About SmallTrain

  SmallTrain is an open-source deep learning framework developed by Geek Guild Co., Ltd. for the rapid development of deep neural networks by reducing the effort of acquiring a large amount of data without compromising the accuracy of the problems that it encounters.

What is SmallTrain?



SmallTrain is a wrapper able to use TensorFlow and Pytorch, It’s designed as a hi-accuracy model.
SmallTrain is a deep learning library adopted the-state-of-the-art algorithms from the newest paper and our own algorithms.
SmallTrain is a platform where you can easily build your own trained model with additional learning with your small data can create your high-accuracy model with only a few lines of programming is required.
See the different sections of the documentation and site for more ideas.

Ways to create AI models:



Ways to create AI models:
- Build your model by learning all the data science algorithms and mathematics.
- Build using libraries such as Tensorflow that does the algorithms and mathematical part
- Build using wrappers that use the library for ease of development(for rapid development)

The easiest of these three ways is to use the third wrapper. You can use SmallTrain as this wrapper.

How are we different from Keras and other wrappers library?




- Keras and other wrapper is suitable for PoC but in complex problems it fails and as a wrapper it has a lot of limitations, whereas SmallTrain is a wrapper that closely relates with models in architecture level but still maintaining the ease of development that Keras and other wrapper provides.

- SmallTrain uses a base neural network of about 100 layers of pyramid architecture neural network build referring to state-of-the-art papers and implemented the cutting edge algorithms.

- SmallTrain supports Tensorflow and PyTorch and is designed in a way that they both are non-intrusive with each other and in many cases aids in development according to the style that user needs.

- SmallTrain will train quickly with the data that you can provide and it is highly versatile and supports all kinds of data, and evolves with various range of data.

How to Use SmallTrain




See the different sections of the documentation and site for more ideas.

Great features of SmallTrain

- Easy to develop for POC to production.
- Almost no programming for building your pre-trained model.
- Available as both TensorFlow and Pytorch wrapper.
- Always adapting to algorithms which evolve.
- Build using state-of-the-art Algorithms from Scientific and mathematical papers.
- Accuracy is always going to be better even with minimal data and training time.
- Licensed under MIT Open Source.

SmallTrain Problem space bubble

- Image recognition
- Time series data forcasting
- Object detection
- Anomaly detection
- Voice recognition