| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 11.680164252309797, | |
| "eval_steps": 300, | |
| "global_step": 9600, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.37, | |
| "learning_rate": 5.96e-05, | |
| "loss": 14.3929, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "eval_cer": 0.9869306169517587, | |
| "eval_loss": 7.0630784034729, | |
| "eval_runtime": 45.4918, | |
| "eval_samples_per_second": 48.822, | |
| "eval_steps_per_second": 6.111, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 9.895209580838324e-05, | |
| "loss": 8.6544, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "eval_cer": 0.8301460695752451, | |
| "eval_loss": 5.885794639587402, | |
| "eval_runtime": 39.1068, | |
| "eval_samples_per_second": 56.793, | |
| "eval_steps_per_second": 7.109, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "learning_rate": 9.575491873396066e-05, | |
| "loss": 4.2963, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "eval_cer": 0.5212857966557756, | |
| "eval_loss": 2.0850701332092285, | |
| "eval_runtime": 39.3849, | |
| "eval_samples_per_second": 56.392, | |
| "eval_steps_per_second": 7.059, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.46, | |
| "learning_rate": 9.256843455945254e-05, | |
| "loss": 2.416, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.46, | |
| "eval_cer": 0.42461080146069574, | |
| "eval_loss": 1.6146297454833984, | |
| "eval_runtime": 39.2918, | |
| "eval_samples_per_second": 56.526, | |
| "eval_steps_per_second": 7.075, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.83, | |
| "learning_rate": 8.936056458511549e-05, | |
| "loss": 2.0879, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.83, | |
| "eval_cer": 0.38155871612531234, | |
| "eval_loss": 1.3970342874526978, | |
| "eval_runtime": 39.3238, | |
| "eval_samples_per_second": 56.48, | |
| "eval_steps_per_second": 7.07, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.19, | |
| "learning_rate": 8.615269461077845e-05, | |
| "loss": 1.8739, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 2.19, | |
| "eval_cer": 0.3361522198731501, | |
| "eval_loss": 1.259849190711975, | |
| "eval_runtime": 39.8822, | |
| "eval_samples_per_second": 55.689, | |
| "eval_steps_per_second": 6.971, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 2.56, | |
| "learning_rate": 8.29448246364414e-05, | |
| "loss": 1.7464, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 2.56, | |
| "eval_cer": 0.3469632904093792, | |
| "eval_loss": 1.1771619319915771, | |
| "eval_runtime": 39.4164, | |
| "eval_samples_per_second": 56.347, | |
| "eval_steps_per_second": 7.053, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 2.92, | |
| "learning_rate": 7.973695466210437e-05, | |
| "loss": 1.7025, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 2.92, | |
| "eval_cer": 0.31635594849125503, | |
| "eval_loss": 1.1029127836227417, | |
| "eval_runtime": 39.436, | |
| "eval_samples_per_second": 56.319, | |
| "eval_steps_per_second": 7.049, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 3.29, | |
| "learning_rate": 7.652908468776732e-05, | |
| "loss": 1.5982, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 3.29, | |
| "eval_cer": 0.3095329617528349, | |
| "eval_loss": 1.049662709236145, | |
| "eval_runtime": 39.3386, | |
| "eval_samples_per_second": 56.459, | |
| "eval_steps_per_second": 7.067, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 3.65, | |
| "learning_rate": 7.332121471343029e-05, | |
| "loss": 1.5632, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 3.65, | |
| "eval_cer": 0.30030751489525276, | |
| "eval_loss": 1.013907790184021, | |
| "eval_runtime": 39.4665, | |
| "eval_samples_per_second": 56.276, | |
| "eval_steps_per_second": 7.044, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 4.02, | |
| "learning_rate": 7.011334473909324e-05, | |
| "loss": 1.5063, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 4.02, | |
| "eval_cer": 0.28858350951374206, | |
| "eval_loss": 0.9488086104393005, | |
| "eval_runtime": 39.4805, | |
| "eval_samples_per_second": 56.256, | |
| "eval_steps_per_second": 7.041, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 4.38, | |
| "learning_rate": 6.69054747647562e-05, | |
| "loss": 1.4507, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 4.38, | |
| "eval_cer": 0.272583125120123, | |
| "eval_loss": 0.9192214012145996, | |
| "eval_runtime": 39.3363, | |
| "eval_samples_per_second": 56.462, | |
| "eval_steps_per_second": 7.067, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 4.75, | |
| "learning_rate": 6.369760479041916e-05, | |
| "loss": 1.4029, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 4.75, | |
| "eval_cer": 0.2764751105131655, | |
| "eval_loss": 0.9218717217445374, | |
| "eval_runtime": 39.3534, | |
| "eval_samples_per_second": 56.437, | |
| "eval_steps_per_second": 7.064, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 5.11, | |
| "learning_rate": 6.050042771599658e-05, | |
| "loss": 1.3126, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 5.11, | |
| "eval_cer": 0.27450509321545263, | |
| "eval_loss": 0.9048557281494141, | |
| "eval_runtime": 39.4768, | |
| "eval_samples_per_second": 56.261, | |
| "eval_steps_per_second": 7.042, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 5.48, | |
| "learning_rate": 5.729255774165954e-05, | |
| "loss": 1.3053, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 5.48, | |
| "eval_cer": 0.2652315971554872, | |
| "eval_loss": 0.8531870245933533, | |
| "eval_runtime": 39.4442, | |
| "eval_samples_per_second": 56.307, | |
| "eval_steps_per_second": 7.048, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 5.84, | |
| "learning_rate": 5.40846877673225e-05, | |
| "loss": 1.2989, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 5.84, | |
| "eval_cer": 0.24903901595233519, | |
| "eval_loss": 0.8212350010871887, | |
| "eval_runtime": 39.4345, | |
| "eval_samples_per_second": 56.321, | |
| "eval_steps_per_second": 7.05, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 6.21, | |
| "learning_rate": 5.087681779298546e-05, | |
| "loss": 1.2403, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 6.21, | |
| "eval_cer": 0.26124351335767826, | |
| "eval_loss": 0.8196715116500854, | |
| "eval_runtime": 39.3565, | |
| "eval_samples_per_second": 56.433, | |
| "eval_steps_per_second": 7.064, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 6.57, | |
| "learning_rate": 4.766894781864842e-05, | |
| "loss": 1.1903, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 6.57, | |
| "eval_cer": 0.2607630213338459, | |
| "eval_loss": 0.8173399567604065, | |
| "eval_runtime": 39.5209, | |
| "eval_samples_per_second": 56.198, | |
| "eval_steps_per_second": 7.034, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 6.94, | |
| "learning_rate": 4.446107784431138e-05, | |
| "loss": 1.2313, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 6.94, | |
| "eval_cer": 0.24995195079761676, | |
| "eval_loss": 0.8241677284240723, | |
| "eval_runtime": 39.4127, | |
| "eval_samples_per_second": 56.352, | |
| "eval_steps_per_second": 7.054, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 7.3, | |
| "learning_rate": 4.125320786997434e-05, | |
| "loss": 1.1554, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 7.3, | |
| "eval_cer": 0.24529117816644244, | |
| "eval_loss": 0.7795117497444153, | |
| "eval_runtime": 39.3516, | |
| "eval_samples_per_second": 56.44, | |
| "eval_steps_per_second": 7.065, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 7.67, | |
| "learning_rate": 3.80453378956373e-05, | |
| "loss": 1.1243, | |
| "step": 6300 | |
| }, | |
| { | |
| "epoch": 7.67, | |
| "eval_cer": 0.2526427061310782, | |
| "eval_loss": 0.782616913318634, | |
| "eval_runtime": 39.456, | |
| "eval_samples_per_second": 56.291, | |
| "eval_steps_per_second": 7.046, | |
| "step": 6300 | |
| }, | |
| { | |
| "epoch": 8.03, | |
| "learning_rate": 3.483746792130026e-05, | |
| "loss": 1.099, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 8.03, | |
| "eval_cer": 0.23010763021333847, | |
| "eval_loss": 0.7462431192398071, | |
| "eval_runtime": 39.3849, | |
| "eval_samples_per_second": 56.392, | |
| "eval_steps_per_second": 7.059, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 8.4, | |
| "learning_rate": 3.1629597946963216e-05, | |
| "loss": 1.0777, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 8.4, | |
| "eval_cer": 0.22544685758216412, | |
| "eval_loss": 0.7633857131004333, | |
| "eval_runtime": 39.2653, | |
| "eval_samples_per_second": 56.564, | |
| "eval_steps_per_second": 7.08, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 8.76, | |
| "learning_rate": 2.8421727972626178e-05, | |
| "loss": 1.0901, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 8.76, | |
| "eval_cer": 0.23880453584470498, | |
| "eval_loss": 0.7462579607963562, | |
| "eval_runtime": 39.2693, | |
| "eval_samples_per_second": 56.558, | |
| "eval_steps_per_second": 7.079, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 9.13, | |
| "learning_rate": 2.5224550898203592e-05, | |
| "loss": 1.0049, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 9.13, | |
| "eval_cer": 0.22155487218912165, | |
| "eval_loss": 0.7342504858970642, | |
| "eval_runtime": 39.3738, | |
| "eval_samples_per_second": 56.408, | |
| "eval_steps_per_second": 7.061, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 9.49, | |
| "learning_rate": 2.2016680923866555e-05, | |
| "loss": 1.0011, | |
| "step": 7800 | |
| }, | |
| { | |
| "epoch": 9.49, | |
| "eval_cer": 0.22674418604651161, | |
| "eval_loss": 0.7101256251335144, | |
| "eval_runtime": 39.3531, | |
| "eval_samples_per_second": 56.438, | |
| "eval_steps_per_second": 7.064, | |
| "step": 7800 | |
| }, | |
| { | |
| "epoch": 9.86, | |
| "learning_rate": 1.8808810949529513e-05, | |
| "loss": 1.0084, | |
| "step": 8100 | |
| }, | |
| { | |
| "epoch": 9.86, | |
| "eval_cer": 0.22189121660580435, | |
| "eval_loss": 0.698137640953064, | |
| "eval_runtime": 39.3187, | |
| "eval_samples_per_second": 56.487, | |
| "eval_steps_per_second": 7.07, | |
| "step": 8100 | |
| }, | |
| { | |
| "epoch": 10.22, | |
| "learning_rate": 1.5600940975192476e-05, | |
| "loss": 0.9547, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 10.22, | |
| "eval_cer": 0.22222756102248703, | |
| "eval_loss": 0.7049764394760132, | |
| "eval_runtime": 39.2625, | |
| "eval_samples_per_second": 56.568, | |
| "eval_steps_per_second": 7.081, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 10.59, | |
| "learning_rate": 1.2393071000855433e-05, | |
| "loss": 0.9304, | |
| "step": 8700 | |
| }, | |
| { | |
| "epoch": 10.59, | |
| "eval_cer": 0.2266961368441284, | |
| "eval_loss": 0.6863571405410767, | |
| "eval_runtime": 39.4555, | |
| "eval_samples_per_second": 56.291, | |
| "eval_steps_per_second": 7.046, | |
| "step": 8700 | |
| }, | |
| { | |
| "epoch": 10.95, | |
| "learning_rate": 9.185201026518392e-06, | |
| "loss": 0.9044, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 10.95, | |
| "eval_cer": 0.2206419373438401, | |
| "eval_loss": 0.6961002945899963, | |
| "eval_runtime": 39.4203, | |
| "eval_samples_per_second": 56.341, | |
| "eval_steps_per_second": 7.052, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 11.32, | |
| "learning_rate": 5.977331052181352e-06, | |
| "loss": 0.9054, | |
| "step": 9300 | |
| }, | |
| { | |
| "epoch": 11.32, | |
| "eval_cer": 0.21146453968864118, | |
| "eval_loss": 0.6892764568328857, | |
| "eval_runtime": 39.4357, | |
| "eval_samples_per_second": 56.32, | |
| "eval_steps_per_second": 7.049, | |
| "step": 9300 | |
| }, | |
| { | |
| "epoch": 11.68, | |
| "learning_rate": 2.7694610778443115e-06, | |
| "loss": 0.9067, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 11.68, | |
| "eval_cer": 0.21718239477224677, | |
| "eval_loss": 0.682004451751709, | |
| "eval_runtime": 39.4317, | |
| "eval_samples_per_second": 56.325, | |
| "eval_steps_per_second": 7.05, | |
| "step": 9600 | |
| } | |
| ], | |
| "logging_steps": 300, | |
| "max_steps": 9852, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 12, | |
| "save_steps": 600, | |
| "total_flos": 1.5092741663533305e+20, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |