No Saved Kaleidosope: an 100% Jitted Neural Network Coding Language with Pythonic Syntax
This is an incremental improvement for machine learning practitioners, offering a new compiler tool with mixed performance outcomes.
The authors developed a jitted compiler for training neural networks with Pythonic syntax, achieving similar speed and performance as PyTorch on CIFAR-10, but showing degraded results in ImageNet and GRU network tasks.
We developed a jitted compiler for training Artificial Neural Networks using C++, LLVM and Cuda. It features object-oriented characteristics, strong typing, parallel workers for data pre-processing, pythonic syntax for expressions, PyTorch like model declaration and Automatic Differentiation. We implement the mechanisms of cache and pooling in order to manage VRAM, cuBLAS for high performance matrix multiplication and cuDNN for convolutional layers. Our experiments with Residual Convolutional Neural Networks on ImageNet, we reach similar speed but degraded performance. Also, the GRU network experiments show similar accuracy, but our compiler have degraded speed in that task. However, our compiler demonstrates promising results at the CIFAR-10 benchmark, in which we reach the same performance and about the same speed as PyTorch. We make the code publicly available at: https://github.com/NoSavedDATA/NoSavedKaleidoscope