Hi buddy,
Does anyone try to implement roi pooling using tvm?
Anyone can help?
1 Like
shape = (1, 32, 38, 50)
in_dtype= 'float16'
x0 = relay.var("data0", shape=(1, 32, 38, 50), dtype=in_dtype)
x1 = relay.var("data1", shape=(2, 5), dtype='float16')
y = relay.vision.roi_pool(x0, x1, (7, 7), 1.0)
func = relay.Function([x0, x1], y)