
I'm currently running an MSc and researching on electricity demand projection using recurrent neural network. The issue is that I don't know anything on neural network but I'm always ready and willing ...

This is a general question for any of the frameworks for both RNN and LSTM.
When we use a Vanilla or plain networks, for a single layer, such as
current_layer = torch.nn.Linear(100,125) means that ...

So, i have the task to classify different job titles into categories. The data is really noisy and consists of around 200 categories containing around 20 job titles. So my thought has been to create a ...

I was trying to figure out how to estimate the number of parameters in an LSTM layer. What is the relationship of number of parameters with the num lstmcells, inputdimension, and hidden outputstate ...

I am going through a Binary Classification tutorial using PyTorch and here, the last layer of the network is torch.Linear() with just one neuron. (Makes Sense) which will give us a single neuron. as ...

I have started Deep Learning few months ago using Tensorflow and tf.keras.
I fully get the concept behind classic Dense Layers or Convolutional/pooling layers where the unit parameter is the number ...

I would like to solve an "optimization problem" by using new and fancy machine learning methods (also to learn more about them) and I haven't found yet which methods / models will help me to solve my ...

I try to adapt "Text generation with an RNN" tutorial to generate new dinosaur names from a list of the existing ones. For training RNN tutorial text is divided into example character sequences of ...

I'm a beginner to Recurrent Neural Network, while learning it I find that the RNN only takes 1 previous value into consideration or 'n' previous value into consideration to predict the next value. But ...

I'm currently playing with this GitHub repo.
I'm having an issue with the file opening/reading part of it. In the prepare.py, there is the line fo = open(sys.argv[1]). Now when I create a file and ...

In Keras we have keras.preprocessing.text to tokenize the text on our requirement and generate a voabulary.
tokenizer = tf.keras.preprocessing.text.Tokenizer(split=' ', oov_token=1)
tokenizer....

To summarize, I write my own RNN from scratch and it seems that it has no problem working. However, it takes a lot of time to train the data, so I want to determine the batch size. The system is ...

def create_example_model():
tf.keras.backend.set_floatx('float64')
model = Sequential()
model.add(LSTM(128, input_shape=((60, len(df_train.columns)))))
model.add(Dense(64, activation='...

im using keras for a multiclass clasffication of textcomments problem, this one, to be precise:
https://www.kaggle.com/c/jigsawtoxiccommentclassificationchallenge
There is six classes, and the ...

I have a question about shifting the input sequence to predict the "next word" in a sequence. I'm providing a toy example below but note my samples are IID and not from a giant corpus of text but ...

Im trying to implement a textclassifier using triplet loss to classify different job descriptions into categories based on this paper. But whatever i do, the classifier yields very bad results.
For ...

I am practicing with RNN. I randomly create 5 integers. If the first integer is an odd number, the y value is 1, otherwise y is 0 (So, only the first x counts). Problem is, when I run this model, it ...

I am training an LSTM NN to forecast the time series of more than 3000 features. The things is each features has a particular time series, but the NN predict a constant value which is different for ...

I am using binary cross entropy, and I have 2 epochs:
batch_size = 32
epochs = 2
History = model.fit(padded_train, y_train, batch_size = batch_size, epochs = epochs, validation_split = 0.1)
Now i ...

Currently, I'm playing with Stocks Predictions task which I try to solve using LSTM/GRU.
Problem: After training LSTM/GRU I get huge drop predicted values
Model training process
Train, test data is ...

I want to build a LSTM model where the input to the (n+1)th timestep is a function of the output at the (n)th timestep. I don't see a way this can be done in the current framework. People have been ...

the GridSearch parameter tuning for batch_size, epochs, optimizer, and units work perfectly for my RNN:
X_train = []
y_train = []
for i in range(60, np.ma.size(training_set_scaled)): #have ...

For a specific problem in reinforcement learning (inspired in this paper), I'm using a RNN which is fed with data of shape (batch_size, time_steps, features) = (1,1,1), for L datapoints, and then a "...

I have implemented a simple LSTM as well as GRU network for timeseries forecast:
def LSTM1(T0, tau0, tau1, optimizer, y0):
model = Sequential()
model.add(Input(shape=(T0,tau0), ...

%matplotlib inline
import tensoflow as tf
import matplotlib.pyplot as plt
from rnn.lstm_recurrent_model import LSTMRecurrentModel
from rnn.lstm_solver import LSTMSolver
from rnn.data_util import ...

model=tf.keras.Sequential([tf.keras.layers.Embedding(encoder.vocab_size,64),
tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(64, return_sequences=True)),
...

I have training data as two columns
1.'Sentences'
2.'Relevant_text' (text in this column is a subset of text in the column 'Sentences')
I tried training a RNN with LSTM directly treating 'Sentences' ...

im currently looking around at some lstm models using keras, and trying to learn about how they function.
Im currently looking at this guide (not necessary to check out): https://www.kaggle.com/...

Given a task of sentence corruption detection(binary classification), I wonder is it possible to use both charcterlevel and wordlevel RNN as the corruption happens both at character level(...

I would like to train RNN to detecting corrupted word. (binary classification)
Currently, I have a approximate ratio of # correct words : # corrupted words = 20:1.
Also, most corruption is like "...

I have a theoretical question based on RNN. As we all know, RNN have hidden states and it holds its previous time step information.
While working with RNN we also batch our input data for better ...

If I am using an LSTM to predict future values of a time series chart which is more or less monotonically increasing. Does tanh work as an activation function for all the LSTM units since it is a ...

I am having a time series prediction problem and building an LSTM like below :
def create_model():
model = Sequential()
model.add(LSTM(50,kernel_regularizer=l2(0.01), recurrent_regularizer=l2(...

I am relatively new in Neural Network and Keras. I am trying different Neural Network architectures and my goal is not to get the best accuracy but more to understand and be able to diagnostic the ...

I'm interested in deriving backpropagation through time by hand for Keras's SimpleRNN.
Some implementations of RNNs use sigmoid functions, and some use hyperbolic tangents.
I'm looking for the set ...

How the model learn without changing its parameters/ weights?
If we train the RNN on some data and then apply it to test data , what changes do we make ? Cause the weights/parameters don't change ...

I found that cudnn's RNN calculation has two sets of interfaces, one is to separate the input data into multiple tensor representations, and the other is to encapsulate all the data into a ...

Actually, i'm still confusing to determine the right size of WHH matrix in RNN, are the dimensions of the WHH matrix determined by the number of recurrent steps or other?
I have read the following ...

I have a dataset of timeseries. each time series has a value and more features.
sample example : 4 time stamps of history, each time stamp is a vector of 3 value [20,2,4].
sometime the first value ...

I am writing a custom LSTM cell and I use dynamic_rnn for the model. I have a custom variable in MyLSTMCell which I add one to it in each epoch.
class MyLSTMCell(tf.nn.rnn_cell.BasicLSTMCell):
...

I've loaded an "untrained" model from github at https://pytorch.org/hub/nvidia_deeplearningexamples_tacotron2/, but when I try to see summary of the model using torchsummary.summary I get the ...

I am using the tensorflow image captioning tutorial to train a model. They have used GRU in decoder but i want to use LSTM based decoder or infact bidirectional LSTM if possible.. GRU works fine but ...

I have to fine tune a model which is already trained with the dataset .I have to fine tune this model with new set of data which are representing the same classes as in my training but having some ...

I'm new to Tensorflow and trying to understand the Tensorflow model of Texttospeech synthesis https://github.com/Rayhanemamah/Tacotron2
Very basic questions. I cloned the model from the above ...

I have the following timeseries aggregated input for an LSTMbased model:
x(0): {y(0,0): {a(0,0), b(0,0)}, y(0,1): {a(0,1), b(0,1)}, ..., y(0,n): {a(0,n), b(0,n)}}
x(1): {y(1,0): {a(1,0), b(1,0)}, y(...

Using Pytorch LSTM architecture trying to build a text generation model. For every batch, I'm using pad_sequence to have min padding for every sequence, therefore I have a variable dims batch (...

In Speaker segmentation, how do we check for OVER SEGMENTATION, as over segmentation leads to high purity but low coverage.
What is the difference between purity and coverage?
How does the tradeoff ...

While working on making a simple RNN using Pytorch nn.linear function. So firstly I initialized my weights as
self.W_x = nn.Linear(self.input_dim, self.hidden_dim, bias=True)
self.W_h = nn.Linear(...

Now I have been looking RNN's for a while.
And the vanilla version seems pretty obvious to me, a model trains by running each word in a sequence through the network one at a time. Here in the rolled ...

im looking a bit at lstm models for a data science project, and ive noticed how 't' often denotes current state. Like in this example:
this is of course the formula for the current state. I am a ...