14 questions
0
votes
0
answers
33
views
How to Train a Network to Distinguish True and False Outputs from a Pretrained ResNet on CIFAR-10
I'm trying to perform image classification on CIFAR-10 using ResNet. By pretraining ResNet, I was able to achieve a performance improvement of 92% on CIFAR-10.
Next, I want to add a fully connected ...
- 7
0
votes
0
answers
62
views
How to train a simple vision transformer model on a custom dataset similar to CIFAR10?
I am attempting to train the following vision transformer model: https://github.com/tintn/vision-transformer-from-scratch on a custom dataset formatted similarly to CIFAR10.
This code is configured to ...
0
votes
0
answers
23
views
cunn_ClassNLLCriterion_updateOutput_kernel: block: [0,0,0], thread: [0,0,0] Assertion `t >= 0 && t < n_classes` failed
I work with a portion of Cifar 100 database, so, I extract only data with classes : [10, 11, 12, 13, 14, 15, 16, 17, 18, 19] to make a multiclass prediction of 10 classes.
class ResNet(nn.Module):
......
- 701
0
votes
0
answers
35
views
load cifar-100 get low test accuracy
when I load cifar-10 with this manner :
from torchvision.datasets import CIFAR10
DATA_ROOT = "~/data/cifar-10"
def load_data() -> Tuple[torch.utils.data.DataLoader, torch.utils.data....
- 701
0
votes
0
answers
46
views
Trained model on cifar10 performs poorly on real images
So I'm trying to train a model using the CIFAR10 dataset.
The problem is that while the performance of the model on validation and test sets are good (about 95-96%), the model fails to predict images ...
- 39
0
votes
0
answers
32
views
Very low test accuracy with Cifar100
def load_cifar100_data_test(datadir,test_bs):
dl_obj = CIFAR100_truncated
transform_test = _data_transforms_cifar100_test()
test_ds = dl_obj(datadir, classes=None, dataidxs=None, train=...
- 701
0
votes
0
answers
30
views
when I change data set I receive: Layer sequential weight shape (3, 3, 1, 64) is not compatible with provided weight shape (4, 4, 3, 64)
I'm new to py and machine-learning.
for university project I'm implementing a simple and basic federated cnn that should work on mnist, cifar10 and cifar100. Aside from the fact that for the moment I'...
- 1
0
votes
0
answers
152
views
How to use resnet instead of VGG to improve Cifar 100 performance
In order to improve the performance of cifar 100, we wanted to improve the code by using resnet. However, the accuracy continues to hover around 50%. There seems to be a problem with my code, but ...
- 1
0
votes
1
answer
22
views
Why keras accuracy and loss are not changing between epochs and how to fix
This is my code:
from tensorflow import keras
import numpy as np
import matplotlib.pyplot as plt
x_train, y_train), (x_test, y_test) = keras.datasets.cifar100.load_data()
num_of_class = 100
y_train =...
0
votes
1
answer
385
views
using mobilnetv2 on cifar10, cifar100, and imagenet accuracy is not enough
I have been making some experiments with mobilenetv2 and used dataset with cifar10, cifar100.
when I used the code, it does not give me accuracy above 80%(accuracy using validation dataset)
when I ...
- 61
0
votes
1
answer
577
views
keras.backend.function() won't accept model.layers[0].input as input tensor
I'm trying to use this tensorflow implementation of openmax and adapt it to CIFAR-100 to use it in my project. As part of openmax instead of softmax you have to get the activation vector of the ...
- 11
1
vote
0
answers
237
views
Unable to bypass SSL certificate verification failure
Attempting to run the following line of code:
data <- dataset_cifar100("fine")
results in:
Error: Exception: URL fetch failure on https://www.cs.toronto.edu/~kriz/cifar-100-python.tar....
2
votes
0
answers
185
views
Dataloader with a different batch size for each iteration for a deep learning project
For my deep learning project I need a cifar10 data loader (python) for my model which uses a varied batch size each iteration, unfortunately torch.utils.data.DataLoader uses a constant batch size ...
- 39
0
votes
1
answer
68
views
Run Condition GAN model on the CIFAR-10 dataset, unable to clip (np.clip(X_test,-1,1))
everyone,
I want to load the dataset CIFAR-10, and do some pre-processing on those loaded images, when I run the following codes, my kernel crashes for no reason:
import numpy as np
%matplotlib inline
...
- 21