easytexminer.core¶

Trainer¶

class easytexminer.core.trainer.Trainer(model, train_dataset, valid_dataset, evaluator=None)[source]¶

property model_module¶

property learning_rate¶

set_model_and_optimizer(model, cfg)[source]¶

resume_from_ckpt(model_module, cfg)[source]¶

set_tensorboard()[source]¶

set_evaluator(evaluator=None, eval_metrics=None)[source]¶

set_data_loader(train_dataset, valid_dataset, cfg)[source]¶

log_train_infos()[source]¶

before_epoch(_epoch)[source]¶

after_epoch()[source]¶

before_iter()[source]¶

optimizer_step()[source]¶

after_iter(_step, _epoch, loss_dict)[source]¶

after_train()[source]¶

save_checkpoint(save_best=False)[source]¶

train()[source]¶

Evaluator¶

class easytexminer.core.evaluator.Evaluator(metrics=('accuracy',))[source]¶

evaluate(model, valid_loader=None, valid_dataset=None, eval_batch_size=32, teacher_model=None)[source]¶

evaluate_text_classify(model, valid_loader)[source]¶

evaluate_multi_label_text_classify(model, valid_loader)[source]¶

evaluate_language_modeling(model, valid_loader)[source]¶

evaluate_sequence_labeling(model, valid_loader)[source]¶

evaluate_none_task(model, teacher_model, valid_loader)[source]¶