| { | |
| "best_global_step": 296, | |
| "best_metric": 0.9902256964114682, | |
| "best_model_checkpoint": "outputs/modernbert_multitask\\checkpoint-296", | |
| "epoch": 2.0, | |
| "eval_steps": 500, | |
| "global_step": 296, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.1694915254237288, | |
| "grad_norm": 19.44745635986328, | |
| "learning_rate": 1.8260869565217393e-05, | |
| "loss": 2.5678, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.3389830508474576, | |
| "grad_norm": 13.687152862548828, | |
| "learning_rate": 1.8907363420427555e-05, | |
| "loss": 1.1327, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.5084745762711864, | |
| "grad_norm": 6.3025336265563965, | |
| "learning_rate": 1.7719714964370545e-05, | |
| "loss": 0.4353, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.6779661016949152, | |
| "grad_norm": 8.157584190368652, | |
| "learning_rate": 1.653206650831354e-05, | |
| "loss": 0.3109, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.847457627118644, | |
| "grad_norm": 4.141303539276123, | |
| "learning_rate": 1.5344418052256533e-05, | |
| "loss": 0.1846, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_cat_accuracy": 0.990108803165183, | |
| "eval_cat_macro_f1": 0.990110753164417, | |
| "eval_cat_macro_precision": 0.9901580517709222, | |
| "eval_cat_macro_recall": 0.9901043700333245, | |
| "eval_cat_weighted_f1": 0.990113014130802, | |
| "eval_imp_accuracy": 0.9678536102868447, | |
| "eval_imp_f1": 0.9258836944127709, | |
| "eval_imp_precision": 0.8787878787878788, | |
| "eval_imp_recall": 0.9783132530120482, | |
| "eval_imp_roc_auc": 0.9973991797932239, | |
| "eval_loss": 0.11633610725402832, | |
| "eval_overall_score": 0.9579972237885939, | |
| "eval_runtime": 18.6093, | |
| "eval_samples_per_second": 108.656, | |
| "eval_steps_per_second": 13.595, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 1.0135593220338983, | |
| "grad_norm": 10.889383316040039, | |
| "learning_rate": 1.4156769596199527e-05, | |
| "loss": 0.1544, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.1830508474576271, | |
| "grad_norm": 8.103240013122559, | |
| "learning_rate": 1.296912114014252e-05, | |
| "loss": 0.0556, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 1.352542372881356, | |
| "grad_norm": 6.277998447418213, | |
| "learning_rate": 1.1781472684085512e-05, | |
| "loss": 0.0573, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.5220338983050847, | |
| "grad_norm": 0.4849356710910797, | |
| "learning_rate": 1.0593824228028504e-05, | |
| "loss": 0.051, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 1.6915254237288135, | |
| "grad_norm": 7.279757022857666, | |
| "learning_rate": 9.406175771971498e-06, | |
| "loss": 0.0636, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 1.8610169491525423, | |
| "grad_norm": 0.6995746493339539, | |
| "learning_rate": 8.21852731591449e-06, | |
| "loss": 0.0383, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_cat_accuracy": 0.9925816023738873, | |
| "eval_cat_macro_f1": 0.9925873151530334, | |
| "eval_cat_macro_precision": 0.9926163922277681, | |
| "eval_cat_macro_recall": 0.9925801130506106, | |
| "eval_cat_weighted_f1": 0.9925866323108367, | |
| "eval_imp_accuracy": 0.9950544015825915, | |
| "eval_imp_f1": 0.9878640776699029, | |
| "eval_imp_precision": 0.9951100244498777, | |
| "eval_imp_recall": 0.980722891566265, | |
| "eval_imp_roc_auc": 0.9997166013150299, | |
| "eval_loss": 0.0532318614423275, | |
| "eval_overall_score": 0.9902256964114682, | |
| "eval_runtime": 18.9313, | |
| "eval_samples_per_second": 106.807, | |
| "eval_steps_per_second": 13.364, | |
| "step": 296 | |
| } | |
| ], | |
| "logging_steps": 25, | |
| "max_steps": 444, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |