def test_maxpool_1d(data_shape_nwc, pool_size, strides, padding):
"""Test a subgraph with a single maxpool_1d operator."""
ishape = data_shape_nwc
input0 = relay.var("input", relay.TensorType(ishape, "int8"))
out = relay.op.nn.max_pool1d(input0,
pool_size,
layout="NWC",
strides=strides,
padding=padding)
mod = tvm.IRModule.from_expr(relay.Function([input0], out))
inputs = {
"input": np.random.randint(low=-128,
high=127,
size=ishape,
dtype="int8")
}
output_list = generate_ref_data(mod, inputs)
compile_and_run(
AOTTestModel(module=mod, inputs=inputs, outputs=output_list),
runner=AOT_CORSTONE300_RUNNER,
interface_api="c",
use_unpacked_api=True,
target_opts={
"-keys": "arm_cpu",
"-mcpu": "cortex-m7",
},
)
def test_conv2d_int8_tflite(ifm_shape, kernel_shape, strides, dilation,
padding, activation):
interface_api = "c"
use_unpacked_api = True
test_runner = AOT_USMP_CORSTONE300_RUNNER
dtype = "int8"
tflite_model, relay_mod, params = create_conv2d_tflite_relay_models(
ifm_shape, kernel_shape, strides, dilation, padding, activation, dtype)
cmsisnn_mod = cmsisnn.partition_for_cmsisnn(relay_mod, params)
# validate pattern matching
assert_partitioned_function(relay_mod, cmsisnn_mod)
# validate CMSIS-NN output against TFLite output
input_map, output_map, output_tolerance = generate_ref_data_tflite(
tflite_model)
compile_and_run(
AOTTestModel(
module=cmsisnn_mod,
inputs=input_map,
outputs=output_map,
params=params,
output_tolerance=output_tolerance,
),
test_runner,
interface_api,
use_unpacked_api,
)
def test_avgpool_1d(data_shape_ncw, pool_size, strides, padding):
"""Test a subgraph with a single avgpool_1d operator."""
ishape = data_shape_ncw
input0 = relay.var("input", relay.TensorType(ishape, "int32"))
out0 = relay.op.nn.avg_pool1d(input0, pool_size, layout="NCW", strides=strides, padding=padding)
ref_mod = tvm.IRModule.from_expr(relay.Function([input0], out0))
input1 = relay.var("input", relay.TensorType(ishape, "int16"))
out1 = relay.op.nn.avg_pool1d(input1, pool_size, layout="NCW", strides=strides, padding=padding)
mod = tvm.IRModule.from_expr(relay.Function([input1], out1))
input_data = np.random.randint(low=-10, high=10, size=ishape, dtype="int32")
inputs = {"input": input_data}
output_list = generate_ref_data(ref_mod, inputs)
compile_and_run(
AOTTestModel(
module=mod, inputs={"input": input_data.astype(dtype="int16")}, outputs=output_list
),
runner=AOT_CORSTONE300_RUNNER,
interface_api="c",
use_unpacked_api=True,
target_opts={
"-keys": "arm_cpu",
"-mcpu": "cortex-m7",
},
)
def test_op_int8(zero_point, scale):
interface_api = "c"
use_unpacked_api = True
test_runner = AOT_USMP_CORSTONE300_RUNNER
dtype = "int8"
shape = [1, 16, 16, 3]
model = make_model(shape, dtype, dtype, zero_point, scale)
orig_mod = make_module(model)
cmsisnn_mod = cmsisnn.partition_for_cmsisnn(orig_mod)
# validate pattern matching
assert_partitioned_function(orig_mod, cmsisnn_mod)
# validate the output
in_min, in_max = get_range_for_dtype_str(dtype)
np.random.seed(0)
input_data = np.random.randint(in_min,
high=in_max,
size=shape,
dtype=dtype)
inputs = {"in0": input_data}
params = {}
output_list = generate_ref_data(orig_mod["main"], inputs, params)
compile_and_run(
AOTTestModel(module=cmsisnn_mod,
inputs=inputs,
outputs=output_list,
params=params),
test_runner,
interface_api,
use_unpacked_api,
)
def test_dense(M, K, N):
"""Test a subgraph with a single dense operator."""
ishape = (M, K)
wshape = (N, K)
input0 = relay.var("input", relay.TensorType(ishape, "int8"))
dense_f = relay.op.nn.batch_flatten(input0)
weight0 = relay.const(
np.random.randint(low=-10, high=10, size=wshape, dtype="int8"))
out = relay.op.nn.dense(dense_f, weight0, out_dtype="int32")
mod = tvm.IRModule.from_expr(relay.Function([input0], out))
inputs = {
"input": np.random.randint(low=-128,
high=127,
size=ishape,
dtype="int8")
}
output_list = generate_ref_data(mod, inputs)
compile_and_run(
AOTTestModel(module=mod, inputs=inputs, outputs=output_list),
runner=AOT_CORSTONE300_RUNNER,
interface_api="c",
use_unpacked_api=True,
target_opts={
"-keys": "arm_cpu",
"-mcpu": "cortex-m7",
},
)
def test_cnn_small():
# download the model
base_url = "https://github.com/ARM-software/ML-zoo/raw/master/models/keyword_spotting/cnn_small/tflite_int8"
file_to_download = "cnn_s_quantized.tflite"
model_file = download_testdata("{}/{}".format(base_url, file_to_download), file_to_download)
with open(model_file, "rb") as f:
tflite_model_buf = f.read()
input_shape = (1, 490)
in_min, in_max = get_range_for_dtype_str("int8")
input_data = np.random.randint(in_min, high=in_max, size=input_shape).astype(np.float32)
orig_mod, params = convert_to_relay(tflite_model_buf, input_data, "input")
cmsisnn_mod = cmsisnn.partition_for_cmsisnn(orig_mod, params)
# validate CMSIS-NN output against CPU output
interface_api = "c"
use_unpacked_api = True
test_runner = AOT_CORSTONE300_RUNNER
inputs = {"input": input_data}
params = {}
output_list = generate_ref_data(orig_mod["main"], inputs, params)
compile_and_run(
AOTTestModel(module=cmsisnn_mod, inputs=inputs, outputs=output_list, params=params),
test_runner,
interface_api,
use_unpacked_api,
)
def test_conv2d(data_shape_nhwc, kernel_size, num_filter, strides, padding,
dilation, dtype):
"""Test a subgraph with a single conv2d operator."""
ishape = data_shape_nhwc
wshape = (*kernel_size, data_shape_nhwc[-1], num_filter)
weight_data = np.random.randint(low=-10, high=10, size=wshape, dtype=dtype)
input0 = relay.var("input", relay.TensorType(ishape, dtype))
weight0 = relay.const(weight_data)
out0 = relay.op.nn.conv2d(
input0,
weight0,
kernel_size=kernel_size,
strides=strides,
padding=padding,
dilation=(dilation, dilation),
data_layout="NHWC",
kernel_layout="HWIO",
out_dtype="int32",
out_layout="NHWC",
)
ref_mod = tvm.IRModule.from_expr(relay.Function([input0], out0))
input1 = relay.var("input", relay.TensorType(ishape, dtype))
weight1 = relay.const(np.moveaxis(weight_data, 2, -1))
out1 = relay.op.nn.conv2d(
input1,
weight1,
kernel_size=kernel_size,
strides=strides,
padding=padding,
dilation=(dilation, dilation),
data_layout="NHWC",
kernel_layout="HWOI",
out_dtype="int32",
out_layout="NHWC",
)
mod = tvm.IRModule.from_expr(relay.Function([input1], out1))
inputs = {
"input": np.random.randint(low=-128,
high=127,
size=ishape,
dtype=dtype)
}
output_list = generate_ref_data(ref_mod, inputs)
compile_and_run(
AOTTestModel(module=mod, inputs=inputs, outputs=output_list),
runner=AOT_CORSTONE300_RUNNER,
interface_api="c",
use_unpacked_api=True,
target_opts={
"-keys": "arm_cpu",
"-mcpu": "cortex-m7",
},
)
def test_op_int8(op, input_0_scale, input_0_zero_point, input_1_scale, input_1_zero_point):
interface_api = "c"
use_unpacked_api = True
test_runner = AOT_CORSTONE300_RUNNER
dtype = "int8"
shape = [1, 16, 16, 3]
model = make_model(
op,
shape,
dtype,
dtype,
input_0_scale,
input_0_zero_point,
input_1_scale,
input_1_zero_point,
)
orig_mod = make_module(model)
cmsisnn_mod = cmsisnn.partition_for_cmsisnn(orig_mod)
# validate pattern matching
attrs = [
cmsisnn_mod[var.name_hint].attrs
for var in cmsisnn_mod.get_global_vars()
if cmsisnn_mod[var.name_hint].attrs
]
assert any(attrs), "At least one function with external attributes was expected."
compilers = [
key == "Compiler" and value == "cmsis-nn" for attr in attrs for key, value in attr.items()
]
assert any(compilers), "Module does not contain function for cmsisnn target."
assert count_num_calls(orig_mod) == count_num_calls(
cmsisnn_mod
), "Number of calls changed during partitioning"
# validate the output
in_min, in_max = get_range_for_dtype_str(dtype)
inputs = {
"input_0": np.random.randint(in_min, high=in_max, size=shape, dtype=dtype),
"input_1": np.random.randint(in_min, high=in_max, size=shape, dtype=dtype),
}
output_list = generate_ref_data(orig_mod["main"], inputs)
compile_and_run(
AOTTestModel(
module=cmsisnn_mod,
inputs=inputs,
outputs=output_list,
output_tolerance=1,
),
test_runner,
interface_api,
use_unpacked_api,
)
def test_op_int8(
in_shape,
pool_size,
strides,
padding,
relu_type,
pool_type,
zero_point,
scale,
):
interface_api = "c"
use_unpacked_api = True
test_runner = AOT_USMP_CORSTONE300_RUNNER
dtype = "int8"
model = make_model(
pool_type,
in_shape,
pool_size,
strides,
padding,
dtype,
scale,
zero_point,
relu_type,
)
orig_mod = make_module(model)
cmsisnn_mod = cmsisnn.partition_for_cmsisnn(orig_mod)
# validate pattern matching
assert_partitioned_function(orig_mod, cmsisnn_mod)
# validate the output
in_min, in_max = get_range_for_dtype_str(dtype)
np.random.seed(0)
inputs = {
"input":
np.random.randint(in_min, high=in_max, size=in_shape, dtype="int8"),
}
output_list = generate_ref_data(orig_mod["main"], inputs)
compile_and_run(
AOTTestModel(
module=cmsisnn_mod,
inputs=inputs,
outputs=output_list,
params=None,
output_tolerance=1,
),
test_runner,
interface_api,
use_unpacked_api,
)
def test_constant_input_int8(op, input_0, input_1):
interface_api = "c"
use_unpacked_api = True
test_runner = AOT_USMP_CORSTONE300_RUNNER
dtype = "int8"
shape = [1, 16, 16, 3]
input_0_scale = 0.256
input_0_zero_point = 33
input_1_scale = 0.128
input_1_zero_point = -24
model = make_model(
op,
input_0,
input_1,
input_0_scale,
input_0_zero_point,
input_1_scale,
input_1_zero_point,
)
orig_mod = make_module(model)
cmsisnn_mod = cmsisnn.partition_for_cmsisnn(orig_mod)
# validate pattern matching
assert_partitioned_function(orig_mod, cmsisnn_mod)
# validate the output
in_min, in_max = get_range_for_dtype_str(dtype)
inputs = {}
if isinstance(input_0, tvm.relay.expr.Var):
inputs.update({
"input_0":
np.random.randint(in_min, high=in_max, size=shape, dtype=dtype)
})
if isinstance(input_1, tvm.relay.expr.Var):
inputs.update({
"input_1":
np.random.randint(in_min, high=in_max, size=shape, dtype=dtype)
})
output_list = generate_ref_data(orig_mod["main"], inputs)
compile_and_run(
AOTTestModel(
module=cmsisnn_mod,
inputs=inputs,
outputs=output_list,
output_tolerance=1,
),
test_runner,
interface_api,
use_unpacked_api,
)
def test_op_int8(op, relu_type, input_0_scale, input_0_zero_point,
input_1_scale, input_1_zero_point):
interface_api = "c"
use_unpacked_api = True
test_runner = AOT_USMP_CORSTONE300_RUNNER
dtype = "int8"
shape = [1, 16, 16, 3]
model = make_model(
op,
generate_variable("input_0"),
generate_variable("input_1"),
input_0_scale,
input_0_zero_point,
input_1_scale,
input_1_zero_point,
relu_type,
)
orig_mod = make_module(model)
cmsisnn_mod = cmsisnn.partition_for_cmsisnn(orig_mod)
# validate pattern matching
assert_partitioned_function(orig_mod, cmsisnn_mod)
# validate the output
in_min, in_max = get_range_for_dtype_str(dtype)
inputs = {
"input_0": np.random.randint(in_min,
high=in_max,
size=shape,
dtype=dtype),
"input_1": np.random.randint(in_min,
high=in_max,
size=shape,
dtype=dtype),
}
output_list = generate_ref_data(orig_mod["main"], inputs)
compile_and_run(
AOTTestModel(
module=cmsisnn_mod,
inputs=inputs,
outputs=output_list,
output_tolerance=1,
),
test_runner,
interface_api,
use_unpacked_api,
)
def check_aot_executor_result(
mod, map_inputs, out_shape, result, tol=1e-5, target="llvm", device=tvm.cpu()
):
# Late import to avoid breaking test with USE_MICRO=OFF.
from aot.aot_test_utils import AOTTestModel, AOT_DEFAULT_RUNNER, compile_and_run
interface_api = "packed"
use_unpacked_api = False
test_runner = AOT_DEFAULT_RUNNER
compile_and_run(
AOTTestModel(module=mod, inputs=map_inputs, outputs=[result]),
test_runner,
interface_api,
use_unpacked_api,
)
def test_empty_function():
ORIGINAL_MODEL = """
#[version = "0.0.5"]
def @main(%data : Tensor[(16, 29), int8]) -> Tensor[(16, 29), int8] {
add(%data, %data)
}
"""
CMSISNN_MODEL = """
#[version = "0.0.5"]
def @tvmgen_default_cmsis_nn_main_1(%i1: Tensor[(16, 29), int8], Inline=1, Compiler="cmsis-nn", global_symbol="tvmgen_default_cmsis_nn_main_1", Primitive=1) -> Tensor[(16, 29), int8] {
add(%i1, %i1)
}
def @main(%data : Tensor[(16, 29), int8]) -> Tensor[(16, 29), int8] {
%1 = @tvmgen_default_cmsis_nn_main_1(%data) /* ty=Tensor[(16, 29), int8] */;
%1
}
"""
orig_mod = tvm.parser.fromtext(ORIGINAL_MODEL)
cmsisnn_mod = tvm.parser.fromtext(CMSISNN_MODEL)
params = {}
# validate the output
interface_api = "c"
use_unpacked_api = True
test_runner = AOT_USMP_CORSTONE300_RUNNER
dtype = "int8"
in_min, in_max = get_range_for_dtype_str(dtype)
rng = np.random.default_rng(12345)
inputs = {"data": rng.integers(in_min, high=in_max, size=(16, 29), dtype=dtype)}
outputs = generate_ref_data(orig_mod["main"], inputs, params)
compile_and_run(
AOTTestModel(
module=cmsisnn_mod,
inputs=inputs,
outputs=outputs,
params=params,
output_tolerance=0,
),
test_runner,
interface_api,
use_unpacked_api,
)
def test_cnn_small():
# download the model
base_url = "https://github.com/ARM-software/ML-zoo/raw/48a22ee22325d15d2371a6df24eb7d67e21dcc97/models/keyword_spotting/cnn_small/tflite_int8"
file_to_download = "cnn_s_quantized.tflite"
file_saved = "cnn_s_quantized_15Dec2021.tflite"
model_file = download_testdata("{}/{}".format(base_url, file_to_download), file_saved)
with open(model_file, "rb") as f:
tflite_model_buf = f.read()
input_shape = (1, 490)
dtype = "int8"
in_min, in_max = get_range_for_dtype_str(dtype)
rng = np.random.default_rng(12345)
input_data = rng.integers(in_min, high=in_max, size=input_shape, dtype=dtype)
orig_mod, params = convert_to_relay(tflite_model_buf, input_data, "input")
cmsisnn_mod = cmsisnn.partition_for_cmsisnn(orig_mod, params)
# validate CMSIS-NN output against CPU output
interface_api = "c"
use_unpacked_api = True
test_runner = AOT_CORSTONE300_RUNNER
inputs = {"input": input_data}
params = {}
output_list = generate_ref_data(orig_mod["main"], inputs, params)
compile_and_run(
AOTTestModel(
module=cmsisnn_mod,
inputs=inputs,
outputs=output_list,
params=params,
output_tolerance=1,
),
test_runner,
interface_api,
use_unpacked_api,
)
def test_cnn_small():
# download the model
base_url = "https://github.com/ARM-software/ML-zoo/raw/ee35139af86bdace5e502b09fe8b9da9cb1f06bb/models/keyword_spotting/cnn_small/tflite_int8"
file_to_download = "cnn_s_quantized.tflite"
model_file = download_testdata("{}/{}".format(base_url, file_to_download),
file_to_download)
with open(model_file, "rb") as f:
tflite_model_buf = f.read()
input_shape = (1, 490)
rng = np.random.default_rng(12345)
input_data = rng.random(input_shape, dtype=np.float32)
orig_mod, params = convert_to_relay(tflite_model_buf, input_data, "input")
cmsisnn_mod = cmsisnn.partition_for_cmsisnn(orig_mod, params)
# validate CMSIS-NN output against CPU output
interface_api = "c"
use_unpacked_api = True
test_runner = AOT_CORSTONE300_RUNNER
inputs = {"input": input_data}
params = {}
output_list = generate_ref_data(orig_mod["main"], inputs, params)
compile_and_run(
AOTTestModel(
module=cmsisnn_mod,
inputs=inputs,
outputs=output_list,
params=params,
output_tolerance=1,
),
test_runner,
interface_api,
use_unpacked_api,
)
def test_conv2d_int8_tflite(ifm_shape, kernel_shape, strides, dilation,
padding, activation):
interface_api = "c"
use_unpacked_api = True
test_runner = AOT_USMP_CORSTONE300_RUNNER
dtype = "int8"
from tvm.relay.testing.tflite import TFLiteModel
tfl_model = TFLiteModel(dtype)
conv2d_function = tfl_model.create_conv2d_single(kernel_shape, strides,
padding, dilation,
activation)
tfl_model.create_tflite_model(conv2d_function, [ifm_shape])
relay_mod, relay_params = tfl_model.convert_to_relay()
cmsisnn_mod = cmsisnn.partition_for_cmsisnn(relay_mod, relay_params)
# validate pattern matching
assert_partitioned_function(relay_mod, cmsisnn_mod)
# validate CMSIS-NN output against TFLite output
input_map, output_map, output_tolerance = tfl_model.generate_reference_data(
)
compile_and_run(
AOTTestModel(
module=cmsisnn_mod,
inputs=input_map,
outputs=output_map,
params=relay_params,
output_tolerance=output_tolerance,
),
test_runner,
interface_api,
use_unpacked_api,
)
def test_depthwise_int8(
ifm_shape,
kernel_size,
padding,
strides,
dilation,
enable_bias,
relu_type,
input_zero_point,
input_scale,
kernel_scale,
out_channels,
depth_multiplier,
):
interface_api = "c"
use_unpacked_api = True
test_runner = AOT_CORSTONE300_RUNNER
dtype = "int8"
groups = 1
weight_format = "HWIO"
kernel_h = kernel_size[0]
kernel_w = kernel_size[1]
kernel_shape = (kernel_h, kernel_w, ifm_shape[3] // groups, out_channels)
kernel_zero_point = 0
in_min, in_max = get_range_for_dtype_str(dtype)
groups = ifm_shape[3]
weight_format = "HWOI"
kernel_shape = (kernel_h, kernel_w, ifm_shape[3], depth_multiplier)
out_channels = ifm_shape[3] * depth_multiplier
ks_len = len(kernel_scale)
kernel_scale = [kernel_scale[i % ks_len] for i in range(out_channels)]
output_scale, output_zero_point = get_conv2d_qnn_params(
kernel_shape,
input_scale,
input_zero_point,
kernel_scale,
kernel_zero_point,
dtype,
dtype,
dtype,
True,
)
model, params = make_model(
ifm_shape,
kernel_shape,
input_zero_point,
input_scale,
kernel_zero_point,
kernel_scale,
output_zero_point,
output_scale,
padding,
strides,
dilation,
groups,
dtype,
dtype,
out_channels,
weight_format,
enable_bias,
relu_type,
)
orig_mod = make_module(model)
cmsisnn_mod = cmsisnn.partition_for_cmsisnn(orig_mod, params)
# validate pattern matching
attrs = [
cmsisnn_mod[var.name_hint].attrs
for var in cmsisnn_mod.get_global_vars()
if cmsisnn_mod[var.name_hint].attrs
]
assert any(attrs), "At least one function with external attributes was expected."
compilers = [
key == "Compiler" and value == "cmsis-nn" for attr in attrs for key, value in attr.items()
]
assert any(compilers), "Module does not contain function for cmsis-nn target."
assert count_num_calls(orig_mod) == count_num_calls(
cmsisnn_mod
), "Number of calls changed during partitioning"
# validate the output
rng = np.random.default_rng(12345)
inputs = {"input": rng.integers(in_min, high=in_max, size=ifm_shape, dtype=dtype)}
output_list = generate_ref_data(orig_mod["main"], inputs, params)
compile_and_run(
AOTTestModel(
module=cmsisnn_mod,
inputs=inputs,
outputs=output_list,
params=params,
output_tolerance=1,
),
test_runner,
interface_api,
use_unpacked_api,
)
def test_op_int8(
in_shape,
enable_bias,
input_zero_point,
input_scale,
kernel_scale,
out_channels,
relu_type,
):
interface_api = "c"
use_unpacked_api = True
test_runner = AOT_USMP_CORSTONE300_RUNNER
dtype = "int8"
kernel_zero_point = 0
kernel_shape = [out_channels, in_shape[1]]
conv2d_kernel_shape = (1, 1, kernel_shape[0], kernel_shape[1])
in_min, in_max = get_range_for_dtype_str(dtype)
output_scale, output_zero_point = get_conv2d_qnn_params(
conv2d_kernel_shape,
input_scale,
input_zero_point,
kernel_scale,
kernel_zero_point,
dtype,
)
model, params = make_model(
in_shape,
kernel_shape,
input_zero_point,
kernel_zero_point,
input_scale,
kernel_scale,
output_zero_point,
output_scale,
dtype,
dtype,
out_channels,
enable_bias,
)
orig_mod = make_module(model)
cmsisnn_mod = cmsisnn.partition_for_cmsisnn(orig_mod, params)
# validate pattern matching
assert_partitioned_function(orig_mod, cmsisnn_mod)
# validate the output
rng = np.random.default_rng(12345)
inputs = {
"input": rng.integers(in_min, high=in_max, size=in_shape, dtype=dtype)
}
output_list = generate_ref_data(orig_mod["main"], inputs, params)
compile_and_run(
AOTTestModel(
module=cmsisnn_mod,
inputs=inputs,
outputs=output_list,
params=params,
output_tolerance=1,
),
test_runner,
interface_api,
use_unpacked_api,
)
def test_op_int8(
in_shape,
enable_bias,
input_zero_point,
input_scale,
kernel_scale,
out_channels,
relu_type,
):
interface_api = "c"
use_unpacked_api = True
test_runner = AOT_CORSTONE300_RUNNER
dtype = "int8"
kernel_zero_point = 0
kernel_shape = [out_channels, in_shape[1]]
conv2d_kernel_shape = (1, 1, kernel_shape[0], kernel_shape[1])
in_min, in_max = get_range_for_dtype_str(dtype)
output_scale, output_zero_point = get_conv2d_qnn_params(
conv2d_kernel_shape,
input_scale,
input_zero_point,
kernel_scale,
kernel_zero_point,
dtype,
)
model, params = make_model(
in_shape,
kernel_shape,
input_zero_point,
kernel_zero_point,
input_scale,
kernel_scale,
output_zero_point,
output_scale,
dtype,
dtype,
out_channels,
enable_bias,
)
orig_mod = make_module(model)
cmsisnn_mod = cmsisnn.partition_for_cmsisnn(orig_mod, params)
# validate pattern matching
attrs = [
cmsisnn_mod[var.name_hint].attrs
for var in cmsisnn_mod.get_global_vars()
if cmsisnn_mod[var.name_hint].attrs
]
assert any(
attrs), "At least one function with external attributes was expected."
compilers = [
key == "Compiler" and value == "cmsis-nn" for attr in attrs
for key, value in attr.items()
]
assert any(
compilers), "Module does not contain function for cmsisnn target."
assert count_num_calls(orig_mod) == count_num_calls(
cmsisnn_mod), "Number of calls changed during partitioning"
# validate the output
rng = np.random.default_rng(12345)
inputs = {
"input": rng.integers(in_min, high=in_max, size=in_shape, dtype=dtype)
}
output_list = generate_ref_data(orig_mod["main"], inputs, params)
compile_and_run(
AOTTestModel(
module=cmsisnn_mod,
inputs=inputs,
outputs=output_list,
params=params,
output_tolerance=1,
),
test_runner,
interface_api,
use_unpacked_api,
)
def test_depthwise_int8(
ifm_shape,
kernel_size,
padding,
strides,
dilation,
enable_bias,
relu_type,
input_zero_point,
input_scale,
kernel_scale,
out_channels,
depth_multiplier,
):
interface_api = "c"
use_unpacked_api = True
test_runner = AOT_CORSTONE300_RUNNER
dtype = "int8"
groups = 1
weight_format = "HWIO"
kernel_h = kernel_size[0]
kernel_w = kernel_size[1]
kernel_shape = (kernel_h, kernel_w, ifm_shape[3] // groups, out_channels)
kernel_zero_point = 0
in_min, in_max = get_range_for_dtype_str(dtype)
groups = ifm_shape[3]
weight_format = "HWOI"
kernel_shape = (kernel_h, kernel_w, ifm_shape[3], depth_multiplier)
out_channels = ifm_shape[3] * depth_multiplier
ks_len = len(kernel_scale)
kernel_scale = [kernel_scale[i % ks_len] for i in range(out_channels)]
output_scale, output_zero_point = get_conv2d_qnn_params(
kernel_shape,
input_scale,
input_zero_point,
kernel_scale,
kernel_zero_point,
dtype,
dtype,
dtype,
True,
)
model, params = make_model(
ifm_shape,
kernel_shape,
input_zero_point,
input_scale,
kernel_zero_point,
kernel_scale,
output_zero_point,
output_scale,
padding,
strides,
dilation,
groups,
dtype,
dtype,
out_channels,
weight_format,
enable_bias,
relu_type,
)
orig_mod = make_module(model)
cmsisnn_mod = cmsisnn.partition_for_cmsisnn(orig_mod, params)
# validate pattern matching
assert_partitioned_function(orig_mod, cmsisnn_mod)
# validate the output
rng = np.random.default_rng(12345)
inputs = {"input": rng.integers(in_min, high=in_max, size=ifm_shape, dtype=dtype)}
output_list = generate_ref_data(orig_mod["main"], inputs, params)
compile_and_run(
AOTTestModel(
module=cmsisnn_mod,
inputs=inputs,
outputs=output_list,
params=params,
output_tolerance=1,
),
test_runner,
interface_api,
use_unpacked_api,
)
