SHAKTI: A 2.5 Billion Parameter Small Language Model Optimized for Edge AI and Low-Resource Environments Paper ⢠2410.11331 ⢠Published Oct 15, 2024 ⢠8