Dragon-Alpha&cu32: A Java-based Tensor Computing Framework With its High-Performance CUDA Library
This work addresses the problem of limited Java-based deep learning frameworks for developers in that ecosystem, though it is incremental as it builds on existing concepts with performance optimizations.
The authors tackled the underutilization of Java in deep learning by developing Dragon-Alpha, a Java-based tensor computing framework with a high-performance CUDA library (cu32), which reduced training time by 75.38% to 97.32% and memory usage by 29.2% to 66.4% compared to PyTorch and cuDNN on networks like AlexNet and ResNet using Cifar-10.
Java is very powerful, but in Deep Learning field, its capabilities probably has not been sufficiently exploited. Compared to the Java-based deep-learning-frameworks, the Python-based (PyTorch, TensorFlow, etc) are undoubtedly the mainstream, due to their easy-to-use, flexibility and better ecosystem. Dragon-Alpha is a Java-based Tensor Computing Framework, with easy-to-use, high-scalability and high-performance, trying to break Java's dilemma in deep learning field and make it more effective. Dragon-Alpha supports different levels of APIs, and can be used as a deep-learning-framework through its user-friendly high-level APIs. Dragon-Alpha has potential to aggregate computing-power across heterogeneous platforms and devices, based on its multi-layer architecture and Java's big-data ecosystem. Dragon-Alpha has its asynchronized APIs to improve parallelism, and highly-optimized CUDA library cu32 which adopts unique convolution\deconvolution operators for small feature maps. The experiments show that, compared to PyTorch&cuDNN, Dragon-Alpha&cu32 costs less time and memory (75.38% to 97.32%, 29.2% to 66.4%), to train some typical neural networks (AlexNet, VGG, GoogleNet, ResNet) on Cifar-10.