def burn():
sz = 128
img_shp = [sz, sz, sz, sz]
kern_shp = [sz // 2, sz, 3, 3]
out_shp = get_conv_output_shape(img_shp, kern_shp, "valid", (1, 1))
img = tensor4("img")
kern = tensor4("kern")
out = tensor4("out")
def rand(shp):
return np.random.rand(*shp).astype(config.floatX)
img = aesara.shared(rand(img_shp))
kern = aesara.shared(rand(kern_shp))
out = aesara.shared(rand(out_shp))
# beta 1 is needed to force the reuse of out, otherwise, it is
# replaced by a GpuAllocEmpty
o1 = dnn._dnn_conv(img, kern, conv_mode="conv", out=out, beta=1.0)
mode = aesara.compile.get_default_mode().including(
"local_remove_all_assert")
f = aesara.function([], [o1], mode=mode)
aesara.printing.debugprint(f)
print("Start computation")
for i in range(10000):
f.fn()
print("Computation stopped")
def infer_shape(self, fgraph, node, input_shape):
imshp = input_shape[0]
kshp = input_shape[1]
res = get_conv_output_shape(
imshp, kshp, self.border_mode, self.subsample, self.filter_dilation
)
return [res]
def array_like_conv_output(self, inputs_shape, filters_shape, border_mode,
subsample, dilation, dtype):
# Return a random array with inferred convolution output shape.
out_shp = get_conv_output_shape(inputs_shape, filters_shape,
border_mode, subsample, dilation)
out_shp = assert_conv_shape(out_shp)
return np.random.random(out_shp).astype(dtype)
def get_if_valid_conv_output_shape(case_tuple):
# Filter function to keep only cases that produce valid convolution output shapes.
out_shp = get_conv_output_shape(
case_tuple[0], # input shape
case_tuple[1], # filter shape
case_tuple[4], # border mode
case_tuple[2], # subsample
case_tuple[3],
) # dilation
try:
return assert_conv_shape(out_shp)
except ValueError:
return False
def local_conv2d_gradinputs_cpu(fgraph, node):
if (
not isinstance(node.op, AbstractConv2d_gradInputs)
or node.inputs[0].dtype == "float16"
):
return None
kern, topgrad, shape = node.inputs
if not isinstance(kern.type, TensorType) or not isinstance(
topgrad.type, TensorType
):
return None
if node.op.border_mode not in ["full", "valid"]:
return None
if not node.op.filter_flip:
# Not tested yet
return None
if node.op.num_groups > 1 or node.op.unshared:
return None
# Conv 3d implementation, needed when subsample > 2
if node.op.border_mode == "valid" and node.op.subsample != (1, 1):
# The op don't support that anymore.
return False
# Conv2d Implementation
dx, dy = node.op.subsample
if dx not in (1, 2) or dy not in (1, 2):
# Not implemented in the gradient of ConvOp
return None
if node.op.imshp is None:
op_imshp = (None, None, None, None)
else:
op_imshp = node.op.imshp
if node.op.kshp is None:
op_kshp = (None, None, None, None)
else:
op_kshp = node.op.kshp
if None in op_imshp or None in op_kshp:
if (dx, dy) != (1, 1):
return None
mode = "valid"
if not node.op.border_mode == "full":
mode = "full"
filters = kern.dimshuffle((1, 0, 2, 3))
filters = filters[:, :, ::-1, ::-1]
outshp = get_conv_output_shape(
op_imshp,
op_kshp,
node.op.border_mode,
node.op.subsample,
node.op.filter_dilation,
)[2:]
fulloutshp = get_conv_output_shape(op_imshp, op_kshp, node.op.border_mode, (1, 1))[
2:
]
nkern = op_imshp[1]
imshp = (op_kshp[0], outshp[0], outshp[1])
imshp_logical = (op_kshp[0], fulloutshp[0], fulloutshp[1])
din = ConvOp(
imshp,
op_kshp[2:],
nkern,
op_imshp[0],
1,
1,
output_mode=mode,
unroll_batch=None,
unroll_kern=None,
unroll_patch=None,
imshp_logical=imshp_logical,
kshp_logical=None,
version=-1,
direction_hint="bprop inputs",
)
din = din(topgrad, filters)
copy_stack_trace(node.outputs[0], din)
din = aesara.tensor.patternbroadcast(din, node.outputs[0].broadcastable)
copy_stack_trace(node.outputs[0], din)
return [din]
def local_conv2d_gradweight_cpu(fgraph, node):
if (
not isinstance(node.op, AbstractConv2d_gradWeights)
or node.inputs[0].dtype == "float16"
):
return None
img, topgrad, shape = node.inputs
if not isinstance(img.type, TensorType) or not isinstance(topgrad.type, TensorType):
return None
if node.op.border_mode not in ["full", "valid"]:
return None
if not node.op.filter_flip:
# Not tested yet
return
if node.op.num_groups > 1 or node.op.unshared:
return None
if node.op.border_mode == "valid" and (node.op.subsample != (1, 1)):
return None
dx, dy = node.op.subsample
if dx not in (1, 2) or dy not in (1, 2):
# Not implemented in the gradient of ConvOp
return None
if node.op.imshp is None:
op_imshp = (None, None, None, None)
else:
op_imshp = node.op.imshp
if node.op.kshp is None:
op_kshp = (None, None, None, None)
else:
op_kshp = node.op.kshp
if None in op_imshp or None in op_kshp:
if (dx, dy) != (1, 1):
# We cannot infer the shapes
return None
# Determine gradient on kernels
assert len(op_imshp) == 4 and len(op_kshp) == 4
outshp = get_conv_output_shape(
op_imshp,
op_kshp,
node.op.border_mode,
node.op.subsample,
node.op.filter_dilation,
)[2:]
fulloutshp = get_conv_output_shape(op_imshp, op_kshp, node.op.border_mode, (1, 1))[
2:
]
newimg = img.dimshuffle((1, 0, 2, 3))
newtopgrad = topgrad.dimshuffle((1, 0, 2, 3))
if node.op.border_mode == "valid":
(img, filters) = (newimg, newtopgrad)
kshp_logical = fulloutshp
kshp_logical_top_aligned = False
imshp_logical = None
(bsize, nkern) = (op_imshp[1], op_kshp[0])
imshp = (op_imshp[0], op_imshp[2], op_imshp[3])
kshp = outshp
elif node.op.border_mode == "full":
(img, filters) = (newtopgrad, newimg)
kshp_logical = None
kshp_logical_top_aligned = True
imshp_logical = (op_imshp[0], fulloutshp[0], fulloutshp[1])
(bsize, nkern) = (op_kshp[0], op_imshp[1])
imshp = (op_imshp[0], outshp[0], outshp[1])
kshp = op_imshp[2:]
else:
raise NotImplementedError("Only [full,valid] modes are currently supported.")
# Flip the kernels
filters = filters[:, :, ::-1, ::-1]
dw = ConvOp(
imshp,
kshp,
nkern,
bsize,
1,
1,
output_mode="valid",
unroll_batch=None,
unroll_kern=None,
unroll_patch=None,
imshp_logical=imshp_logical,
kshp_logical=kshp_logical,
kshp_logical_top_aligned=kshp_logical_top_aligned,
direction_hint="bprop weights",
)
res = dw(img, filters)
copy_stack_trace(node.outputs[0], res)
if node.op.border_mode == "valid":
res = res.dimshuffle((1, 0, 2, 3))
res = res[:, :, ::-1, ::-1]
res = aesara.tensor.patternbroadcast(res, node.outputs[0].broadcastable)
copy_stack_trace(node.outputs[0], res)
return [res]
def local_abstractconv_cudnn_alt(fgraph, node):
if not isinstance(node.op, (AbstractConv2d, AbstractConv2d_gradWeights,
AbstractConv2d_gradInputs)):
return
if version(raises=False) < 6000 and node.op.filter_dilation != (1, 1):
return None
if node.op.unshared:
return None
if isinstance(node.op.border_mode, tuple) and any(
isinstance(p, tuple) for p in node.op.border_mode):
# Asymmetric padding not yet supported
return None
inp1 = node.inputs[0]
inp2 = node.inputs[1]
if not dnn_available(inp1.type.context_name):
return
op = node.op
border_mode = node.op.border_mode
subsample = node.op.subsample
filter_dilation = node.op.filter_dilation
num_groups = node.op.num_groups
precision, _ = get_precision(None, [inp1, inp2])
if node.op.filter_flip:
conv_mode = "conv"
else:
conv_mode = "cross"
if isinstance(op, AbstractConv2d):
if border_mode == "half" or subsample != (1, 1) or num_groups != 1:
return None
if border_mode == "full":
direction_hint = "bprop inputs"
elif border_mode == "valid" and filter_dilation == (1, 1):
direction_hint = "bprop weights"
else:
return None
rval = dnn_conv(
inp1,
inp2,
border_mode=border_mode,
subsample=subsample,
dilation=filter_dilation,
direction_hint=direction_hint,
conv_mode=conv_mode,
num_groups=num_groups,
)
elif isinstance(op, AbstractConv2d_gradWeights):
if (border_mode == "valid" and subsample == (1, 1)
and filter_dilation == (1, 1) and num_groups == 1):
img = gpu_contiguous(inp1)
topgrad = gpu_contiguous(inp2)
ctx_name = infer_context_name(img, topgrad)
img = gpu_contiguous(img.dimshuffle(1, 0, 2, 3))
topgrad = gpu_contiguous(topgrad.dimshuffle(1, 0, 2, 3))
ishape = [shape_i_op(i)(img) for i in range(img.ndim)]
tshape = [shape_i_op(i)(topgrad) for i in range(topgrad.ndim)]
out_shp = get_conv_output_shape(
ishape,
tshape,
border_mode=border_mode,
subsample=subsample,
filter_dilation=filter_dilation,
)
out_shp = assert_conv_shape(out_shp)
out = GpuAllocEmpty(dtype=img.dtype,
context_name=ctx_name)(*out_shp)
desc = GpuDnnConvDesc(
border_mode=border_mode,
subsample=subsample,
dilation=filter_dilation,
conv_mode="cross",
precision=precision,
)(out.shape)
conv = GpuDnnConv(algo=None, num_groups=num_groups)(img, topgrad,
out, desc)
if conv_mode == "conv":
conv = conv[:, :, ::-1, ::-1]
rval = as_gpuarray_variable(conv.dimshuffle(1, 0, 2, 3), ctx_name)
else:
return None
elif isinstance(op, AbstractConv2d_gradInputs):
if border_mode == "valid" and subsample == (1, 1) and num_groups == 1:
kerns = gpu_contiguous(inp1.dimshuffle(1, 0, 2, 3))
topgrad = gpu_contiguous(inp2)
ctx_name = infer_context_name(kerns, topgrad)
conv_mode = "cross" if conv_mode == "conv" else "conv"
desc = GpuDnnConvDesc(
border_mode="full",
subsample=subsample,
dilation=filter_dilation,
conv_mode=conv_mode,
precision=precision,
)(kerns.shape)
tshape = [shape_i_op(i)(topgrad) for i in range(topgrad.ndim)]
kshape = [shape_i_op(i)(kerns) for i in range(kerns.ndim)]
shape = get_conv_output_shape(
tshape,
kshape,
border_mode="full",
subsample=subsample,
filter_dilation=filter_dilation,
)
shape = assert_conv_shape(shape)
out = GpuAllocEmpty(dtype=topgrad.dtype,
context_name=ctx_name)(*shape)
rval = GpuDnnConv(algo=None, num_groups=num_groups)(topgrad, kerns,
out, desc)
else:
return None
return [rval]
args.input_shape,
args.filter_shape,
args.subsample,
args.dilation,
args.border_mode,
args.conv_mode,
args.alpha,
args.beta,
)
if args.print_infos:
CheckDnn.print_infos(count_tests=False)
print("======================")
print("Running", test, algo, dtype, precision, *parameters)
if test == FWD:
tests.run_conv_fwd(algo, dtype, precision, parameters)
expected_output_shape = get_conv_output_shape(
args.input_shape,
args.filter_shape,
args.border_mode,
args.subsample,
args.dilation,
)
elif test == BWD_FILTER:
tests.run_conv_gradweight(algo, dtype, precision, parameters)
expected_output_shape = args.filter_shape
elif test == BWD_DATA:
tests.run_conv_gradinput(algo, dtype, precision, parameters)
expected_output_shape = args.input_shape
print("Computed shape:", expected_output_shape)
print("... OK")
