[Fix][Frontend][TOPI] minor bugs

[hzfan]

Fix 2 minor bugs in:

• reshape array of shape [1] to []
• pytorch linear

cc @yzhliu @comaniac

[masahi]

We can use _op.nn.dense without transposing its second input. Also see https://discuss.tvm.apache.org/t/pytorch-unable-to-import-a-simple-torch-nn-linear-to-relay/10383/4

[comaniac]

Maybe I missed something, does that mean the current linear layers converted from PyTorch are wrong?

[masahi]

Yes, see my comment in the discuss thread above. Until PT 1.8, nn.Linear converts to aten::addmm, so this code path never hit unless users explicitly call something that converts to aten::linear. Since PT 1.9, nn.Linear converts to aten::linear, so this becomes a major issue.

[comaniac]

Ah I see, so this bug doesn't explode until PT 1.9. Thanks for the explanation!

[hzfan]

@masahi My concern is that linear is supposed to support inputs with arbitrary ranks like 1d, 2d, 3d, 4d.... The different input dimensions are normalized in self.matmul. If _op.nn.dense is directly used, the usability of linear would be limited to inputs whose ranks are 2.

[comaniac]

How about we just make an exception: when the input rank is 2 then use dense; otherwise use matmul?

[masahi]

Doesn't op.nn.dense also support ND input?

tvm/python/tvm/relay/op/nn/nn.py

Lines 1528 to 1530 in 850abb0

	data : tvm.relay.Expr
	The input data to the operator,
	of shape `(d_1, d_2, ..., d_n, units_in)`.

[masahi]

also if we want to support ND input, please add more tests.

[hzfan]

@masahi Tests with 3d input are added. For now I use dense when both inputs are 2d and fallback to matmul otherwise.

I tried op.nn.dense in ND scenario, but it fails at

tvm/python/tvm/topi/x86/dense_alter_op.py

Lines 29 to 37 in 4b9d43e

	@dense_alter_layout.register(["cpu", "arm_cpu"])
	def _alter_dense_layout(attrs, inputs, tinfos, out_type):
	target = tvm.target.Target.current(allow_none=False)
	dispatch_ctx = autotvm.task.DispatchContext.current
	data_tensor, weight_tensor = tinfos
	out_dtype = out_type.dtype
	M, K = get_const_tuple(data_tensor.shape)
	N, _ = get_const_tuple(weight_tensor.shape)

where data_tensor.shape is unpacked into a tuple of size 2. I also check that in tests/python/relay/test_op_level1.py dense is tested numerically for 2d inputs only. Not sure if I miss anything.

[masahi]

Yeah it looks like relay doc is wrong and topi dense indeed only supports 2D #8412

[junrushao]

I agree with Masa, and otherwise looks good to me :-)

[hzfan]

Any idea why ci always gets aborted in Build:CPU? Did I miss anything?

[comaniac]

Any idea why ci always gets aborted in Build:CPU? Did I miss anything?

It was timeout (4hrs). Re-trigger CI may resolve this issue

[masahi]

It should be fixed by #8658

[junrushao]

Thanks @hzfan @masahi @comaniac! It is merged :-)