Update README.md
Browse files
README.md
CHANGED
|
@@ -1,66 +1,74 @@
|
|
| 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 |
-
**
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
tags:
|
| 4 |
+
- assam
|
| 5 |
+
- custom-dataset
|
| 6 |
+
- llama-architecture
|
| 7 |
+
- hardware-contrained-infra
|
| 8 |
+
---
|
| 9 |
+
# asmllm1.8-v1-instruct
|
| 10 |
+
|
| 11 |
+
**A specialized language model for comprehensive understanding of Assam, India**
|
| 12 |
+
*Optimized for efficiency on constrained hardware*
|
| 13 |
+
|
| 14 |
+
[](https://opensource.org/licenses/MIT)
|
| 15 |
+
|
| 16 |
+
## Overview
|
| 17 |
+
|
| 18 |
+
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:
|
| 19 |
+
|
| 20 |
+
- Culturally-aware responses about Assam
|
| 21 |
+
- Efficient resource utilization
|
| 22 |
+
- Clear, concise answer formulation
|
| 23 |
+
- Broad coverage of Assam-related topics
|
| 24 |
+
|
| 25 |
+
## Key Features
|
| 26 |
+
|
| 27 |
+
- **Hardware Efficiency**: Optimized for deployment on laptops and mid-range GPUs
|
| 28 |
+
- **Cultural Context**: Trained on Assam-specific data corpus
|
| 29 |
+
- **Practical Focus**: Prioritizes clarity over unnecessary complexity
|
| 30 |
+
- **Diverse Knowledge Base**: Covers geography, history, economy, and culture
|
| 31 |
+
|
| 32 |
+
## ποΈ Training Data
|
| 33 |
+
|
| 34 |
+
Curated dataset of 2,000 records on the state of Assam,India covering:
|
| 35 |
+
|
| 36 |
+
π **Geography**
|
| 37 |
+
- Physical features
|
| 38 |
+
- Climate patterns
|
| 39 |
+
- Natural resources
|
| 40 |
+
|
| 41 |
+
π¨ **Cultural Heritage**
|
| 42 |
+
- Traditional festivals (Bihu, Ambubachi Mela)
|
| 43 |
+
- Indigenous arts and crafts
|
| 44 |
+
- Local dialects and literature
|
| 45 |
+
|
| 46 |
+
πΌ **Economic Landscape**
|
| 47 |
+
- Tea production
|
| 48 |
+
- Oil and natural gas reserves
|
| 49 |
+
- Agricultural practices
|
| 50 |
+
|
| 51 |
+
π° **Historical Context**
|
| 52 |
+
- Ancient kingdoms (Kamarupa, Chutiya)
|
| 53 |
+
- Colonial history (British era)
|
| 54 |
+
- Architectural landmarks (Kamakhya Temple, Rang Ghar)
|
| 55 |
+
|
| 56 |
+
π **Educational Framework**
|
| 57 |
+
- Major institutions
|
| 58 |
+
- Educational initiatives
|
| 59 |
+
|
| 60 |
+
## βοΈ Model Specifications
|
| 61 |
+
|
| 62 |
+
| Parameter | Value |
|
| 63 |
+
|----------------------|--------------------|
|
| 64 |
+
| π§ **Architecture** | Transformer-based |
|
| 65 |
+
| π’ **Parameters** | 1.8B |
|
| 66 |
+
| π **Context Window**| 1000 tokens |
|
| 67 |
+
| π― **Precision** | bfloat16 |
|
| 68 |
+
| π» **Hardware** | RTX 2050 (4GB VRAM)|
|
| 69 |
+
|
| 70 |
+
## π¨π» Developed By
|
| 71 |
+
|
| 72 |
+
**Name:** Anand Dey
|
| 73 |
+
**Role:** Developer
|
| 74 |
+
**Contact:** [π§ [email protected]](mailto:[email protected])
|