def run_conv_valid(
self,
inputs_shape,
filters_shape,
border_mode="valid",
filter_dilation=(1, 1, 1),
subsample=(1, 1, 1),
verify_grad=False,
):
inputs_shape = [inputs_shape[i] for i in (0, 4, 1, 2, 3)]
filters_shape = [filters_shape[i] for i in (0, 4, 1, 2, 3)]
inputs_val = np.random.random(inputs_shape).astype(config.floatX)
filters_val = np.random.random(filters_shape).astype(config.floatX)
inputs = gpuarray_shared_constructor(inputs_val)
filters = gpuarray_shared_constructor(filters_val)
conv_ref = Corr3dMM(
border_mode=border_mode,
filter_dilation=filter_dilation,
subsample=subsample,
)(ref_cast(inputs), ref_cast(filters))
f_ref = aesara.function([], conv_ref, mode=mode_without_gpu)
conv = GpuCorr3dMM(
border_mode=border_mode,
filter_dilation=filter_dilation,
subsample=subsample,
)(inputs, filters)
f = aesara.function([], conv, mode=mode_with_gpu)
res_ref = f_ref()
res = f()
utt.assert_allclose(res_ref, res)
if verify_grad:
utt.verify_grad(
GpuCorr3dMM(
border_mode=border_mode,
filter_dilation=filter_dilation,
subsample=subsample,
),
[inputs_val, filters_val],
mode=mode_with_gpu,
)
def local_abstractconv3d_gemm(fgraph, node):
# If config.blas__ldflags is empty, Aesara will use
# a NumPy C implementation of [sd]gemm_.
if config.cxx == "" or node.inputs[0].dtype == "float16":
return
if not isinstance(node.op, AbstractConv3d):
return None
img, kern = node.inputs
if not isinstance(img.type, TensorType) or not isinstance(kern.type, TensorType):
return None
# need to flip the kernel if necessary
if node.op.filter_flip:
kern = kern[:, :, ::-1, ::-1, ::-1]
rval = Corr3dMM(
border_mode=node.op.border_mode,
subsample=node.op.subsample,
filter_dilation=node.op.filter_dilation,
num_groups=node.op.num_groups,
)(img, kern)
copy_stack_trace(node.outputs[0], rval)
return [rval]