BanglaLLM/Bangla-s1k-qwen-2.5-3B-Instruct

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:3.1BQuant:BF16Ctx Length:32kPublished:Mar 8, 2025Architecture:Transformer0.0K Warm

BanglaLLM/Bangla-s1k-qwen-2.5-3B-Instruct is a 3.1 billion parameter instruction-tuned causal language model, fine-tuned from Qwen/Qwen2.5-3B-Instruct. Developed by BanglaLLM, this model specializes in understanding and generating content in Bengali. It is specifically optimized for tasks requiring strong performance in the Bengali language, leveraging the s1k-Bangla-qwen dataset for its fine-tuning.

Loading preview...

Overview

BanglaLLM/Bangla-s1k-qwen-2.5-3B-Instruct is a 3.1 billion parameter instruction-tuned language model, built upon the Qwen2.5-3B-Instruct architecture. This model has been specifically fine-tuned by BanglaLLM using the BanglaLLM/s1k-Bangla-qwen dataset, focusing on enhancing its capabilities for the Bengali language.

Key Capabilities

  • Bengali Language Proficiency: Optimized for understanding and generating text in Bengali.
  • Instruction Following: Designed to follow instructions effectively, making it suitable for various NLP tasks.
  • Fine-tuned Performance: Leverages the TRL framework for supervised fine-tuning (SFT) to improve task-specific performance.

Training Details

The model was trained using the TRL library (version 0.12.0) with Transformers (4.46.1) and PyTorch (2.5.1). The fine-tuning process utilized the s1k-Bangla-qwen dataset, which is tailored for Bengali language tasks, ensuring specialized performance in this domain.

Good For

  • Applications requiring strong Bengali language generation.
  • Instruction-based tasks in Bengali.
  • Research and development in Bengali NLP.