Blog
Fine-Tuning Small Language Models for Edge RoboticsExploring LoRA adaptation, QLoRA, and 4-bit quantization benchmarks for real-time inference on resource-constrained embedded systems.
By Md. Shakhoyat Rahman Shujon
Oct 10, 2026
K-MINDS Research
Research notes, reading summaries, and published writeups—a living archive of the ideas and experiments shaping our work.
Exploring LoRA adaptation, QLoRA, and 4-bit quantization benchmarks for real-time inference on resource-constrained embedded systems.
A comparative evaluation of self-supervised speech representations (Wav2Vec 2.0 & HuBERT) for low-resource Bengali speech recognition.