Back to search
Top Contributors
1
commits
Designed a neural network from scratch with Octave GNU, capable of detecting handwritten numerical digits from 0-9 with 99% accuracy. The network is trained with 5000 examples from the MNIST data set. Cost function and gradient are calculated to minimize error of the network’s output.
Stars
1
Forks
0
Watchers
1
Open Issues
0
Overall repository health assessment
No package.json found
This might not be a Node.js project
1
commits