search

zhongkaifu / RNNSharp

RNNSharp is a toolkit of deep recurrent neural network which is widely used for many different kinds of tasks, such as sequence labeling, sequence-to-sequence and so on. It's written by C# language and based on .NET framework 4.6 or above versions. RNNSharp supports many different types of networks, such as forward and bi-directional network, sequence-to-sequence network, and different types of layers, such as LSTM, Softmax, sampled Softmax and others.
BSD 3-Clause "New" or "Revised" License
285 stars 92 forks source link

Impact of run-time feature #26

Closed My-Khan closed 7 years ago

My-Khan commented 8 years ago

Hello Sir, hope you will be fine Would you please like to let me know the impact of run-time features on results, simply why we use run-time features?

zhongkaifu commented 8 years ago

Run-time features means RNNSharp uses the previous token's output (T-1) as feature for current token (T). Currently, this feature is only enabled for forward RNN, but not b-directional RNN.

My-Khan commented 8 years ago

Many thanks sir, but sir still confused, Actually whats the logic behind its usage? SIMPLY why run-time features are being considered?

zhongkaifu commented 8 years ago

The idea of run-time feature is to provide previous output token as feature to current token. It extends feature we have.

My-Khan commented 7 years ago

ok sir got it