Implementation of the linear transformer-based decoder experiments from 5G LDPC Linear Transformer for Channel Decoding using the Sionna link-level simulator.
This work introduces a novel, fully differentiable linear-time complexity transformer decoder and a transformer decoder to correct 5G New Radio (NR) LDPC codes. We propose a scalable approach to decode linear block codes with O(n) complexity rather than O(n^2) for regular transformers. The architectures' performances are compared to Belief Propagation (BP), the production-level decoding algorithm used for 5G New Radio (NR) LDPC codes. We achieve bit error rate performance that matches a regular Transformer decoder and surpases one iteration BP, also achieving competitive time performance against BP, even for larger block codes. We utilize Sionna, Nvidia's 5G & 6G physical layer research software, for reproducible results.
LTD_model_5G_LDPC.ipynbrunning Decoder on 5G LDPC codes.
In src/ folder:
decoder.pylinear and regular transformer decoder model.decoder5G.pydecoder model for 5G LDPC codes.e2e_model.pyend-to-end channel between Encoder and Decoder.args.pyclass holding arguments for the models and training.utils.pysave, load, train/test functions for transformer models.utils5G.py5G pcm pruning function.time_comparison.pycomparison and plotting functions to evaluate speed of models.
[A] M. Hernandez, F. Pinero, "5G LDPC Linear Transformer for Channel Decoding", 2025
[C] Yoni Choukroun, Lior Wolf, "Error Correction Code Transformer", 2022
© 2025 Mario Hernandez Torres.
This project is licensed under the Apache License 2.0.
Distributed on an "AS IS" basis, without warranties or conditions of any kind.