Syed Abdul Gaffar Shakhadri, Kruthika KR, Rakshit Aralimatti
We introduce Shakti, a 2.5 billion parameter language model specifically optimized for resource-constrained environments such as edge devices, including smartphones, wearables, and IoT systems. Shakti combines high-performance NLP with optimized efficiency and precision, making it ideal for real-time AI applications where computational resources and memory are limited. With support for vernacular languages and domain-specific tasks, Shakti excels in industries such as healthcare, finance, and customer service. Benchmark evaluations demonstrate that Shakti performs competitively against larger models while maintaining low latency and on-device efficiency, positioning it as a leading solution for edge AI.
| Task | Dataset | Metric | Value | Model |
|---|---|---|---|---|
| Question Answering | HellaSwag | Accuracy | 52.4 | Shakti-LLM (2.5B) |
| Question Answering | MedQA | Accuracy | 60.3 | Shakti-LLM (2.5B) |
| Question Answering | BBH | Accuracy | 58.2 | Shakti-LLM (2.5B) |
| Question Answering | MML | Accuracy | 71.8 | qwen-LLM 7B |
| Question Answering | TruthfulQA | Accuracy | 68.4 | Shakti-LLM (2.5B) |
| Question Answering | PIQA | Accuracy | 86.2 | Shakti-LLM (2.5B) |
| Question Answering | BoolQ | Accuracy | 61.1 | Shakti-LLM (2.5B) |
| Question Answering | TriviaQA | EM | 58.2 | Shakti-LLM (2.5B) |