ramyaa1113/gemma2b-webxr-showroom-v2
The ramyaa1113/gemma2b-webxr-showroom-v2 is a 2.5 billion parameter Gemma 2B-based causal language model, fine-tuned by Rajalakshmi Mahadevan (Ramya). Optimized for AI-assisted interactions, it excels at providing product explanations and guidance within WebXR virtual showrooms and immersive 3D environments. This model is specifically designed to power conversational assistants for virtual retail and interactive product exploration.
Loading preview...
Model Overview
gemma2b-webxr-showroom-v2 is a 2.5 billion parameter conversational model, fine-tuned from Google's Gemma 2B by Rajalakshmi Mahadevan (Ramya). It is specifically designed to function as an AI showroom assistant within immersive WebXR applications, enhancing 3D product experiences.
Key Capabilities
- Product Explanations: Generates detailed descriptions of product features and functionalities.
- Interactive Showroom Guidance: Assists users in navigating and understanding virtual environments.
- Conversational Product Queries: Handles user questions related to products in an immersive context.
- Virtual Retail Assistance: Powers AI assistants for e-commerce and product demonstration in WebXR.
Training and Optimization
The model was trained on a custom dataset of approximately 15,000 samples, curated to simulate real user interactions in virtual showrooms, including product explanations, feature descriptions, and retail dialogues. Training was conducted on Kaggle GPU notebooks using mixed precision (fp16).
Intended Use Cases
This model is ideal for integration into:
- WebXR virtual product showrooms
- Immersive e-commerce experiences
- AI guides within 3D environments
- Conversational interfaces for immersive applications
It is part of the IntelliShop XR project, demonstrating AI's role in enhancing WebXR product exploration. The model is domain-tuned for interactive product assistance, and its performance may degrade for unrelated topics. Formal benchmark evaluation is ongoing, with current focus on qualitative testing within XR interaction scenarios.