【Bugs】ValueError not enough values to unpack (expected 2, got 1)

【Bugs】ValueError not enough values to unpack (expected 2, got 1) 【Bugs】ValueError: not enough values to unpack (expected 2, got 1) 1. Bug 场景

Pytorch环境下用Huggingface获取bert.outputs代码output = model(**input)[0]的时候出现bug

>>>model(**input)
Traceback (most recent call last):
  File "", line 1, in 
  File "D:anaconda3envspt16libsite-packagestorchnnmodulesmodule.py", line 722, in _call_impl
    result = self.forward(*input, **kwargs)
  File "D:anaconda3envspt16libsite-packagestransformersmodelsbertmodeling_bert.py", line 949, in forward
    batch_size, seq_length = input_shape
ValueError: not enough values to unpack (expected 2, got 1)

input是shape形如1D的dict的形式：

{'attention_mask': tensor([1, 1, 1], device='cuda:0'), 'input_ids': tensor([ 101, 2088, 102], device='cuda:0'), 'token_type_ids': tensor([0, 0, 0], device='cuda:0')}

但是我们加载的model是model = BertModel.from_pretrained(checkpoint)的形式，所以input_shape必须是形如2D的tensor.所以在此用.unsqueeze(0)方法实现。

2. Debug

将1D dict转化成2D dict

for k in input.keys():
    input[k]= input[k].unsqueeze(0)

此时model(input)[0]不再报错。