "loss": 0.3299, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 5.291005291005291, | |
| "eval_cer": 0.1159011484510154, | |
| "eval_loss": 0.27692675590515137, | |
| "eval_runtime": 100.2812, | |
| "eval_samples_per_second": 30.145, | |
| "eval_steps_per_second": 3.769, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 5.9523809523809526, | |
| "grad_norm": 1.3598448038101196, | |
| "learning_rate": 0.00019056761268781302, | |
| "loss": 0.3024, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 5.9523809523809526, | |
| "eval_cer": 0.11369178715096608, | |
| "eval_loss": 0.26659879088401794, | |
| "eval_runtime": 100.6471, | |
| "eval_samples_per_second": 30.036, | |
| "eval_steps_per_second": 3.756, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 6.613756613756614, | |
| "grad_norm": 1.0627385377883911, | |
| "learning_rate": 0.00017491652754590985, | |
| "loss": 0.273, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 6.613756613756614, | |
| "eval_cer": 0.11135041998707731, | |
| "eval_loss": 0.2516258955001831, | |
| "eval_runtime": 100.6836, | |
| "eval_samples_per_second": 30.025, | |
| "eval_steps_per_second": 3.754, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 7.275132275132275, | |
| "grad_norm": 2.335207223892212, | |
| "learning_rate": 0.00015926544240400667, | |
| "loss": 0.2573, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 7.275132275132275, | |
| "eval_cer": 0.11219108890942313, | |
| "eval_loss": 0.26705658435821533, | |
| "eval_runtime": 100.0417, | |
| "eval_samples_per_second": 30.217, | |
| "eval_steps_per_second": 3.778, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 7.936507936507937, | |
| "grad_norm": 2.2953412532806396, | |
| "learning_rate": 0.0001436143572621035, | |
| "loss": 0.2373, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 7.936507936507937, | |
| "eval_cer": 0.11012068115025742, | |
| "eval_loss": 0.26839447021484375, | |
| "eval_runtime": 101.1747, | |
| "eval_samples_per_second": 29.879, | |
| "eval_steps_per_second": 3.736, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 8.597883597883598, | |
| "grad_norm": 1.1853502988815308, | |
| "learning_rate": 0.00012796327212020032, | |
| "loss": 0.2196, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 8.597883597883598, | |
| "eval_cer": 0.10729992427032022, | |
| "eval_loss": 0.2642601728439331, | |
| "eval_runtime": 100.6315, | |
| "eval_samples_per_second": 30.04, | |
| "eval_steps_per_second": 3.756, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 9.25925925925926, | |
| "grad_norm": 4.087908744812012, | |
| "learning_rate": 0.00011234348914858095, | |
| "loss": 0.2033, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 9.25925925925926, | |
| "eval_cer": 0.10725129053101096, | |
| "eval_loss": 0.27656930685043335, | |
| "eval_runtime": 100.8711, | |
| "eval_samples_per_second": 29.969, | |
| "eval_steps_per_second": 3.747, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 9.920634920634921, | |
| "grad_norm": 2.103665351867676, | |
| "learning_rate": 9.669240400667779e-05, | |
| "loss": 0.1958, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 9.920634920634921, | |
| "eval_cer": 0.10560469107154023, | |
| "eval_loss": 0.2580866515636444, | |
| "eval_runtime": 100.9128, | |
| "eval_samples_per_second": 29.957, | |
| "eval_steps_per_second": 3.746, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 10.582010582010582, | |
| "grad_norm": 4.133511066436768, | |
| "learning_rate": 8.104131886477462e-05, | |
| "loss": 0.1758, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 10.582010582010582, | |
| "eval_cer": 0.10472233608692934, | |
| "eval_loss": 0.2675287127494812, | |
| "eval_runtime": 100.8159, | |
| "eval_samples_per_second": 29.985, | |
| "eval_steps_per_second": 3.749, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 11.243386243386244, | |
| "grad_norm": 1.4247872829437256, | |
| "learning_rate": 6.539023372287144e-05, | |
| "loss": 0.1671, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 11.243386243386244, | |
| "eval_cer": 0.1042429463708809, | |
| "eval_loss": 0.2661576569080353, | |
| "eval_runtime": 100.6921, | |
| "eval_samples_per_second": 30.022, | |
| "eval_steps_per_second": 3.754, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 11.904761904761905, | |
| "grad_norm": 1.5974005460739136, | |
| "learning_rate": 4.9739148580968275e-05, | |
| "loss": 0.1589, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 11.904761904761905, | |
| "eval_cer": 0.10398588232024622, | |
| "eval_loss": 0.26505181193351746, | |
| "eval_runtime": 100.6474, | |
| "eval_samples_per_second": 30.036, | |
| "eval_steps_per_second": 3.756, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 12.566137566137566, | |
| "grad_norm": 0.8991349935531616, | |
| "learning_rate": 3.408806343906511e-05, | |
| "loss": 0.1433, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 12.566137566137566, | |
| "eval_cer": 0.10245044569348238, | |
| "eval_loss": 0.2700433135032654, | |
| "eval_runtime": 101.5429, | |
| "eval_samples_per_second": 29.771, | |
| "eval_steps_per_second": 3.723, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 13.227513227513228, | |
| "grad_norm": 1.4518108367919922, | |
| "learning_rate": 1.8436978297161933e-05, | |
| "loss": 0.1394, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 13.227513227513228, | |
| "eval_cer": 0.10264498065071943, | |
| "eval_loss": 0.27115562558174133, | |
| "eval_runtime": 108.9106, | |
| "eval_samples_per_second": 27.757, | |
| "eval_steps_per_second": 3.471, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 13.88888888888889, | |
| "grad_norm": 1.4964288473129272, | |
| "learning_rate": 2.817195325542571e-06, | |
| "loss": 0.1318, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 13.88888888888889, | |
| "eval_cer": 0.10274919580638213, | |
| "eval_loss": 0.2701418697834015, | |
| "eval_runtime": 100.2825, | |
| "eval_samples_per_second": 30.145, | |
| "eval_steps_per_second": 3.769, | |
| "step": 10500 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 10584, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 14, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 2.0915180948213604e+19, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |