All Questions
224 questions
0
votes
0
answers
114
views
Haversine Loss function: No gradients provided for any variable?
I'm trying to implement a custom loss function, the haversine loss that computes the distance between two points on a sphere. The coordinates of the classes are saved in a csv, hence I have to convert ...
0
votes
0
answers
72
views
Image Classification Model halting at epoch 1/25, tensorflow keras model
I have been following a tutorial of a deep learning neural network model project and for some reason my model halts at the first epoch when trying to train on a dataset of images. I am not sure if ...
- 1
0
votes
0
answers
81
views
Audio Classification using CNN predicting wrong all the time
I am trying to do an audio classification model that can take an audio as input and predict the genre of the music. The issue is that the accuracy is high, it is performing well on the test set but ...
0
votes
0
answers
175
views
Keras Fit Method 'Unrecognized Data Type' for NLP Classification Problem with TextVectorization
I have an LSTM based model. I am using a text vectorization layer and an embedding layer prior to the LSTM layers. I am also using two Dense layers with a softmax output. I am trying to go for ...
0
votes
1
answer
1k
views
can't load trained keras model with custom regularization class
I'm training the PointNet3D object classification model with my own dataset following the Tutorial here in Keras: https://keras.io/examples/vision/pointnet/#point-cloud-classification-with-pointnet
...
- 1,895
0
votes
1
answer
122
views
Multitask learning to classify on dog images
I'm trying to train a multitask classification model (mobilenet). Basically, a single model that given an image of a dog, it classifies both the color and the breed. Each dataset simply has ...
- 5,617
-2
votes
1
answer
58
views
Should my classification object be black or white inline [closed]
I have a simple question. I am doing 2D shape recognition using Keras. I have these sample images I have generated using cv2:
Quarter Circle
Cross
Rectangle.
I was wondering if I should do a black ...
- 128
-1
votes
1
answer
131
views
Improving Multi-Class classification model? [closed]
I'm creating a CNN for skin lesion classification.
I recently added a weighter loss function to my model to try and improve its accuracy, but even with the new weighted losses, my model still only ...
- 3
0
votes
1
answer
155
views
Training Loss for Classification Model Isn't Decreasing [closed]
I'm currently building a video classification model for engagement detection but I'm having some trouble training it. The model takes in two tensors as inputs: a 10x48x48x1 tensor which holds a stack ...
- 21
0
votes
1
answer
247
views
Keras Grad-CAM layer definition
So I am training a Keras model that has a base model of VGG16. As seen below.
Model summary:
I am also generating a Grad-CAM heatmap using the visualized layer as an input. because i cannot choose ...
- 13
0
votes
1
answer
100
views
Classification of multi target and multi label using tensorflow and keras
I would like to classify objects in an image such as cars. There could be multiple cars in the image and I would like to get brand, color, type of each car in the image.
I created a multibranch CNN to ...
- 317
0
votes
1
answer
205
views
Train a classifier on specific labels of MNIST dataset with TensorFlow
I would like to train a classifier on MNIST dataset but with limited labels. For eg. I would like to train a classifier only on labels [1, 4, 5, 6, 8, 9] from all the labels [0-9]. I am getting the ...
- 528
0
votes
0
answers
78
views
TensorFlow output in the wrong range
I am training a simple single-layer NN for binary classification with TensorFlow. My training data has the labels 0 and 1, the training of the NN seems ok for now, I get accuracy values of ~80% and a ...
0
votes
1
answer
84
views
Using an LSTM to predict a category
My dataset contains a curve of data points with a column labeled "BUY". "BUY" indicates that the curve reached a local maxima or minima at this point (not fully on accurate due to ...
- 13
0
votes
1
answer
185
views
Why does Keras Precision metric work on One-Hot-Vectors but their Accuracy metric does not?
This doesn't seem to make sense to me. When I use the keras.metrics.Precision metric it works fine on my multi-class model ( i.e. on one-hot-vector outputs ), but when I use keras.metrics.Accuracy it ...
- 879