[Question]: text_summarization->pegasus->train.py 使用cpu训练报错

[QQR1]

请提出你的问题

Exception in thread Thread-4: Traceback (most recent call last): File "E:\Users\renqi\anaconda3\envs\paddle\lib\threading.py", line 932, in _bootstrap_inner self.run() File "E:\Users\renqi\anaconda3\envs\paddle\lib\threading.py", line 870, in run self._target(*self._args, **self._kwargs) File "E:\Users\renqi\anaconda3\envs\paddle\lib\site-packages\paddle\fluid\dataloader\dataloader_iter.py", line 217, in _thread_loop batch = self._dataset_fetcher.fetch(indices, File "E:\Users\renqi\anaconda3\envs\paddle\lib\site-packages\paddle\fluid\dataloader\fetcher.py", line 134, in fetch data = self.collate_fn(data) File "E:\Users\renqi\anaconda3\envs\paddle\lib\site-packages\paddlenlp\data\data_collator.py", line 383, in __call__ decoder_input_ids = self.model.prepare_decoder_input_ids_from_labels(labels=features["labels"]) File "E:\Users\renqi\anaconda3\envs\paddle\lib\site-packages\paddlenlp\transformers\pegasus\modeling.py", line 673, in prepare_decoder_input_ids_from_labels return shift_tokens_right(labels, self.pegasus.pad_token_id, self.pegasus.config["decoder_start_token_id"]) File "E:\Users\renqi\anaconda3\envs\paddle\lib\site-packages\paddlenlp\transformers\pegasus\modeling.py", line 45, in shift_tokens_right shifted_input_ids = paddle.where(shifted_input_ids == -100, pad_token_id, shifted_input_ids) File "E:\Users\renqi\anaconda3\envs\paddle\lib\site-packages\paddle\tensor\search.py", line 640, in where broadcast_zeros = paddle.add(zeros_like_x, zeros_like_y) File "E:\Users\renqi\anaconda3\envs\paddle\lib\site-packages\paddle\tensor\math.py", line 278, in add return _C_ops.elementwise_add(x, y) ValueError: (InvalidArgument) The type of data we are trying to retrieve does not match the type of data currently contained in the container. [Hint: Expected dtype() == paddle::experimental::CppTypeToDataType<T>::Type(), but received dtype():7 != paddle::experimental::CppTypeToDataType<T>::Type():5.] (at ..\paddle\phi\core\dense_tensor.cc:137) [operator < elementwise_add > error]

[gongel]

你好，收到，我们复现一下

[gongel]

你好，我们测试PaddlePaddle==2.4.1，PaddleNLP==2.4.9没有复现出你的问题，可以跑通代码。可以试试上述的版本，也辛苦提供下你的版本哈。

[xiaofeng-ling]

我这边也出现了这个问题
Windows11 x64
Python 3.7.9
PaddlePaddle==2.4.1
PaddleNLP==2.4.9

使用的是 PaddleNLP-2.4.9的release包中的训练代码，路径PaddleNLP-2.4.9\applications\text_summarization\pegasus\train.py
python.exe train.py --model_name_or_path=IDEA-CCNL/Randeng-Pegasus-238M-Summary-Chinese --train_file data/train.json --eval_file data/test.json --output_dir pegasus_out --max_source_length 128 --max_target_length 64 --epoch 20 --logging_steps 1 --save_steps 10000 --train_batch_size 128 --eval_batch_size 128 --learning_rate 5e-5 --warmup_proportion 0.02 --weight_decay=0.01 --device=cpu
报错内容一致

[gongel]

各位，已修复该问题#4431，感谢反馈