EXPLORE CONNECTIONS
RILQ: Rank-Insensitive LoRA-based Quantization Error Compensation for Boosting 2-bit Large Language Model Accuracy
Follow the relationships that help you find your next source.
Connections are not prepared for this record yet, or its source metadata has changed. The original record remains available while background snapshots are rebuilt.