[TOPI][OP] Support Faster-RCNN Proposal OP on CPU

[FrozenGene]

Currently, our TVM only implement proposal op on CUDA. However, we should support Proposal op on CPU too.

@vinx13 @tqchen please help to review it.

[kevinthesun]

In topi cpu, all vision ops are written with hybrid script instead of ir builder now. Should we keep things consistent?

[FrozenGene]

I tried to play Hybrid Script, but seems that we have some restriction so that I can not complete Hybrid Script replacement. For example:

@hybrid.script
def outer_product(a, b, x):
    first_shape = a.shape[0]
    first_shape //= 2
    c = output_tensor((3, 3), 'float32')
    for i in const_range(first_shape):
        c[0, 0] = x[i // 2]
    return c

a = tvm.placeholder((3, ), name='a')
b = tvm.placeholder((3, ), name='b')
x = [1, 2, 3]
c = outer_product(a, b, tvm.convert(x))

It will report ValueError: Const range should start from a const and iterate const times

We should change into

   first_shape = a.shape[0] // 2
    c = output_tensor((3, 3), 'float32')
    for i in const_range(first_shape):
        c[0, 0] = x[i // 2]
    return c

This make code strange. Do we have any good ways to solve it?

Another thing is GPU / CPU relay on some common functions like generate_anchor / reg_bbox and so on. Like in generate_anchor, we use math.floor, hybrid script doesn't support. So if I have to move the generate_anchor implementation into predict_bbox_ir, I meet one problem is floor. When I add it from tvm.intrin import floor, I meet problem TypeError: 'FloatImm' object does not support indexing, you could try this simple reproduce case:

@hybrid.script
def outer_product(a, b, x):
    first_shape = a.shape[0]
    #first_shape = first_shape // 2
    c = output_tensor((3, 3), 'float32')
    for i in const_range(first_shape):
        floor(3.1 / 2.0)
    return c

So must we use hybrid script? Seems hybrid script has some restriction. Welcome feedback to above problems, if we could solve, I will try to use hybrid script to continue.

[FrozenGene]

ping @kevinthesun @vinx13

[vinx13]

I met these issue too when I trying to migrate proposal CUDA op to hybrid, I think it’s okay to use ir builder for the time being

[vinx13]

Thanks @FrozenGene @kevinthesun this is merged

[junrushao]

@were Could you also look into the hybrid script issue?