74
NepaliGPT-2.0 is an advanced text generation model developed by PrinceLab Pvt. Ltd. This model is part of the ongoing effort to democratize artificial intelligence through open source and open science. Built on the robust meta-llama/Llama-3.1-8B architecture, NepaliGPT-2.0 has been finetuned to enhance its performance in generating text in both English and Nepali. Utilizing cutting-edge technologies such as Transformers and Safetensors, this model ensures efficient and accurate text generation. The training process was accelerated using Unsloth and Hugging Face's TRL library, making it twice as fast as conventional methods. With a model size of 8 billion parameters and employing the BF16 tensor type, NepaliGPT-2.0 is designed for high-performance applications. The model is available under the MIT license, promoting open collaboration and innovation. Although not currently deployed by any inference provider, users can train and deploy this model within their own environments, benefiting from its powerful capabilities in text generation.
Built with