hishab/titulm-llama-3.2-1b-v1.1

The hishab/titulm-llama-3.2-1b-v1.1 is a 1 billion parameter Llama 3.2-based auto-regressive language model continually pre-trained by Hishab on extensive Bangla datasets. This model is specifically optimized for high-quality Bangla text generation and understanding tasks, demonstrating superior performance in Bangla language evaluation benchmarks. It supports a context length of 32768 tokens and is primarily intended for Bangla text generation and language understanding applications, with secondary support for English.