37f56db5af
* [example] added code for training speech command dataset * update kes_model.py * update kes_model.py * format * format * add more comments to explain the new classifier designed for speech command classification task * add copyrigh info * update copyrigh info of classifier.py
48 lines
1.5 KiB
Python
48 lines
1.5 KiB
Python
# Copyright (c) 2021 Jingyong Hou
|
|
#
|
|
# 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 torch
|
|
import torch.nn as nn
|
|
|
|
|
|
class GlobalClassifier(nn.Module):
|
|
"""Add a global average pooling before the classifier"""
|
|
def __init__(self, classifier: nn.Module):
|
|
super(GlobalClassifier, self).__init__()
|
|
self.classifier = classifier
|
|
|
|
def forward(self, x: torch.Tensor):
|
|
x = torch.mean(x, dim=1)
|
|
return self.classifier(x)
|
|
|
|
|
|
class LastClassifier(nn.Module):
|
|
"""Select last frame to do the classification"""
|
|
def __init__(self, classifier: nn.Module):
|
|
super(LastClassifier, self).__init__()
|
|
self.classifier = classifier
|
|
|
|
def forward(self, x: torch.Tensor):
|
|
x = x[:, -1, :]
|
|
return self.classifier(x)
|
|
|
|
class ElementClassifier(nn.Module):
|
|
"""Classify all the frames in an utterance"""
|
|
def __init__(self, classifier: nn.Module):
|
|
super(ElementClassifier, self).__init__()
|
|
self.classifier = classifier
|
|
|
|
def forward(self, x: torch.Tensor):
|
|
return self.classifier(x)
|