All Projects → capnramses → neural_net_handwriting

capnramses / neural_net_handwriting

Licence: MIT License
neural network for handwriting recognition from scratch in C

Programming Languages

c
50402 projects - #5 most used programming language
shell
77523 projects

Projects that are alternatives of or similar to neural net handwriting

Mathematical-Handwriting-recognition
Convert hand written mathematical expressions and formula to Latext using Machine Learning
Stars: ✭ 50 (+194.12%)
Mutual labels:  handwriting-recognition
PyCasia
A python library to work with the CASIA Chinese handwriting database.
Stars: ✭ 38 (+123.53%)
Mutual labels:  handwriting-recognition
nhcr
Offline Nepali Handwritten Character Recognition Using Artificial Neural Networks
Stars: ✭ 19 (+11.76%)
Mutual labels:  mlp
iinkJS
✏️ ☁️ iinkJS is the fastest way to integrate rich handwriting recognition features in your webapp.
Stars: ✭ 65 (+282.35%)
Mutual labels:  handwriting-recognition
air writing
Online Hand Writing Recognition using BLSTM
Stars: ✭ 26 (+52.94%)
Mutual labels:  handwriting-recognition
Deep-Learning-Models
Deep Learning Models implemented in python.
Stars: ✭ 17 (+0%)
Mutual labels:  mlp
tensorflow-example
Tensorflow-example:使用MNIST训练模型,并识别手写数字图片
Stars: ✭ 26 (+52.94%)
Mutual labels:  handwriting-recognition
PyRecognizer
"A neural network to rule them all, a neural network to find them, a neural network to bring them all and verify if is you !!" (Face recognition tool)
Stars: ✭ 28 (+64.71%)
Mutual labels:  mlp
SpeakerDiarization RNN CNN LSTM
Speaker Diarization is the problem of separating speakers in an audio. There could be any number of speakers and final result should state when speaker starts and ends. In this project, we analyze given audio file with 2 channels and 2 speakers (on separate channels).
Stars: ✭ 56 (+229.41%)
Mutual labels:  mlp
mlp stock
Stock price prediction using ensemble MLP in PyTorch.
Stars: ✭ 25 (+47.06%)
Mutual labels:  mlp
MachineLearning
Implementations of machine learning algorithm by Python 3
Stars: ✭ 16 (-5.88%)
Mutual labels:  mlp
PaperSynth
Handwritten text to synths!
Stars: ✭ 18 (+5.88%)
Mutual labels:  handwriting-recognition
kdd99-scikit
Solutions to kdd99 dataset with Decision tree and Neural network by scikit-learn
Stars: ✭ 50 (+194.12%)
Mutual labels:  mlp
rnnt decoder cuda
An efficient implementation of RNN-T Prefix Beam Search in C++/CUDA.
Stars: ✭ 60 (+252.94%)
Mutual labels:  handwriting-recognition
Handwriting-Recognition
Software to recognize handwriting
Stars: ✭ 46 (+170.59%)
Mutual labels:  handwriting-recognition
form-segmentation
Let's explore how we can extract text from forms
Stars: ✭ 42 (+147.06%)
Mutual labels:  handwriting-recognition
image-classification
A collection of SOTA Image Classification Models in PyTorch
Stars: ✭ 70 (+311.76%)
Mutual labels:  mlp
autonomio
Core functionality for the Autonomio augmented intelligence workbench.
Stars: ✭ 27 (+58.82%)
Mutual labels:  mlp
hwrt
A toolset for handwriting recognition
Stars: ✭ 61 (+258.82%)
Mutual labels:  handwriting-recognition
handwriting.js
A simple API access to the handwriting recognition service of Google IME
Stars: ✭ 89 (+423.53%)
Mutual labels:  handwriting-recognition

neural_net_handwriting

A demo project to brush up my machine learning skills by writing an image recognition neural network from scratch in C.

input sample

This uses a typical 3-layer multi-layer perceptron. Training time is about 3 seconds for 80% correct, and 30s for 96% correct against the competition MNIST set of hand-written digits, which is pretty speedy for a simple C program.

  • Hand-written linear algebra code in C, refined with asm inspection. Should have somewhat better cache efficiency than typical maths starter libraries for matrix multiplication.
  • Hand-written CSV parser for reading the data.
  • Hand-written image output for testing the data.

Note that I didn't include the test data files because they are pretty big. I used the CSV versions which you can download from https://pjreddie.com/projects/mnist-in-csv/.

cloc stats

Note that the project description data, including the texts, logos, images, and/or trademarks, for each open source project belongs to its rightful owner. If you wish to add or remove any projects, please contact us at [email protected].