Shaista-pro: An Urdu-Optimized 8B Language Model

Shaista-pro is an 8 billion parameter causal language model developed by ReySajju742, fine-tuned from the Qalb-1.0-8B-Instruct base model, which itself is built upon Llama-3 8B. This model is specifically designed to excel in Urdu language tasks, offering full model weights (16.1 GB) for comprehensive deployment.

Key Capabilities

• High-Quality Urdu Text Generation: Optimized for creating fluent and contextually relevant Urdu text.
• Logical Reasoning in Urdu: Demonstrates strong performance in understanding and responding to complex logical queries.
• Urdu Poetry Generation: Capable of generating creative and coherent Urdu poetry.
• Instruction Following: Effectively processes and executes instructions provided in Urdu.
• Multilingual Support: While primarily focused on Urdu, it also offers secondary support for English.

Good for

• Developers and researchers working on Urdu-centric NLP applications.
• Applications requiring nuanced Urdu text generation, such as content creation or chatbots.
• Projects involving logical reasoning or creative writing in the Urdu language.
• Integrating a robust Urdu language model into various platforms with its full weight availability.