bugfix: RuntimeError: view size is not compatible with input tensor's size and stride (at least one dimension spans across

报错原因：这是因为view()需要Tensor中的元素地址是连续的，但可能出现Tensor不连续的情况，所以先用 .contiguous()。将其在内存中变成连续分布即可。
解决方案：output = self.dense(out.transpose(1, 2).contiguous().view(B, L, -1))

visual.py

@@ -66,7 +66,7 @@ class Attention(nn.Module):
         out = attention_fn_default(
             q, k, v
         )
-        output = self.dense(out.transpose(1, 2).view(B, L, -1))
+        output = self.dense(out.transpose(1, 2).contiguous().view(B, L, -1))
         output = self.output_dropout(output)
         return output