lstm-gru-pytorch-master
所属分类：数值算法/人工智能
开发工具：Python
文件大小：145KB
下载次数：15
上传日期：2020-03-07 21:47:57
上 传 者：
guojunj
说明： 使用python和pytorch实现LSTM和GRU网络，并用MNIST数据集进行训练学习
(Use Python and python to implement LSTM and Gru network, and use MNIST data set for training and learning)
文件列表:
loss_gru.txt (150000, 2019-01-21)
loss_lstm.txt (150000, 2019-01-21)
lstm_gru.ipynb (40699, 2019-01-21)
# Implementation of LSTM and GRU cells for PyTorch
This repository is an implementation of the LSTM and GRU cells without using the PyTorch LSTMCell and GRUCell.
It is tested on the MNIST dataset for classification.
The 28x28 MNIST images are treated as sequences of 28x1 vector.
The RNN consist of
- A linear layer that maps 28-dimensional input to and 128-dimensional hidden layer
- One intermediate recurrent neural network (LSTM or GRU)
- A fully connected layer which maps the 128 dimensional input to 10-dimensional vector of class labels.
## Requirements
Python>=3.5
PyTorch== 0.4.0
近期下载者：
相关文件：
收藏者：