Questions tagged [deep-learning]
Deep Learning is an area of machine learning whose goal is to learn complex functions using special neural network architectures that are "deep" (consist of many layers). This tag should be used for questions about implementation of deep learning architectures. General machine learning questions should be tagged "machine learning". Including a tag for the relevant software library (e.g., "keras", "tensorflow","pytorch","fast.ai" etc) is helpful.
deep-learning
27,912
questions
0
votes
0
answers
2
views
Why is the gradient wrt input zero?
This snippet tries to take gradient wrt the input to a trained model
import torch
import torch.nn as nn
DATASET_SIZE = 10
TUNING_EPOCH = 10
RANDOMIZATION = "rand"
shape = (10000,)
bsize = 1
...
-2
votes
0
answers
6
views
Two variants of Nesterov Accelerated Gradient: are they equivalent?
I was puzzled to find that the description of the Nesterov Accelerated Gradient on Paperswithcode, namely:
v_t = beta * v_t-1 + eta * ∇ J(theta - beta * v_t-1)
theta_t = theta_t-1 + v_t
is slighlty ...
-1
votes
0
answers
12
views
What Is the best ML API for a Deep Audio Classification Model on the Raspberry Pi 0 2W?
I'd like to get your opinions on which ML API would be best for building an audio classification model. This model will be deployed on a small Raspberry Pi (Raspberry Pi 0 2W; could be something ...
0
votes
0
answers
13
views
Unet from segmentation_models_pytorch stalling in training
I have been following a tutorial on training a segmentation model on a custom dataset, but it refuses to make any progrees in training the model.
This is my model setup
import ...
0
votes
0
answers
27
views
data augmentation in semantic segmentation
I am tryin to perform data augmentation on my dataset but I am getting blank white images instead, works fine on masks but there is a problem with images.
How to resolve this issue?
this is original ...
-1
votes
0
answers
13
views
ResourceExhaustedError: OOM when allocating tensor with shape[17656320,1024] and type float
This question has been asked a couple of times and I tried this, this, this etc, but couldn't solve my problem. I have just 4 images of approximately 1MB which I'm trying to fit into a deep learning ...
0
votes
1
answer
19
views
Inconsistent Model Predictions When Using Entire Validation Dataset vs. Batch Sampling in TensorFlow
I am training a deep learning model using TensorFlow and Keras on an image classification task. My model achieves high validation accuracy when evaluated using the validation_ds dataset. However, when ...
-2
votes
0
answers
9
views
It is difficult to implement crack segmentation using the SAM large model. How to use Lora to fine tune SAM to achieve crack segmentation?
I don't know how to write the code during training. I want to get relevant code to study
It is difficult to implement crack segmentation using the SAM large model. How to use Lora to fine tune SAM to ...
0
votes
0
answers
22
views
Why are gradients None for all variables in my TensorFlow Siamese Network during training?
I'm training a Siamese Network in TensorFlow for image registration. The network takes a pair of images (fixed and moving) and outputs parameters for an affine transformation that aligns the moving ...
-1
votes
0
answers
13
views
Deep NN converging at same cost
I have a deep neural network, and I am trying to use it to classify signed numbers in 64x64 pcitures of numbers from 0-9. The cost seems to always converge around 2.3, which is 10% accuracy or just ...
-2
votes
0
answers
18
views
single class object detection using cnn getting false positive
Here, i trying to build a Manhole object detection using cnn , In this model i got 95% accuracy and after training. im getting false positive like , if i test image with manhole(trained object) for ...
0
votes
0
answers
33
views
Processing Frames instead of jpg images for Deep Learning Model
I've been trying to implement live translation with a sign language cnn model i made with Keras and OpenCV and im running into an issue with converting the frames to a format the model is able to ...
-1
votes
0
answers
26
views
Practical difference between EMA decay and LR decay
I cannot wrap my head around the differences - in practice - between EMA decay and LR decay.
It feels to me like they both accomplish the same thing, just in different ways (the following is likely ...
0
votes
0
answers
5
views
How to interpret the following pix2pix training losses?
The plots of loss functions obtained are as follows:
enter image description here
How to interpret this graph? I don't know this results is great or bad..
How to improve this pix2pix model?
Help me ...
0
votes
0
answers
14
views
How to perform Faster Object detection on webcam using onnx runtime web?
This is the javascript script which is performing object detection accurately but processing is taking time which creates lags in frames.There is no smoothness in it.
This is the script from some code ...