Unflatten (dim=1, unflattened_size= (1, 35, 35)), nn. ... This is the reason why onnx2keras translates a conv2d layer with padding into two different layers ...
確定! 回上一頁