WANG Ning

当TrainingArguments中的report_to参数不指定时，将会报错

3

 文档中https://paddlenlp.readthedocs.io/zh/latest/trainer.html?highlight=TrainingArguments#trainingarguments 说明可以不指定，应该是哪里赋值出了问题

函数paddlenlp.data.data_collator.DataCollatorWithPadding 与描述不符

1

描述中是pad为batch中最长，实际上最长长度为max_model_length。 ``` @dataclass class DataCollatorWithPadding: """ Data collator that will dynamically pad the inputs to the longest sequence in the batch. Args: tokenizer (`paddlenlp.transformers.PretrainedTokenizer`): The tokenizer used for encoding the...

I havent found an example of LCD or other sensors, what peripherals can JArduino support ? And further, what kind of boards can it support?

Hello! Can you provide masking code about motif & sub-sequence strategy? We want to apply it on rna pretraining while there are not too many details about these two strategy...

ss_label dimension

2

Hi, I am wandering why ss_label is a 600x3 matrix? I know that 600 is sequence length after cutting, and what the 3 columns represent? BTW, your work is awesome!

If the program waked by p.run() does not finish, will the code continue to execute?

代码typo errror

1

请检查 https://github.com/PaddlePaddle/X2Paddle/blob/develop/x2paddle/project_convertor/pytorch/torch2paddle/io.py#L172 应该是把self打错成seld了

rnacentral_active.fasta

Hi developer, Great jobs! Could you provide raw sequences of your dataset?