Enhancing the Decoding Rates of BATS Codes by Learning with Guided Information

This repo is for the paper [1] accepted by ISIT2022. Please refer to the accompanying paper [1] for detailed descriptions.

Requirement

• python >= 3.7
• galois ~= 0.0.21
• torch ~= 1.10.0
• GPU support optional

Quick Start

python train.py --n_pkt 64 --n_batch 10 --batch_size 8 --num_hops 10 --loss_rate 0.1

trained weights available upon request. contact: Jack Qing ([email protected])

Reference

[1] J. Qing, H. H. F. Yin, and R. W. Yeung, “Enhancing the Decoding Rates of BATS Codes by Learning with Guided Information” in Proc. ISIT’22, Espoo, Finland, Jun. 2022.