mirror of
https://github.com/k2-fsa/icefall.git
synced 2025-08-26 18:24:18 +00:00
216 lines
6.5 KiB
Python
216 lines
6.5 KiB
Python
#!/usr/bin/env python3
|
|
# Copyright 2021 Xiaomi Corp. (authors: Fangjun Kuang
|
|
# Mingshuang Luo)
|
|
#
|
|
# See ../../../../LICENSE for clarification regarding multiple authors
|
|
#
|
|
# Licensed 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.
|
|
|
|
|
|
import math
|
|
import torch.nn as nn
|
|
|
|
|
|
def conv3x3(in_planes, out_planes, stride=1):
|
|
return nn.Conv2d(
|
|
in_planes,
|
|
out_planes,
|
|
kernel_size=3,
|
|
stride=stride,
|
|
padding=1,
|
|
bias=False,
|
|
)
|
|
|
|
|
|
class BasicBlock(nn.Module):
|
|
expansion = 1
|
|
|
|
def __init__(self, inplanes, planes, stride=1, downsample=None):
|
|
super(BasicBlock, self).__init__()
|
|
self.conv1 = conv3x3(inplanes, planes, stride)
|
|
self.bn1 = nn.BatchNorm2d(planes)
|
|
self.relu = nn.ReLU(inplace=True)
|
|
self.conv2 = conv3x3(planes, planes)
|
|
self.bn2 = nn.BatchNorm2d(planes)
|
|
self.downsample = downsample
|
|
self.stride = stride
|
|
|
|
def forward(self, x):
|
|
residual = x
|
|
out = self.conv1(x)
|
|
out = self.bn1(out)
|
|
out = self.relu(out)
|
|
|
|
out = self.conv2(out)
|
|
out = self.bn2(out)
|
|
if self.downsample is not None:
|
|
residual = self.downsample(x)
|
|
|
|
out += residual
|
|
out = self.relu(out)
|
|
|
|
return out
|
|
|
|
|
|
class ResNet(nn.Module):
|
|
def __init__(self, block, layers):
|
|
self.inplanes = 64
|
|
super(ResNet, self).__init__()
|
|
self.layer1 = self._make_layer(block, 64, layers[0])
|
|
self.layer2 = self._make_layer(block, 128, layers[1], stride=2)
|
|
self.layer3 = self._make_layer(block, 256, layers[2], stride=2)
|
|
self.layer4 = self._make_layer(block, 512, layers[3], stride=2)
|
|
self.avgpool = nn.AdaptiveAvgPool2d(1)
|
|
for m in self.modules():
|
|
if isinstance(m, nn.Conv2d):
|
|
n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
|
|
m.weight.data.normal_(0, math.sqrt(2.0 / n))
|
|
elif isinstance(m, nn.BatchNorm2d):
|
|
m.weight.data.fill_(1)
|
|
m.bias.data.zero_()
|
|
elif isinstance(m, nn.BatchNorm1d):
|
|
m.weight.data.fill_(1)
|
|
m.bias.data.zero_()
|
|
|
|
def _make_layer(self, block, planes, blocks, stride=1):
|
|
downsample = None
|
|
if stride != 1 or self.inplanes != planes * block.expansion:
|
|
downsample = nn.Sequential(
|
|
nn.Conv2d(
|
|
self.inplanes,
|
|
planes * block.expansion,
|
|
kernel_size=1,
|
|
stride=stride,
|
|
bias=False,
|
|
),
|
|
nn.BatchNorm2d(planes * block.expansion),
|
|
)
|
|
|
|
layers = []
|
|
layers.append(block(self.inplanes, planes, stride, downsample))
|
|
self.inplanes = planes * block.expansion
|
|
for i in range(1, blocks):
|
|
layers.append(block(self.inplanes, planes))
|
|
|
|
return nn.Sequential(*layers)
|
|
|
|
def forward(self, x):
|
|
x = self.layer1(x)
|
|
x = self.layer2(x)
|
|
x = self.layer3(x)
|
|
x = self.layer4(x)
|
|
x = self.avgpool(x)
|
|
x = x.view(x.size(0), -1)
|
|
return x
|
|
|
|
|
|
class VisualNet2(nn.Module):
|
|
def __init__(self, num_classes):
|
|
"""
|
|
Args:
|
|
num_classes:
|
|
The output dimension of the visualnet2 model.
|
|
"""
|
|
super().__init__()
|
|
self.num_classes = num_classes
|
|
self.inputDim = 512
|
|
self.conv3d = nn.Conv3d(
|
|
3,
|
|
64,
|
|
kernel_size=(5, 7, 7),
|
|
stride=(1, 2, 2),
|
|
padding=(2, 3, 3),
|
|
bias=False,
|
|
)
|
|
self.bn = nn.BatchNorm3d(64, track_running_stats=True)
|
|
self.relu = nn.ReLU(True)
|
|
self.maxpool = nn.MaxPool3d(
|
|
kernel_size=(1, 3, 3), stride=(1, 2, 2), padding=(0, 1, 1)
|
|
)
|
|
|
|
# resnet
|
|
self.resnet18 = ResNet(BasicBlock, [2, 2, 2, 2])
|
|
|
|
# grus
|
|
self.gru1 = nn.GRU(512, 512, 1, bidirectional=True)
|
|
self.gru2 = nn.GRU(1024, 512, 1, bidirectional=True)
|
|
|
|
# dropout
|
|
self.dropout = nn.Dropout(p=0.5)
|
|
|
|
# fc
|
|
self.linear = nn.Linear(1024, self.num_classes)
|
|
|
|
# initialize
|
|
self._initialize_weights()
|
|
|
|
def forward(self, x):
|
|
frameLen = x.size(2)
|
|
x = self.conv3d(x)
|
|
x = self.bn(x)
|
|
x = self.relu(x)
|
|
x = self.maxpool(x)
|
|
|
|
x = x.transpose(1, 2)
|
|
x = x.contiguous()
|
|
x = x.view(-1, 64, x.size(3), x.size(4))
|
|
x = self.resnet18(x)
|
|
|
|
x = self.dropout(x)
|
|
x = x.view(-1, frameLen, self.inputDim)
|
|
|
|
x = x.permute(1, 0, 2)
|
|
x, h = self.gru1(x)
|
|
x, h = self.gru2(self.dropout(x))
|
|
x = self.linear(x)
|
|
x = x.permute(1, 0, 2)
|
|
x = nn.functional.log_softmax(x, dim=-1)
|
|
return x
|
|
|
|
def _initialize_weights(self):
|
|
for m in self.modules():
|
|
if isinstance(m, nn.Conv3d):
|
|
n = (
|
|
m.kernel_size[0]
|
|
* m.kernel_size[1]
|
|
* m.kernel_size[2]
|
|
* m.out_channels
|
|
)
|
|
m.weight.data.normal_(0, math.sqrt(2.0 / n))
|
|
if m.bias is not None:
|
|
m.bias.data.zero_()
|
|
|
|
elif isinstance(m, nn.Conv2d):
|
|
n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
|
|
m.weight.data.normal_(0, math.sqrt(2.0 / n))
|
|
if m.bias is not None:
|
|
m.bias.data.zero_()
|
|
|
|
elif isinstance(m, nn.Conv1d):
|
|
n = m.kernel_size[0] * m.out_channels
|
|
m.weight.data.normal_(0, math.sqrt(2.0 / n))
|
|
if m.bias is not None:
|
|
m.bias.data.zero_()
|
|
|
|
elif isinstance(m, nn.BatchNorm3d):
|
|
m.weight.data.fill_(1)
|
|
m.bias.data.zero_()
|
|
|
|
elif isinstance(m, nn.BatchNorm2d):
|
|
m.weight.data.fill_(1)
|
|
m.bias.data.zero_()
|
|
|
|
elif isinstance(m, nn.BatchNorm1d):
|
|
m.weight.data.fill_(1)
|
|
m.bias.data.zero_()
|