ReadMe.txt

-Further Testing: Includes testing with matrices and other not stable tests
-Baseline's files: basic testing files, + Completed stream communication and multi-dimensional matrices in struct for packet sending with bram-ddr read/write
-Basic Convolution Calcs: Includes the simplest version of the algorithms used
-Completed: Includes the latest versions of Accelerators(written in HLS) - Version 1 - Official & Tested
-Vitis_files: Inludes the latest FPGA implementations of the accelerators
-Matlab,C Windows Unet: Indcludes the UNET architecture, able to run on Windows(C) as a dummy neural network and the matlab dummy for benchmarking purposes. 
-Benchmarks: All the results -raw- , (for better visualization and comparison check the thesis .pdf file)
-2021: Version 2 Architecture with huge performance improvments, high FPGA utilization. HLS optimization - No comments hard to read - CAUTION! - Use the Version 1 and read thesis so you can understand the version 2. (Not Vitis Tested - reccomended resolution for Version 2: 256x256)


More information about Neural Network and functions read Higher level language implementations + Thesis by Labis Skoufis.