File size: 2,283 Bytes
d4776f3 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 |
# asmllm1.8-v1-instruct
**A specialized language model for comprehensive understanding of Assam, India**
*Optimized for efficiency on constrained hardware*
[](https://opensource.org/licenses/MIT)
## Overview
asmllm1.8-v1-instruct is a 1.8 billion parameter language model specifically designed to provide accurate, culturally relevant information about the Indian state of Assam. Developed for practical deployment on consumer-grade hardware (tested on RTX 2050), this model offers:
- Culturally-aware responses about Assam
- Efficient resource utilization
- Clear, concise answer formulation
- Broad coverage of Assam-related topics
## Key Features
- **Hardware Efficiency**: Optimized for deployment on laptops and mid-range GPUs
- **Cultural Context**: Trained on Assam-specific data corpus
- **Practical Focus**: Prioritizes clarity over unnecessary complexity
- **Diverse Knowledge Base**: Covers geography, history, economy, and culture
## 🗃️ Training Data
Curated dataset of 2,000 records on the state of Assam,India covering:
🌄 **Geography**
- Physical features
- Climate patterns
- Natural resources
🎨 **Cultural Heritage**
- Traditional festivals (Bihu, Ambubachi Mela)
- Indigenous arts and crafts
- Local dialects and literature
💼 **Economic Landscape**
- Tea production
- Oil and natural gas reserves
- Agricultural practices
🏰 **Historical Context**
- Ancient kingdoms (Kamarupa, Chutiya)
- Colonial history (British era)
- Architectural landmarks (Kamakhya Temple, Rang Ghar)
🎓 **Educational Framework**
- Major institutions
- Educational initiatives
## ⚙️ Model Specifications
| Parameter | Value |
|----------------------|--------------------|
| 🧠 **Architecture** | Transformer-based |
| 🔢 **Parameters** | 1.8B |
| 📜 **Context Window**| 1000 tokens |
| 🎯 **Precision** | bfloat16 |
| 💻 **Hardware** | RTX 2050 (4GB VRAM)|
## 👨💻 Developed By
**Name:** Anand Dey
**Role:** Developer
**Contact:** [📧 [email protected]](mailto:[email protected]) |