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
A question regarding usage or RNNSharp for a particular problem #39
I'm wandering whether RNNSharp can be used for the following problem:
I have a sequence of 3-number vectors {[a1, a2, a3], [b1, b2, b3]... [n1, n2, n3]}. and a matching output, which is a vector of 5 numbers [A1, A2, A3, A4, A5]. And my training input/output would be a list of such pairs of sequences, where output always has the same size (5-number vector) while input size varies (number of 3-number vectors is different).
and I would use the samples above to train the network so that it can output a 5-number vector for a random-size input (random number of 3-number vectors)
Can RNNSharp deal with such a problem? I'm not a novice with NN, but I've never used RNN before, so I'm not sure whether it can be used with a (relative) success here. Also, if RNNSharp is capable of dealing with such a problem, any help of how can I start / setup the network would be a greatly appreciated:) (as I would be using the API in my own project)
Hello,
I'm wandering whether RNNSharp can be used for the following problem:
I have a sequence of 3-number vectors {[a1, a2, a3], [b1, b2, b3]... [n1, n2, n3]}. and a matching output, which is a vector of 5 numbers [A1, A2, A3, A4, A5]. And my training input/output would be a list of such pairs of sequences, where output always has the same size (5-number vector) while input size varies (number of 3-number vectors is different).
So, for example:
[1, 2, 3], [4, 5, 6] -> [1, 2, 3, 4, 5] [1, 2, 3], [4, 5, 6], [7, 8, 9,] -> [5, 6, 7, 8, 9] [1, 2, 3], [4, 5, 6], [7, 8, 9,], [10, 11, 12], [13, 14, 15] -> [10, 11, 12, 13, 14] ....
and I would use the samples above to train the network so that it can output a 5-number vector for a random-size input (random number of 3-number vectors)
e.g.
[1, 2, 3], [4, 5, 6], [7, 8, 9,], [10, 11, 12] -> desired 5-number output
Can RNNSharp deal with such a problem? I'm not a novice with NN, but I've never used RNN before, so I'm not sure whether it can be used with a (relative) success here. Also, if RNNSharp is capable of dealing with such a problem, any help of how can I start / setup the network would be a greatly appreciated:) (as I would be using the API in my own project)
Thanks in advance