test_topi_depthwise_conv2d.py 18.4 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
# Licensed to the Apache Software Foundation (ASF) under one
# or more contributor license agreements.  See the NOTICE file
# distributed with this work for additional information
# regarding copyright ownership.  The ASF licenses this file
# to you under the Apache License, Version 2.0 (the
# "License"); you may not use this file except in compliance
# with the License.  You may obtain a copy of the License at
#
#   http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing,
# software distributed under the License is distributed on an
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
# KIND, either express or implied.  See the License for the
# specific language governing permissions and limitations
# under the License.
17
import tvm
18
from tvm import autotvm
19
import topi
20
import topi.testing
21 22
import numpy as np
from topi.util import get_const_tuple
23
from topi.nn.util import get_pad_tuple
24
from tvm.contrib.pickle_memoize import memoize
25

26
from common import get_all_backend
27

28
def depthwise_conv2d_with_workload_nchw(batch, in_channel, in_height, channel_multiplier, filter_height, stride, padding, dilation=1):
29 30 31
    in_width = in_height
    filter_channel = in_channel
    filter_width = filter_height
32 33 34 35 36 37 38 39 40 41
    stride_h = stride_w = stride

    if dilation == 1:
        # here we transform the padding argument from 'str' to  'tuple' ,
        # because we need this to match the "workload" tuple to the records in TopHub
        pad_h, pad_w, _, _ = get_pad_tuple(padding, (filter_height, filter_width))
        padding_args = (pad_h, pad_w)
    else:
        padding_args = padding

42 43 44 45 46
    # placeholder
    Input = tvm.placeholder((batch, in_channel, in_height, in_width), name='Input')
    Filter = tvm.placeholder((filter_channel, channel_multiplier, filter_height, filter_width), name='Filter')
    Scale = tvm.placeholder((in_channel * channel_multiplier,), name='Scale')
    Shift = tvm.placeholder((in_channel * channel_multiplier,), name='Shift')
47

48 49
    dtype = 'float32'

50
    def check_device(device):
51 52
        ctx = tvm.context(device, 0)
        if not ctx.exist:
53 54
            print("Skip because %s is not enabled" % device)
            return
55
        print("Running on target: %s" % device)
56
        with tvm.target.create(device):
57
            # declare
58 59
            DepthwiseConv2d = topi.nn.depthwise_conv2d_nchw(Input, Filter,
                (stride_h, stride_w), padding_args, dilation, dtype)
60 61
            ScaleShift = topi.nn.scale_shift_nchw(DepthwiseConv2d, Scale, Shift)
            Relu = topi.nn.relu(ScaleShift)
62 63 64 65
            # schedule
            s1 = topi.generic.schedule_depthwise_conv2d_nchw(DepthwiseConv2d)
            s2 = topi.generic.schedule_depthwise_conv2d_nchw(ScaleShift)
            s3 = topi.generic.schedule_depthwise_conv2d_nchw(Relu)
66 67 68 69
        # build the kernels
        f1 = tvm.build(s1, [Input, Filter, DepthwiseConv2d], device)
        f2 = tvm.build(s2, [Input, Filter, Scale, Shift, ScaleShift], device)
        f3 = tvm.build(s3, [Input, Filter, Scale, Shift, Relu], device)
70 71 72 73 74 75 76 77 78 79 80 81 82

        # Prepare pod type for test data closure
        input_shape = get_const_tuple(Input.shape)
        filter_shape = get_const_tuple(Filter.shape)
        scale_shape = get_const_tuple(Scale.shape)
        shift_shape = get_const_tuple(Shift.shape)
        scale_shift_shape = get_const_tuple(ScaleShift.shape)

        # Use memoize, pickle the test data for next time use.
        @memoize("topi.tests.test_topi_depthwise_conv2d.nchw")
        def get_ref_data():
            input_np = np.random.uniform(size=input_shape).astype(dtype)
            filter_np = np.random.uniform(size=filter_shape).astype(dtype)
83
            dilated_filter_np = topi.testing.dilate_python(filter_np, (1, 1, dilation, dilation))
84 85 86 87
            scale_np = np.random.uniform(size=scale_shape).astype(dtype)
            shift_np = np.random.uniform(size=shift_shape).astype(dtype)
            # correctness with scipy
            depthwise_conv2d_scipy = topi.testing.depthwise_conv2d_python_nchw(
88
                input_np, dilated_filter_np, stride, padding)
89 90 91 92 93 94 95 96 97 98
            scale_shift_scipy = np.zeros(shape=scale_shift_shape)
            for c in range(in_channel * channel_multiplier):
                scale_shift_scipy[:,c,:,:] = depthwise_conv2d_scipy[:,c,:,:] * scale_np[c] + shift_np[c]
                relu_scipy = np.maximum(scale_shift_scipy, 0)
            return (input_np, filter_np, scale_np, shift_np,
                    depthwise_conv2d_scipy, scale_shift_scipy, relu_scipy)
        # Get the test data
        (input_np, filter_np, scale_np, shift_np,
         depthwise_conv2d_scipy, scale_shift_scipy, relu_scipy) = get_ref_data()

99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114
        input_tvm = tvm.nd.array(input_np, ctx)
        filter_tvm = tvm.nd.array(filter_np, ctx)
        scale_tvm = tvm.nd.array(scale_np, ctx)
        shift_tvm = tvm.nd.array(shift_np, ctx)
        depthwise_conv2d_tvm = tvm.nd.array(np.zeros(shape=get_const_tuple(DepthwiseConv2d.shape), dtype=DepthwiseConv2d.dtype), ctx)
        scale_shift_tvm = tvm.nd.array(np.zeros(shape=get_const_tuple(ScaleShift.shape), dtype=ScaleShift.dtype), ctx)
        relu_tvm = tvm.nd.array(np.zeros(shape=get_const_tuple(Relu.shape), dtype=Relu.dtype), ctx)
        # launch kernel 1 (depthwise_conv2d)
        timer_1 = f1.time_evaluator(f1.entry_name, ctx, number=1)
        tcost_1 = timer_1(input_tvm, filter_tvm, depthwise_conv2d_tvm).mean
        # launch kernel 2 (depthwise_conv2d + scale_shift)
        timer_2 = f2.time_evaluator(f2.entry_name, ctx, number=1)
        tcost_2 = timer_2(input_tvm, filter_tvm, scale_tvm, shift_tvm, scale_shift_tvm).mean
        # launch kernel 3 (depthwise_conv2d + scale_shift + relu)
        timer_3 = f3.time_evaluator(f3.entry_name, ctx, number=1)
        tcost_3 = timer_3(input_tvm, filter_tvm, scale_tvm, shift_tvm, relu_tvm).mean
115 116 117
        tvm.testing.assert_allclose(depthwise_conv2d_tvm.asnumpy(), depthwise_conv2d_scipy, rtol=1e-5)
        tvm.testing.assert_allclose(scale_shift_tvm.asnumpy(), scale_shift_scipy, rtol=1e-5)
        tvm.testing.assert_allclose(relu_tvm.asnumpy(), relu_scipy, rtol=1e-5)
118

119
    for device in get_all_backend():
120 121
        with autotvm.tophub.context(device):  # load tophub pre-tuned parameters
            check_device(device)
122

123

124
def depthwise_conv2d_with_workload_nhwc(batch, in_channel, in_height, channel_multiplier, filter_height, stride_h, padding, dilation=1):
125 126 127 128
    in_width = in_height
    filter_channel = in_channel
    filter_width = filter_height
    stride_w = stride_h
129 130 131 132 133 134 135 136 137

    if dilation == 1:
        # here we transform the padding argument from 'str' to  'tuple' ,
        # because we need this to match the "workload" tuple to the records in TopHub
        pad_h, pad_w, _, _ = get_pad_tuple(padding, (filter_height, filter_width))
        padding_args = (pad_h, pad_w)
    else:
        padding_args = padding

138 139 140 141 142
    # placeholder
    Input = tvm.placeholder((batch, in_height, in_width, in_channel), name='Input')
    Filter = tvm.placeholder((filter_height, filter_width,filter_channel, channel_multiplier), name='Filter')
    Scale = tvm.placeholder((in_channel * channel_multiplier,), name='Scale')
    Shift = tvm.placeholder((in_channel * channel_multiplier,), name='Shift')
143

144 145
    dtype = 'float32'

146
    def check_device(device):
147 148
        ctx = tvm.context(device, 0)
        if not ctx.exist:
149 150
            print("Skip because %s is not enabled" % device)
            return
151
        print("Running on target: %s" % device)
152 153

        with tvm.target.create(device):
154
            # declare
155 156
            DepthwiseConv2d = topi.nn.depthwise_conv2d_nhwc(Input, Filter,
                (stride_h, stride_w), padding_args, dilation, dtype)
157 158 159
            ScaleShift = topi.nn.scale_shift_nhwc(DepthwiseConv2d, Scale, Shift)
            Relu = topi.nn.relu(ScaleShift)
            # schedule
160 161 162
            s1 = topi.generic.schedule_depthwise_conv2d_nhwc(DepthwiseConv2d)
            s2 = topi.generic.schedule_depthwise_conv2d_nhwc(ScaleShift)
            s3 = topi.generic.schedule_depthwise_conv2d_nhwc(Relu)
163 164 165 166
        # build the kernels
        f1 = tvm.build(s1, [Input, Filter, DepthwiseConv2d], device)
        f2 = tvm.build(s2, [Input, Filter, Scale, Shift, ScaleShift], device)
        f3 = tvm.build(s3, [Input, Filter, Scale, Shift, Relu], device)
167 168 169 170 171 172 173 174 175

        # Prepare pod type for test data closure
        input_shape = get_const_tuple(Input.shape)
        filter_shape = get_const_tuple(Filter.shape)
        scale_shape = get_const_tuple(Scale.shape)
        shift_shape = get_const_tuple(Shift.shape)
        scale_shift_shape = get_const_tuple(ScaleShift.shape)

        # Use memoize, pickle the test data for next time use.
176
        @memoize("topi.tests.test_topi_depthwise_conv2d.nhwc.v2")
177 178 179
        def get_ref_data():
            input_np = np.random.uniform(size=input_shape).astype(dtype)
            filter_np = np.random.uniform(size=filter_shape).astype(dtype)
180
            dilated_filter_np = topi.testing.dilate_python(filter_np, (dilation, dilation, 1, 1))
181 182 183 184
            scale_np = np.random.uniform(size=scale_shape).astype(dtype)
            shift_np = np.random.uniform(size=shift_shape).astype(dtype)
            # correctness with scipy
            depthwise_conv2d_scipy = topi.testing.depthwise_conv2d_python_nhwc(
185
                input_np, dilated_filter_np, stride=[stride_h, stride_w], padding=padding)
186 187 188 189 190 191 192 193 194 195
            scale_shift_scipy = np.zeros(shape=scale_shift_shape)
            for c in range(in_channel * channel_multiplier):
                scale_shift_scipy[:,:,:,c] = depthwise_conv2d_scipy[:,:,:,c] * scale_np[c] + shift_np[c]
                relu_scipy = np.maximum(scale_shift_scipy, 0)
            return (input_np, filter_np, scale_np, shift_np,
                    depthwise_conv2d_scipy, scale_shift_scipy, relu_scipy)
        # Get the test data
        (input_np, filter_np, scale_np, shift_np,
         depthwise_conv2d_scipy, scale_shift_scipy, relu_scipy) = get_ref_data()

196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213
        # prepare data
        input_tvm = tvm.nd.array(input_np, ctx)
        filter_tvm = tvm.nd.array(filter_np, ctx)
        scale_tvm = tvm.nd.array(scale_np, ctx)
        shift_tvm = tvm.nd.array(shift_np, ctx)
        depthwise_conv2d_tvm = tvm.nd.array(np.zeros(shape=get_const_tuple(DepthwiseConv2d.shape), dtype=DepthwiseConv2d.dtype), ctx)
        scale_shift_tvm = tvm.nd.array(np.zeros(shape=get_const_tuple(ScaleShift.shape), dtype=ScaleShift.dtype), ctx)
        relu_tvm = tvm.nd.array(np.zeros(shape=get_const_tuple(Relu.shape), dtype=Relu.dtype), ctx)
        # launch kernel 1 (depthwise_conv2d)
        timer_1 = f1.time_evaluator(f1.entry_name, ctx, number=1)
        tcost_1 = timer_1(input_tvm, filter_tvm, depthwise_conv2d_tvm).mean
        # launch kernel 2 (depthwise_conv2d + scale_shift)
        timer_2 = f2.time_evaluator(f2.entry_name, ctx, number=1)
        tcost_2 = timer_2(input_tvm, filter_tvm, scale_tvm, shift_tvm, scale_shift_tvm).mean
        # launch kernel 3 (depthwise_conv2d + scale_shift + relu)
        timer_3 = f3.time_evaluator(f3.entry_name, ctx, number=1)
        tcost_3 = timer_3(input_tvm, filter_tvm, scale_tvm, shift_tvm, relu_tvm).mean
        relu_scipy = np.maximum(scale_shift_scipy, 0)
214 215 216
        tvm.testing.assert_allclose(depthwise_conv2d_tvm.asnumpy(), depthwise_conv2d_scipy, rtol=1e-5)
        tvm.testing.assert_allclose(scale_shift_tvm.asnumpy(), scale_shift_scipy, rtol=1e-5)
        tvm.testing.assert_allclose(relu_tvm.asnumpy(), relu_scipy, rtol=1e-5)
217

218
    for device in get_all_backend():
219 220
        with autotvm.tophub.context(device):  # load tophub pre-tuned parameters
            check_device(device)
221

222 223 224 225 226 227 228 229 230 231 232 233 234
def _transform_data(data, bn):
    # NCHW -> NCHW[x]c
    batch_size, channel, height, width = data.shape
    data = np.reshape(data, (batch_size, channel//bn, bn, height, width))
    data = np.transpose(data, (0, 1, 3, 4, 2))
    return data

def _transform_kernel(kernel, bn):
    # channel, channel_multiplier, kh, kw -> out_channel_chunk, kh, kw, out_channel_block
    channel, channel_multiplier, kh, kw = kernel.shape
    out_channel = channel * channel_multiplier
    kernel = np.reshape(kernel, (out_channel//bn, bn, kh, kw))
    kernel = np.transpose(kernel, (0, 2, 3, 1))
235 236
    out_channel_chunk, kh, kw, out_channel_block = kernel.shape
    return kernel.reshape(out_channel_chunk, 1, kh, kw, 1, out_channel_block)
237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265

def depthwise_conv2d_with_workload_NCHWc(batch, in_channel, in_height, channel_multiplier, filter_height, stride, padding, dilation=1):
    in_width = in_height
    filter_channel = in_channel
    filter_width = filter_height
    stride_h = stride_w = stride

    assert dilation == 1, "depthwise_conv2d_NCHWc currently does not support dilation."
    pad_h, pad_w, _, _ = get_pad_tuple(padding, (filter_height, filter_width))
    padding_args = (pad_h, pad_w)

    out_channel = filter_channel * channel_multiplier
    # for testing functionality,
    # we choose arbitrary block size that can divide the channel,
    # regardless of the performance.
    oc_block = 1
    for bn in range(16, 0, -1):
        if out_channel % bn == 0:
            oc_block = bn
            break

    ic_block = 1
    for bn in range(oc_block, 0, -1):
        if in_channel % bn == 0:
            ic_block = bn
            break

    # placeholder
    Input = tvm.placeholder((batch, in_channel//ic_block, in_height, in_width, ic_block), name='Input')
266
    Filter = tvm.placeholder((out_channel//oc_block, 1, filter_height, filter_width, 1, oc_block), name='Filter')
267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316
    in_layout = "NCHW%dc" % ic_block
    out_layout = "NCHW%dc" % oc_block
    dtype = 'float32'

    def check_device(device):
        ctx = tvm.context(device, 0)
        if not ctx.exist:
            print("Skip because %s is not enabled" % device)
            return
        print("Running on target: %s" % device)
        with tvm.target.create(device):
            # declare
            DepthwiseConv2d = topi.nn.depthwise_conv2d_NCHWc(Input, Filter,
                                                             (stride_h, stride_w),
                                                             padding_args,
                                                             (dilation, dilation),
                                                             in_layout,
                                                             out_layout, dtype)
            # TODO: add scale_shift implement for NCHWc and add test here
            Relu = topi.nn.relu(DepthwiseConv2d)
            # schedule
            s1 = topi.generic.schedule_depthwise_conv2d_nchw(DepthwiseConv2d)
            s2 = topi.generic.schedule_depthwise_conv2d_nchw(Relu)
        # build the kernels
        f1 = tvm.build(s1, [Input, Filter, DepthwiseConv2d], device)
        f2 = tvm.build(s2, [Input, Filter, Relu], device)

        # Prepare pod type for test data closure
        input_shape = (batch, in_channel, in_height, in_width)
        filter_shape = (filter_channel, channel_multiplier, filter_height, filter_width)

        # Use memoize, pickle the test data for next time use.
        @memoize("topi.tests.test_topi_depthwise_conv2d.NCHWc")
        def get_ref_data():
            input_np = np.random.uniform(size=input_shape).astype(dtype)
            filter_np = np.random.uniform(size=filter_shape).astype(dtype)
            # correctness with scipy
            depthwise_conv2d_scipy = topi.testing.depthwise_conv2d_python_nchw(
                input_np, filter_np, stride, padding)
            relu_scipy = np.maximum(depthwise_conv2d_scipy, 0)
            return (_transform_data(input_np, ic_block),
                    _transform_kernel(filter_np, oc_block),
                    _transform_data(depthwise_conv2d_scipy, oc_block),
                    _transform_data(relu_scipy, oc_block))

        # Get the test data
        (input_np, filter_np, depthwise_conv2d_scipy, relu_scipy) = get_ref_data()

        input_tvm = tvm.nd.array(input_np, ctx)
        filter_tvm = tvm.nd.array(filter_np, ctx)
317

318 319 320 321
        depthwise_conv2d_tvm = tvm.nd.array(np.zeros(shape=get_const_tuple(DepthwiseConv2d.shape),
                                                     dtype=DepthwiseConv2d.dtype), ctx)
        relu_tvm = tvm.nd.array(np.zeros(shape=get_const_tuple(Relu.shape), dtype=Relu.dtype), ctx)
        # launch kernel 1 (depthwise_conv2d)
322
        print(filter_tvm.shape)
323 324 325 326 327 328 329 330 331 332 333
        f1(input_tvm, filter_tvm, depthwise_conv2d_tvm)
        # launch kernel 2 (depthwise_conv2d + relu)
        f2(input_tvm, filter_tvm, relu_tvm)
        tvm.testing.assert_allclose(depthwise_conv2d_tvm.asnumpy(), depthwise_conv2d_scipy, rtol=1e-5)
        tvm.testing.assert_allclose(relu_tvm.asnumpy(), relu_scipy, rtol=1e-5)

    # test llvm only for now since depthwise_conv2d_NCHWc implement is missing in other backend.
    for device in ["llvm"]:
        with autotvm.tophub.context(device):  # load tophub pre-tuned parameters
            check_device(device)

334 335

def test_depthwise_conv2d():
336 337 338 339 340 341 342 343 344 345 346 347
    # mobilenet workloads
    depthwise_conv2d_with_workload_nchw(1, 32, 112, 1, 3, 1, "SAME")
    depthwise_conv2d_with_workload_nchw(1, 64, 112, 1, 3, 2, "SAME")
    depthwise_conv2d_with_workload_nchw(1, 128, 56, 1, 3, 1, "SAME")
    depthwise_conv2d_with_workload_nchw(1, 128, 56, 1, 3, 2, "SAME")
    depthwise_conv2d_with_workload_nchw(1, 256, 28, 1, 3, 1, "SAME")
    depthwise_conv2d_with_workload_nchw(1, 256, 28, 1, 3, 2, "SAME")
    depthwise_conv2d_with_workload_nchw(1, 512, 14, 1, 3, 1, "SAME")
    depthwise_conv2d_with_workload_nchw(1, 512, 14, 1, 3, 2, "SAME")
    depthwise_conv2d_with_workload_nchw(1, 1024, 7, 1, 3, 1, "SAME")

    # NCHW
348 349 350 351
    depthwise_conv2d_with_workload_nchw(1, 728, 32, 1, 3, 1, "SAME")
    depthwise_conv2d_with_workload_nchw(4, 256, 64, 2, 5, 2, "SAME")
    depthwise_conv2d_with_workload_nchw(1, 728, 32, 1, 3, 1, "VALID")
    depthwise_conv2d_with_workload_nchw(4, 256, 64, 2, 5, 2, "VALID")
352 353
    # dilation = 2
    depthwise_conv2d_with_workload_nchw(1, 728, 64, 1, 3, 1, "SAME", dilation=2)
354 355

    # NHWC
356 357 358 359
    depthwise_conv2d_with_workload_nhwc(1, 728, 32, 1, 3, 1, "SAME")
    depthwise_conv2d_with_workload_nhwc(4, 256, 64, 2, 5, 2, "SAME")
    depthwise_conv2d_with_workload_nhwc(1, 728, 32, 1, 3, 1, "VALID")
    depthwise_conv2d_with_workload_nhwc(4, 256, 64, 2, 5, 2, "VALID")
360
    # dilation = 2
361 362
    # disabled because it uses too large shared memory on cuda
    # depthwise_conv2d_with_workload_nhwc(1, 728, 64, 1, 3, 1, "SAME", dilation=2)
363

364 365 366 367 368 369 370
    # NCHW[x]c
    depthwise_conv2d_with_workload_NCHWc(1, 728, 32, 1, 3, 1, "SAME")
    depthwise_conv2d_with_workload_NCHWc(4, 256, 64, 2, 5, 2, "SAME")
    depthwise_conv2d_with_workload_NCHWc(1, 728, 32, 1, 3, 1, "VALID")
    depthwise_conv2d_with_workload_NCHWc(4, 256, 64, 2, 5, 2, "VALID")


371 372
if __name__ == "__main__":
    test_depthwise_conv2d()