exelban/tensorflow-cifar-10 — an AI tool — sits at 90 GitHub stars in the AI Tools space. Cifar-10 CNN implementation using TensorFlow library with 20% error.
Snapshot summary built from the project's own GitHub metadata — there's no written TopGit review yet. The page will update automatically when a full review is published.
WHY NO REVIEW YET
TopGit writes full reviews for the most-starred, most-requested repositories. This page is a snapshot until then — see the READ ME tab for the original README in full.
Cifar-10 convolutional network implementation example using TensorFlow library.
Requirement
Library
Version
Python
^3.6.5
Tensorflow
^1.6.0
Numpy
^1.14.2
Pickle
^4.0
Accuracy
Best accurancy what I receive was 79.12% on test data set. You must to understand that network cant always learn with the same accuracy. But almost always accuracy more than 78%.
This repository is just example of implemantation convolution neural network. Here I implement a simple neural network for image recognition with good accuracy.
If you want to get more that 80% accuracy, You need to implement more complicated nn models (such as ResNet, GoogleLeNet, mobileNetV2 ect).
Usage
Download code
git clone https://github.com/exelban/tensorflow-cifar-10
cd tensorflow-cifar-10
Check if you have nessessary packages
pip3 install numpy tensorflow pickle
Train network
By default network will be run 60 epoch (60 times on all training data set).
You can change that by editing _EPOCH in train.py file.
Also by default it process 128 files in each step.
If you training network on CPU or GPU (lowest that 1060 6GB) change _BATCH_SIZE in train.py to a smaller value.
Trying to restore last checkpoint ...
Restored checkpoint from: ./tensorboard/cifar-10-v1.0.0/-23460
Accuracy on Test-Set: 78.81% (7881 / 10000)
Training time
Here you can see how much time takes 60 epoch:
Device
Batch size
Time
Accuracy [%]
NVidia GTX 1070
128
8m 4s
79.12
Intel i7 7700HQ
128
3h 30m
78.91
Please send me (or open issue) your time and accuracy. I will add it to the list.
Model
What's new
v1.0.1
- Set random seed
- Added more information about elapsed time on epoch and full training
v1.0.0
- Removed all references to cifar 100
- Small fixes in data functions
- Almost fully rewrited train.py
- Simplyfy cnn model
- Changed optimizer to AdamOptimizer
- Changed Licence to MIT
- Removed confusion matrix (don't like to have unnecessary dependencies)
- Improved accuracy on testing data set (up to 79%)
- Small fixes in train.py
- Changed saver functions (now session will be saved only if accuracy in this session will be better than the last saved)
- Updated packages
v0.0.1
- Make tests on AWS instances
- Model fixes
- Remove cifar-100 dataset
Does exelban/tensorflow-cifar-10 have a project website?
No homepage URL was recorded for exelban/tensorflow-cifar-10 in TopGit's last sync. The README tab above frequently contains screenshots and demo links, or check the repository description on GitHub.
How active is development on exelban/tensorflow-cifar-10?
The most recent commit recorded on exelban/tensorflow-cifar-10 was 7.0 years ago, based on the GitHub push timestamp. The repository has 48 forks — one of the better signals of community interest.
Is exelban/tensorflow-cifar-10 open source?
Yes — exelban/tensorflow-cifar-10 ships under the MIT license, which makes its source code freely readable (and, depending on license terms, forkable and reusable). Source: github.com/exelban/tensorflow-cifar-10.
What license does exelban/tensorflow-cifar-10 use?
exelban/tensorflow-cifar-10 is released under the MIT license. Always verify the LICENSE file directly on GitHub for the authoritative terms — license strings can be edited out of sync with a project's actual stance.
What topics is exelban/tensorflow-cifar-10 associated with?
GitHub's repository topics for exelban/tensorflow-cifar-10: "cifar10", "cnn", "conv-2d", "convolutional-neural-networks", "example", "model", "network", "neural", "simple", "tensorflow", "tensorflow-library". TopGit's editorial category is AI Tools.
Where do I read more about exelban/tensorflow-cifar-10?
This TopGit page is a snapshot — the READ ME tab shows the project's own README content (links stripped, images preserved). The GitHub repository at github.com/exelban/tensorflow-cifar-10 is the definitive source.
Read full README in the tab above.
Curious whether tensorflow-cifar-10 is right for you?
Let ChatGPT, Claude, or Perplexity look into it — click below and see what AI actually says about tensorflow-cifar-10.