New PYNQ Jupyter Notebook demonstrates Neural Network Transfer Learning using Astronomical Images of Merging Galaxies で知った、ZipMP-PYNQ
プレゼンテーション資料:FPGA-accelerated Dense Linear Machine Learning:A Precision-Convergence Trade-off
論文は、The ZipML Framework for Training Models with End-to-End Low Precision : The Cans, the Cannots, and a Little Bit of Deep Learning
こっちは、36頁もの。
論文は、The ZipML Framework for Training Models with End-to-End Low Precision : The Cans, the Cannots, and a Little Bit of Deep Learning
こっちは、36頁もの。
このStratix VをPYNQに実装か。。。
引用 The top files for training with single-precision floating-point data and quantized data are floatFSGD.vhd and qFSGD.vhd, respectively. The top level interface for both modules is AXI-Stream.
VHDLコードですね。
PYNQ用の bitstream もあります。
ここには、Notebooksが