[Relay][Frontend] Add MXNet test example for relay

src/relay/pass/canonicalize_ops.cc

@@ -22,7 +22,7 @@ class BiasAddSimplifier : public ExprMutator {
       CHECK_EQ(call->args.size(), 2);
       const BiasAddAttrs* param = call->attrs.as<BiasAddAttrs>();
 
-      auto ttype = call->args[0]->type_as<TensorTypeNode>();
+      auto ttype = n->args[0]->type_as<TensorTypeNode>();
       size_t n_dim = ttype->shape.size();
       Expr expanded_bias = ExpandBiasToMatchAxis(call->args[1], n_dim, {param->axis});
       Expr ret = Add(call->args[0], expanded_bias);

tests/python/relay/frontend/mxnet/model_zoo/__init__.py

@@ -0,0 +1,46 @@
+"""MXNet and Relay model zoo."""
+from __future__ import absolute_import
+from . import mlp, resnet, vgg, dqn, dcgan, squeezenet, inception_v3
+import tvm.relay.testing
+
+_num_class = 1000
+_batch = 2
+
+# mlp fc
+mx_mlp = mlp.get_symbol(_num_class)
+relay_mlp = tvm.relay.testing.mlp.get_workload(_batch, _num_class)[0]
+
+# vgg fc
+mx_vgg = {}
+relay_vgg = {}
+for num_layers in [11, 13, 16, 19]:
+    mx_vgg[num_layers] = vgg.get_symbol(_num_class, num_layers)
+    relay_vgg[num_layers] = tvm.relay.testing.vgg.get_workload(
+        _batch, _num_class, num_layers=num_layers)[0]
+
+# resnet fc
+mx_resnet = {}
+relay_resnet = {}
+for num_layers in [18, 34, 50, 101, 152, 200, 269]:
+    mx_resnet[num_layers] = resnet.get_symbol(_num_class, num_layers, '3,224,224')
+    relay_resnet[num_layers] = tvm.relay.testing.resnet.get_workload(
+        _batch, _num_class, num_layers=num_layers)[0]
+
+# squeezenet
+mx_squeezenet = {}
+relay_squeezenet = {}
+for version in ['1.0', '1.1']:
+    mx_squeezenet[version] = squeezenet.get_symbol(version=version)
+    relay_squeezenet[version] = tvm.relay.testing.squeezenet.get_workload(_batch, version=version)[0]
+
+# inception
+mx_inception_v3 = inception_v3.get_symbol()
+relay_inception_v3 = tvm.relay.testing.inception_v3.get_workload(_batch)[0]
+
+# dqn
+mx_dqn = dqn.get_symbol()
+relay_dqn = tvm.relay.testing.dqn.get_workload(_batch)[0]
+
+# dcgan generator
+mx_dcgan = dcgan.get_symbol()
+relay_dcgan = tvm.relay.testing.dcgan.get_workload(_batch)[0]

tests/python/relay/frontend/mxnet/model_zoo/dcgan.py

@@ -0,0 +1,66 @@
+# pylint: disable=unused-argument
+"""
+The MXNet symbol of DCGAN generator
+
+Adopted from:
+https://github.com/tqchen/mxnet-gan/blob/master/mxgan/generator.py
+
+Reference:
+Radford, Alec, Luke Metz, and Soumith Chintala.
+"Unsupervised representation learning with deep convolutional generative adversarial networks."
+arXiv preprint arXiv:1511.06434 (2015).
+"""
+
+import mxnet as mx
+
+def deconv2d(data, ishape, oshape, kshape, name, stride=(2, 2)):
+    """a deconv layer that enlarges the feature map"""
+    target_shape = (oshape[-2], oshape[-1])
+    pad_y = (kshape[0] - 1) // 2
+    pad_x = (kshape[1] - 1) // 2
+    adj_y = (target_shape[0] + 2 * pad_y - kshape[0]) % stride[0]
+    adj_x = (target_shape[1] + 2 * pad_x - kshape[1]) % stride[1]
+
+    net = mx.sym.Deconvolution(data,
+                               kernel=kshape,
+                               stride=stride,
+                               pad=(pad_y, pad_x),
+                               adj=(adj_y, adj_x),
+                               num_filter=oshape[0],
+                               no_bias=True,
+                               name=name)
+    return net
+
+def deconv2d_bn_relu(data, prefix, **kwargs):
+    """a block of deconv + batch norm + relu"""
+    eps = 1e-5 + 1e-12
+
+    net = deconv2d(data, name="%s_deconv" % prefix, **kwargs)
+    net = mx.sym.BatchNorm(net, eps=eps, name="%s_bn" % prefix)
+    net = mx.sym.Activation(net, name="%s_act" % prefix, act_type='relu')
+    return net
+
+def get_symbol(oshape=(3, 64, 64), ngf=128, code=None):
+    """get symbol of dcgan generator"""
+    assert oshape[-1] == 64, "Only support 64x64 image"
+    assert oshape[-2] == 64, "Only support 64x64 image"
+
+    code = mx.sym.Variable("data") if code is None else code
+    net = mx.sym.FullyConnected(code, name="g1", num_hidden=ngf*8*4*4, no_bias=True, flatten=False)
+    net = mx.sym.Activation(net, act_type='relu')
+    # 4 x 4
+    net = mx.sym.reshape(net, shape=(-1, ngf * 8, 4, 4))
+    # 8 x 8
+    net = deconv2d_bn_relu(
+        net, ishape=(ngf * 8, 4, 4), oshape=(ngf * 4, 8, 8), kshape=(4, 4), prefix="g2")
+    # 16x16
+    net = deconv2d_bn_relu(
+        net, ishape=(ngf * 4, 8, 8), oshape=(ngf * 2, 16, 16), kshape=(4, 4), prefix="g3")
+    # 32x32
+    net = deconv2d_bn_relu(
+        net, ishape=(ngf * 2, 16, 16), oshape=(ngf, 32, 32), kshape=(4, 4), prefix="g4")
+    # 64x64
+    net = deconv2d(
+        net, ishape=(ngf, 32, 32), oshape=oshape[-3:], kshape=(4, 4), name="g5_deconv")
+    net = mx.sym.Activation(net, act_type='tanh')
+    return net

tests/python/relay/frontend/mxnet/model_zoo/dqn.py

@@ -0,0 +1,27 @@
+"""
+The mxnet symbol of Nature DQN
+
+Reference:
+Mnih, Volodymyr, et al.
+"Human-level control through deep reinforcement learning."
+Nature 518.7540 (2015): 529.
+"""
+
+import mxnet as mx
+
+def get_symbol(num_action=18):
+    data = mx.sym.Variable(name='data')
+    net = mx.sym.Convolution(data, kernel=(8, 8), stride=(4, 4),
+                             num_filter=32, name='conv1')
+    net = mx.sym.Activation(net, act_type='relu', name='relu1')
+    net = mx.sym.Convolution(net, kernel=(4, 4), stride=(2, 2),
+                             num_filter=64, name='conv2')
+    net = mx.sym.Activation(net, act_type='relu', name='relu2')
+    net = mx.sym.Convolution(net, kernel=(3, 3), stride=(1, 1),
+                             num_filter=64, name='conv3')
+    net = mx.sym.Activation(net, act_type='relu', name='relu3')
+    net = mx.sym.FullyConnected(net, num_hidden=512, name='fc4')
+    net = mx.sym.Activation(net, act_type='relu', name='relu4')
+    net = mx.sym.FullyConnected(net, num_hidden=num_action, name='fc5', flatten=False)
+
+    return net

tests/python/relay/frontend/mxnet/model_zoo/inception_v3.py

@@ -0,0 +1,170 @@
+"""
+Inception V3, suitable for images with around 299 x 299
+
+Reference:
+Szegedy, Christian, et al. "Rethinking the Inception Architecture for Computer Vision." arXiv preprint arXiv:1512.00567 (2015).
+
+Adopted from https://github.com/apache/incubator-mxnet/blob/
+             master/example/image-classification/symbols/inception-v3.py
+"""
+import mxnet as mx
+import numpy as np
+
+def Conv(data, num_filter, kernel=(1, 1), stride=(1, 1), pad=(0, 0), name=None, suffix=''):
+    conv = mx.sym.Convolution(data=data, num_filter=num_filter, kernel=kernel, stride=stride, pad=pad, no_bias=True, name='%s%s_conv2d' %(name, suffix))
+    bn = mx.sym.BatchNorm(data=conv, eps=2e-5, name='%s%s_batchnorm' % (name, suffix))
+    act = mx.sym.Activation(data=bn, act_type='relu', name='%s%s_relu' %(name, suffix))
+    return act
+
+
+def Inception7A(data,
+                num_1x1,
+                num_3x3_red, num_3x3_1, num_3x3_2,
+                num_5x5_red, num_5x5,
+                pool, proj,
+                name):
+    tower_1x1 = Conv(data, num_1x1, name=('%s_conv' % name))
+    tower_5x5 = Conv(data, num_5x5_red, name=('%s_tower' % name), suffix='_conv')
+    tower_5x5 = Conv(tower_5x5, num_5x5, kernel=(5, 5), pad=(2, 2), name=('%s_tower' % name), suffix='_conv_1')
+    tower_3x3 = Conv(data, num_3x3_red, name=('%s_tower_1' % name), suffix='_conv')
+    tower_3x3 = Conv(tower_3x3, num_3x3_1, kernel=(3, 3), pad=(1, 1), name=('%s_tower_1' % name), suffix='_conv_1')
+    tower_3x3 = Conv(tower_3x3, num_3x3_2, kernel=(3, 3), pad=(1, 1), name=('%s_tower_1' % name), suffix='_conv_2')
+    pooling = mx.sym.Pooling(data=data, kernel=(3, 3), stride=(1, 1), pad=(1, 1), pool_type=pool, name=('%s_pool_%s_pool' % (pool, name)))
+    cproj = Conv(pooling, proj, name=('%s_tower_2' %  name), suffix='_conv')
+    concat = mx.sym.Concat(*[tower_1x1, tower_5x5, tower_3x3, cproj], name='ch_concat_%s_chconcat' % name)
+    return concat
+
+# First Downsample
+def Inception7B(data,
+                num_3x3,
+                num_d3x3_red, num_d3x3_1, num_d3x3_2,
+                pool,
+                name):
+    tower_3x3 = Conv(data, num_3x3, kernel=(3, 3), pad=(0, 0), stride=(2, 2), name=('%s_conv' % name))
+    tower_d3x3 = Conv(data, num_d3x3_red, name=('%s_tower' % name), suffix='_conv')
+    tower_d3x3 = Conv(tower_d3x3, num_d3x3_1, kernel=(3, 3), pad=(1, 1), stride=(1, 1), name=('%s_tower' % name), suffix='_conv_1')
+    tower_d3x3 = Conv(tower_d3x3, num_d3x3_2, kernel=(3, 3), pad=(0, 0), stride=(2, 2), name=('%s_tower' % name), suffix='_conv_2')
+    pooling = mx.sym.Pooling(data=data, kernel=(3, 3), stride=(2, 2), pad=(0,0), pool_type="max", name=('max_pool_%s_pool' % name))
+    concat = mx.sym.Concat(*[tower_3x3, tower_d3x3, pooling], name='ch_concat_%s_chconcat' % name)
+    return concat
+
+def Inception7C(data,
+                num_1x1,
+                num_d7_red, num_d7_1, num_d7_2,
+                num_q7_red, num_q7_1, num_q7_2, num_q7_3, num_q7_4,
+                pool, proj,
+                name):
+    tower_1x1 = Conv(data=data, num_filter=num_1x1, kernel=(1, 1), name=('%s_conv' % name))
+    tower_d7 = Conv(data=data, num_filter=num_d7_red, name=('%s_tower' % name), suffix='_conv')
+    tower_d7 = Conv(data=tower_d7, num_filter=num_d7_1, kernel=(1, 7), pad=(0, 3), name=('%s_tower' % name), suffix='_conv_1')
+    tower_d7 = Conv(data=tower_d7, num_filter=num_d7_2, kernel=(7, 1), pad=(3, 0), name=('%s_tower' % name), suffix='_conv_2')
+    tower_q7 = Conv(data=data, num_filter=num_q7_red, name=('%s_tower_1' % name), suffix='_conv')
+    tower_q7 = Conv(data=tower_q7, num_filter=num_q7_1, kernel=(7, 1), pad=(3, 0), name=('%s_tower_1' % name), suffix='_conv_1')
+    tower_q7 = Conv(data=tower_q7, num_filter=num_q7_2, kernel=(1, 7), pad=(0, 3), name=('%s_tower_1' % name), suffix='_conv_2')
+    tower_q7 = Conv(data=tower_q7, num_filter=num_q7_3, kernel=(7, 1), pad=(3, 0), name=('%s_tower_1' % name), suffix='_conv_3')
+    tower_q7 = Conv(data=tower_q7, num_filter=num_q7_4, kernel=(1, 7), pad=(0, 3), name=('%s_tower_1' % name), suffix='_conv_4')
+    pooling = mx.sym.Pooling(data=data, kernel=(3, 3), stride=(1, 1), pad=(1, 1), pool_type=pool, name=('%s_pool_%s_pool' % (pool, name)))
+    cproj = Conv(data=pooling, num_filter=proj, kernel=(1, 1), name=('%s_tower_2' %  name), suffix='_conv')
+    # concat
+    concat = mx.sym.Concat(*[tower_1x1, tower_d7, tower_q7, cproj], name='ch_concat_%s_chconcat' % name)
+    return concat
+
+def Inception7D(data,
+                num_3x3_red, num_3x3,
+                num_d7_3x3_red, num_d7_1, num_d7_2, num_d7_3x3,
+                pool,
+                name):
+    tower_3x3 = Conv(data=data, num_filter=num_3x3_red, name=('%s_tower' % name), suffix='_conv')
+    tower_3x3 = Conv(data=tower_3x3, num_filter=num_3x3, kernel=(3, 3), pad=(0,0), stride=(2, 2), name=('%s_tower' % name), suffix='_conv_1')
+    tower_d7_3x3 = Conv(data=data, num_filter=num_d7_3x3_red, name=('%s_tower_1' % name), suffix='_conv')
+    tower_d7_3x3 = Conv(data=tower_d7_3x3, num_filter=num_d7_1, kernel=(1, 7), pad=(0, 3), name=('%s_tower_1' % name), suffix='_conv_1')
+    tower_d7_3x3 = Conv(data=tower_d7_3x3, num_filter=num_d7_2, kernel=(7, 1), pad=(3, 0), name=('%s_tower_1' % name), suffix='_conv_2')
+    tower_d7_3x3 = Conv(data=tower_d7_3x3, num_filter=num_d7_3x3, kernel=(3, 3), stride=(2, 2), name=('%s_tower_1' % name), suffix='_conv_3')
+    pooling = mx.sym.Pooling(data=data, kernel=(3, 3), stride=(2, 2), pool_type=pool, name=('%s_pool_%s_pool' % (pool, name)))
+    # concat
+    concat = mx.sym.Concat(*[tower_3x3, tower_d7_3x3, pooling], name='ch_concat_%s_chconcat' % name)
+    return concat
+
+def Inception7E(data,
+                num_1x1,
+                num_d3_red, num_d3_1, num_d3_2,
+                num_3x3_d3_red, num_3x3, num_3x3_d3_1, num_3x3_d3_2,
+                pool, proj,
+                name):
+    tower_1x1 = Conv(data=data, num_filter=num_1x1, kernel=(1, 1), name=('%s_conv' % name))
+    tower_d3 = Conv(data=data, num_filter=num_d3_red, name=('%s_tower' % name), suffix='_conv')
+    tower_d3_a = Conv(data=tower_d3, num_filter=num_d3_1, kernel=(1, 3), pad=(0, 1), name=('%s_tower' % name), suffix='_mixed_conv')
+    tower_d3_b = Conv(data=tower_d3, num_filter=num_d3_2, kernel=(3, 1), pad=(1, 0), name=('%s_tower' % name), suffix='_mixed_conv_1')
+    tower_3x3_d3 = Conv(data=data, num_filter=num_3x3_d3_red, name=('%s_tower_1' % name), suffix='_conv')
+    tower_3x3_d3 = Conv(data=tower_3x3_d3, num_filter=num_3x3, kernel=(3, 3), pad=(1, 1), name=('%s_tower_1' % name), suffix='_conv_1')
+    tower_3x3_d3_a = Conv(data=tower_3x3_d3, num_filter=num_3x3_d3_1, kernel=(1, 3), pad=(0, 1), name=('%s_tower_1' % name), suffix='_mixed_conv')
+    tower_3x3_d3_b = Conv(data=tower_3x3_d3, num_filter=num_3x3_d3_2, kernel=(3, 1), pad=(1, 0), name=('%s_tower_1' % name), suffix='_mixed_conv_1')
+    pooling = mx.sym.Pooling(data=data, kernel=(3, 3), stride=(1, 1), pad=(1, 1), pool_type=pool, name=('%s_pool_%s_pool' % (pool, name)))
+    cproj = Conv(data=pooling, num_filter=proj, kernel=(1, 1), name=('%s_tower_2' %  name), suffix='_conv')
+    # concat
+    concat = mx.sym.Concat(*[tower_1x1, tower_d3_a, tower_d3_b, tower_3x3_d3_a, tower_3x3_d3_b, cproj], name='ch_concat_%s_chconcat' % name)
+    return concat
+
+def get_symbol(num_classes=1000, **kwargs):
+    data = mx.sym.Variable(name="data")
+    # stage 1
+    conv = Conv(data, 32, kernel=(3, 3), stride=(2, 2), name="conv")
+    conv_1 = Conv(conv, 32, kernel=(3, 3), name="conv_1")
+    conv_2 = Conv(conv_1, 64, kernel=(3, 3), pad=(1, 1), name="conv_2")
+    pool = mx.sym.Pooling(data=conv_2, kernel=(3, 3), stride=(2, 2), pool_type="max", name="pool")
+    # stage 2
+    conv_3 = Conv(pool, 80, kernel=(1, 1), name="conv_3")
+    conv_4 = Conv(conv_3, 192, kernel=(3, 3), name="conv_4")
+    pool1 = mx.sym.Pooling(data=conv_4, kernel=(3, 3), stride=(2, 2), pool_type="max", name="pool1")
+
+    # # stage 3
+    in3a = Inception7A(pool1, 64,
+                       64, 96, 96,
+                       48, 64,
+                       "avg", 32, "mixed")
+    in3b = Inception7A(in3a, 64,
+                       64, 96, 96,
+                       48, 64,
+                       "avg", 64, "mixed_1")
+    in3c = Inception7A(in3b, 64,
+                       64, 96, 96,
+                       48, 64,
+                       "avg", 64, "mixed_2")
+    in3d = Inception7B(in3c, 384,
+                       64, 96, 96,
+                       "max", "mixed_3")
+    # stage 4
+    in4a = Inception7C(in3d, 192,
+                       128, 128, 192,
+                       128, 128, 128, 128, 192,
+                       "avg", 192, "mixed_4")
+    in4b = Inception7C(in4a, 192,
+                       160, 160, 192,
+                       160, 160, 160, 160, 192,
+                       "avg", 192, "mixed_5")
+    in4c = Inception7C(in4b, 192,
+                       160, 160, 192,
+                       160, 160, 160, 160, 192,
+                       "avg", 192, "mixed_6")
+    in4d = Inception7C(in4c, 192,
+                       192, 192, 192,
+                       192, 192, 192, 192, 192,
+                       "avg", 192, "mixed_7")
+    in4e = Inception7D(in4d, 192, 320,
+                       192, 192, 192, 192,
+                       "max", "mixed_8")
+    # stage 5
+    in5a = Inception7E(in4e, 320,
+                       384, 384, 384,
+                       448, 384, 384, 384,
+                       "avg", 192, "mixed_9")
+    in5b = Inception7E(in5a, 320,
+                       384, 384, 384,
+                       448, 384, 384, 384,
+                       "max", 192, "mixed_10")
+    # pool
+    pool = mx.sym.Pooling(data=in5b, kernel=(8, 8), stride=(1, 1), pool_type="avg", name="global_pool")
+    flatten = mx.sym.Flatten(data=pool, name="flatten")
+    fc1 = mx.sym.FullyConnected(data=flatten, num_hidden=num_classes, name='fc1', flatten=False)
+    softmax = mx.sym.SoftmaxOutput(data=fc1, name='softmax')
+    return softmax

tests/python/relay/frontend/mxnet/model_zoo/mlp.py

@@ -0,0 +1,40 @@
+# 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.
+
+"""
+a simple multilayer perceptron
+"""
+import mxnet as mx
+
+def get_symbol(num_classes=10, **kwargs):
+    data = mx.symbol.Variable('data')
+    data = mx.sym.Flatten(data=data)
+    try:
+        fc1  = mx.symbol.FullyConnected(data = data, name='fc1', num_hidden=128, flatten=False)
+        act1 = mx.symbol.Activation(data = fc1, name='relu1', act_type="relu")
+        fc2  = mx.symbol.FullyConnected(data = act1, name = 'fc2', num_hidden = 64, flatten=False)
+        act2 = mx.symbol.Activation(data = fc2, name='relu2', act_type="relu")
+        fc3  = mx.symbol.FullyConnected(data = act2, name='fc3', num_hidden=num_classes, flatten=False)
+        mlp  = mx.symbol.softmax(data = fc3, name = 'softmax')
+    except:
+        fc1  = mx.symbol.FullyConnected(data = data, name='fc1', num_hidden=128)
+        act1 = mx.symbol.Activation(data = fc1, name='relu1', act_type="relu")
+        fc2  = mx.symbol.FullyConnected(data = act1, name = 'fc2', num_hidden = 64)
+        act2 = mx.symbol.Activation(data = fc2, name='relu2', act_type="relu")
+        fc3  = mx.symbol.FullyConnected(data = act2, name='fc3', num_hidden=num_classes)
+        mlp  = mx.symbol.softmax(data = fc3, name = 'softmax')
+    return mlp