ErtasAI/Llama-3.2-3B-Instruct
ErtasAI/Llama-3.2-3B-Instruct is a 3.21 billion parameter instruction-tuned generative language model developed by Meta. Part of the Llama 3.2 collection, it utilizes an optimized transformer architecture and is fine-tuned for multilingual dialogue use cases, including agentic retrieval and summarization tasks. This model supports a 32768 token context length and is optimized for on-device deployment due to its efficient quantization schemes.
Loading preview...
Overview
ErtasAI/Llama-3.2-3B-Instruct is a 3.21 billion parameter instruction-tuned model from Meta's Llama 3.2 family. It is built upon an optimized transformer architecture and leverages supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) for alignment. The model is designed for multilingual text-in/text-out applications, supporting a context length of 32768 tokens and trained on up to 9 trillion tokens of publicly available data with a December 2023 cutoff.
Key Capabilities
- Multilingual Dialogue: Optimized for multilingual chat and agentic applications, including knowledge retrieval and summarization, with official support for English, German, French, Italian, Portuguese, Hindi, Spanish, and Thai.
- On-Device Deployment: Features advanced quantization schemes (SpinQuant, QLoRA) designed for efficient inference on constrained environments like mobile devices, significantly reducing model size and memory footprint while improving decode and prefill speeds.
- Performance: Outperforms many open-source and closed chat models on common industry benchmarks, demonstrating strong capabilities in MMLU, instruction following, and mathematical reasoning for its size class.
Intended Use Cases
This model is ideal for commercial and research use in:
- Assistant-like chat and agentic applications.
- Knowledge retrieval and summarization.
- Mobile AI-powered writing assistants and query/prompt rewriting.
- On-device applications requiring efficient, smaller language models.