sumitsen/BanglaGptAi1.0
BanglaGptAi1.0 by Sumit Sen is a 0.5 billion parameter instruction-tuned causal language model based on the Qwen2.5 architecture, featuring a 32,768 token context length. It demonstrates enhanced capabilities in coding, mathematics, and instruction following, alongside robust multilingual support for over 29 languages including Bengali. This model is optimized for generating long texts, understanding structured data, and producing structured outputs like JSON, making it suitable for diverse chatbot and content generation tasks.
Loading preview...
Overview
BanglaGptAi1.0 is the inaugural release from Sumit Sen, built upon the Qwen2.5 architecture. This 0.5 billion parameter instruction-tuned causal language model is designed with a full context length of 32,768 tokens and can generate up to 8,192 tokens.
Key Capabilities
- Enhanced Knowledge & Reasoning: Significantly improved performance in coding and mathematics, leveraging specialized expert models.
- Instruction Following: Demonstrates strong instruction following, generating long texts, and understanding structured data such as tables.
- Structured Output Generation: Excels at producing structured outputs, particularly JSON, and is resilient to diverse system prompts for better role-play and chatbot implementation.
- Multilingual Support: Offers comprehensive support for over 29 languages, including Bengali, English, Chinese, French, Spanish, and more.
- Architectural Features: Utilizes transformers with RoPE, SwiGLU, RMSNorm, Attention QKV bias, and tied word embeddings.
Good For
- Developers requiring a model with strong coding and mathematical reasoning abilities.
- Applications needing robust instruction following and the generation of structured data.
- Chatbot implementations that benefit from resilient system prompt handling and role-play capabilities.
- Multilingual content generation, especially for languages like Bengali and others supported.