Question about the get_e2e function

[Shi-Qi-Li]

Hi roosephu!
Thanks for sharing the code! I have a small question about the get_e2e function,

deep_matrix_factorization/main.py

Lines 40 to 49 in 3ad61e3

	def get_e2e(model):
	weight = None
	for fc in model.children():
	assert isinstance(fc, nn.Linear) and fc.bias is None
	if weight is None:
	weight = fc.weight.t()
	else:
	weight = fc(weight)
	return weight

I think it generates the final matrix by multiplying weights from different layers, but I'm a bit confused why the weight of the first layer is transposed?

Thanks!