ViteMamba/TigerLLM-Medical-Bengali

TEXT GENERATIONConcurrency Cost:1Model Size:1BQuant:BF16Ctx Length:32kPublished:Jun 8, 2026License:apache-2.0Architecture:Transformer Open Weights Cold

The ViteMamba/TigerLLM-Medical-Bengali is a 1 billion parameter language model based on the Gemma3 architecture, fine-tuned by ViteMamba. This model specializes in Bengali medical question answering, leveraging QLoRA fine-tuning on a dedicated Bangla medical QA dataset. It is optimized for providing accurate responses to medical queries in Bengali, making it suitable for healthcare-related applications requiring localized language support.

Loading preview...

TigerLLM-Medical-Bengali Overview

The ViteMamba/TigerLLM-Medical-Bengali is a 1 billion parameter language model built upon the Gemma3 architecture, specifically fine-tuned for medical question answering in Bengali. It utilizes QLoRA (4-bit + LoRA) for efficient fine-tuning, enhancing its performance on specialized medical dialogues.

Key Capabilities

  • Bengali Medical QA: Designed to understand and generate responses to medical questions exclusively in Bengali.
  • Efficient Fine-tuning: Leverages QLoRA on a base TigerLLM-1B-it model, making it resource-efficient for deployment.
  • Specialized Dataset: Fine-tuned on a dataset of 901 Bangla medical QA samples, ensuring domain-specific knowledge.
  • Context Length: Supports a context length of 32768 tokens, allowing for more comprehensive medical discussions.

Good For

  • Healthcare Applications: Ideal for chatbots, virtual assistants, or information systems requiring medical knowledge in Bengali.
  • Localized Medical Support: Provides a specialized solution for users seeking medical information in the Bengali language.
  • Research and Development: Useful for researchers and developers working on Bengali NLP in the medical domain.