Skip to content

Commit

Permalink
Address comments
Browse files Browse the repository at this point in the history 
Change-Id: I1de90b238623f3cdf758740b263b3de49b759bd2
  • Loading branch information
@lhutton1
lhutton1 committed Jul 23, 2020
1 parent a129235 commit 96f17ca
Show file tree
Hide file tree
Showing 3 changed files with 39 additions and 28 deletions.
3 changes: 3 additions & 0 deletions src/relay/backend/contrib/arm_compute_lib/codegen.cc
View file
Original file line number Diff line number Diff line change
Expand Up @@ -99,6 +99,8 @@ class ACLJSONSerializer : public backend::contrib::JSONSerializer {
CompositeConvNode nodes{};
const auto* fn = cn->op.as<FunctionNode>();
CHECK(fn);

// Traverse composite convolution function from child to parent
const auto* current_call = fn->body.as<CallNode>();
if (backend::IsOp(current_call, "qnn.requantize")) {
nodes.requantize = current_call;
Expand All @@ -112,6 +114,7 @@ class ACLJSONSerializer : public backend::contrib::JSONSerializer {
nodes.bias = current_call;
current_call = current_call->args[0].as<CallNode>();
}
// Enforce a convolution node exists at this point during traversal
if (nodes.requantize) {
CHECK(backend::IsOp(current_call, "qnn.conv2d"));
} else {
Expand Down
4 changes: 4 additions & 0 deletions src/runtime/contrib/arm_compute_lib/acl_runtime.cc
View file
Original file line number Diff line number Diff line change
Expand Up @@ -230,6 +230,8 @@ class ACLRuntime : public JSONRuntimeBase {
size_t num_inputs = inputs.size();
bool has_bias;
if (node.GetOpName() == "qnn.conv2d") {
CHECK(num_inputs >= 8U && num_inputs <= 9U)
<< "Quantized convolution requires 9 inputs with a bias, 8 inputs without.";
has_bias = num_inputs == 9;
layer->inputs.push_back(MakeACLTensorFromJSONEntry(inputs[0], &inputs[4], &inputs[2]));
layer->inputs.push_back(MakeACLTensorFromJSONEntry(inputs[1], &inputs[5], &inputs[3]));
Expand All @@ -239,6 +241,8 @@ class ACLRuntime : public JSONRuntimeBase {
layer->outputs.push_back(
MakeACLTensorFromJSONNode(node, &inputs[6 + has_bias], &inputs[7 + has_bias]));
} else {
CHECK(num_inputs >= 2U && num_inputs <= 3U)
<< "Convolution requires 3 inputs with a bias, 2 inputs without.";
has_bias = num_inputs == 3;
for (const auto& i : inputs) {
layer->inputs.push_back(MakeACLTensorFromJSONEntry(i));
Expand Down
60 changes: 32 additions & 28 deletions tests/python/contrib/test_arm_compute_lib/test_conv2d.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -241,23 +241,24 @@ def test_conv2d():

device = Device()
np.random.seed(0)
r = random.Random(0)

kernel_hs = [1, 2, 3, 5]
kernel_ws = [1, 2, 3, 5]
pad = [(1, 1), (2, 2), (2, 1)]
strides = [(1, 1), (2, 2)]
dilation = [(1, 1)]
out_channels = [4, 7, 16]
input_shapes = [(10, 10, 14), (12, 15, 16), (20, 20, 20)]
# composite operator (pad, bias, activation)
composite = [(False, False, False), (False, True, False), (False, False, True),
(False, True, True), (True, False, False)]
dtype = "float32"
trials = generate_trials([kernel_hs, kernel_ws, pad, strides, dilation, composite], 3)
trials = generate_trials([kernel_hs, kernel_ws, pad, strides, dilation, out_channels,
input_shapes, composite], 3)

for kernel_h, kernel_w, pad, stride, dilation, composite in trials:
out_channels = r.randint(4, 16)
for kernel_h, kernel_w, pad, stride, dilation, out_channels, input_shapes, composite in trials:
groups = 1
shape = (1,) + tuple(np.random.randint(low=max(kernel_hs + kernel_ws), high=32, size=(3,)))
shape = (1, *input_shapes)
outputs = []
inputs = {
"a": tvm.nd.array(np.random.uniform(-128, 127, shape).astype(dtype)),
Expand All @@ -281,23 +282,24 @@ def test_codegen_conv2d():
return

np.random.seed(0)
r = random.Random(0)

kernel_hs = [1, 2, 3, 5]
kernel_ws = [1, 2, 3, 5]
pad = [(1, 1), (2, 2), (2, 1)]
strides = [(1, 1), (2, 2)]
dilation = [(1, 1)]
out_channels = [4, 7, 16]
input_shapes = [(10, 10, 14), (12, 15, 16), (20, 20, 20)]
# composite operator (pad, bias, activation)
composite = [(False, False, False), (False, True, False), (False, False, True),
(False, True, True), (True, False, False)]
dtype = "float32"
trials = generate_trials([kernel_hs, kernel_ws, pad, strides, dilation, composite], 3)
trials = generate_trials([kernel_hs, kernel_ws, pad, strides, dilation, out_channels,
input_shapes, composite], 3)

for kernel_h, kernel_w, pad, stride, dilation, composite in trials:
out_channels = r.randint(4, 16)
for kernel_h, kernel_w, pad, stride, dilation, out_channels, input_shapes, composite in trials:
groups = 1
shape = (1,) + tuple(r.randint(a=max(kernel_hs + kernel_ws), b=32) for _ in range(3))
shape = (1, *input_shapes)
inputs = {"a"}

args = (shape, kernel_h, kernel_w, pad, stride, dilation, groups, dtype, out_channels)
Expand All @@ -318,32 +320,33 @@ def test_qnn_conv2d():

device = Device()
np.random.seed(0)
r = random.Random(0)

kernel_hs = [1, 2, 3, 5]
kernel_ws = [1, 2, 3, 5]
pad = [(1, 1), (2, 2)]
strides = [(1, 1), (2, 2)]
dilation = [(1, 1)]
out_channels = [4, 7, 16]
input_shapes = [(10, 10, 14), (12, 15, 16), (20, 20, 20)]
# composite operator (pad, bias, activation)
composite = [(False, False, False), (False, True, False), (False, False, True),
(False, True, True), (True, False, False)]
dtype = "uint8"
trials = generate_trials([kernel_hs, kernel_ws, pad, strides, dilation, composite], 3)
trials = generate_trials([kernel_hs, kernel_ws, pad, strides, dilation, out_channels,
input_shapes, composite], 3)

for kernel_h, kernel_w, pad, stride, dilation, composite in trials:
out_channels = r.randint(4, 16)
for kernel_h, kernel_w, pad, stride, dilation, out_channels, input_shapes, composite in trials:
groups = 1
shape = (1,) + tuple(np.random.randint(low=max(kernel_hs + kernel_ws), high=32, size=(3,)))
shape = (1, *input_shapes)
outputs = []
inputs = {
"a": tvm.nd.array(np.random.uniform(0, 255, shape).astype(dtype))
}

input_zp = r.randint(0, 255)
input_sc = r.random() * 2
kernel_zp = r.randint(0, 255)
kernel_sc = r.random() * 2
input_zp = 100
input_sc = 0.5
kernel_zp = 25
kernel_sc = 0.03
output_zp, output_sc = _get_qnn_params(input_zp, input_sc,
kernel_zp, kernel_sc,
kernel_h, kernel_w, shape[3])
Expand All @@ -370,29 +373,30 @@ def test_codegen_qnn_conv2d():
return

np.random.seed(0)
r = random.Random(0)

kernel_hs = [1, 2, 3, 5]
kernel_ws = [1, 2, 3, 5]
pad = [(1, 1), (2, 2), (2, 1)]
strides = [(1, 1), (2, 2)]
dilation = [(1, 1)]
out_channels = [4, 7, 16]
input_shapes = [(10, 10, 14), (12, 15, 16), (20, 20, 20)]
# composite operator (pad, bias, activation)
composite = [(False, False, False), (False, True, False), (False, False, True),
(False, True, True), (True, False, False)]
dtype = "uint8"
trials = generate_trials([kernel_hs, kernel_ws, pad, strides, dilation, composite], 3)
trials = generate_trials([kernel_hs, kernel_ws, pad, strides, dilation, out_channels,
input_shapes, composite], 3)

for kernel_h, kernel_w, pad, stride, dilation, composite in trials:
out_channels = r.randint(4, 16)
for kernel_h, kernel_w, pad, stride, dilation, out_channels, input_shapes, composite in trials:
groups = 1
shape = (1,) + tuple(r.randint(a=max(kernel_hs + kernel_ws), b=32) for _ in range(3))
shape = (1, *input_shapes)
inputs = {"a"}

input_zp = r.randint(0, 255)
input_sc = r.random() * 2
kernel_zp = r.randint(0, 255)
kernel_sc = r.random() * 2
input_zp = 100
input_sc = 0.5
kernel_zp = 25
kernel_sc = 0.03
output_zp, output_sc = _get_qnn_params(input_zp, input_sc,
kernel_zp, kernel_sc,
kernel_h, kernel_w, shape[3])
Expand Down

0 comments on commit 96f17ca

Please sign in to comment.