auto_dialogue_evaluation模型finetune训练报错
Created by: JiaXiao243
在models/PaddleNLP/dialogue_model_toolkit/auto_dialogue_evaluation路径下执行
TASK=human
python -u main.py
--do_train True
--loss_type L2
--use_cuda
--save_path model_files_tmp/${TASK}_finetuned
--init_model model_files/matching_pretrained
--train_path data/label_data/$TASK/train.ids
--val_path data/label_data/$TASK/val.ids
--print_step 1
--save_step 1
--num_scan_data 50
对auto_dialogue_evaluation模型进行finetune报错信息如下: