Publications

(2025). Q-Adam-mini: Memory-Efficient 8-bit Quantized Optimizer for Large Language Model Training. In ICML 2025 Workshop ES-FoMo-III.