|
|
--- |
|
|
license: apache-2.0 |
|
|
language: |
|
|
- en |
|
|
base_model: |
|
|
- EssentialAI/rnj-1-instruct |
|
|
base_model_relation: finetune |
|
|
library_name: peft |
|
|
tags: |
|
|
- tactical-intelligence |
|
|
- vanta-research |
|
|
- scout-8b |
|
|
- conversational |
|
|
- conversational-ai |
|
|
- research |
|
|
- roleplay |
|
|
- chat |
|
|
- chat-ai |
|
|
- chatbot |
|
|
- reasoning |
|
|
- logic |
|
|
- science |
|
|
- STEM |
|
|
- constraint-aware |
|
|
--- |
|
|
|
|
|
<div align="center"> |
|
|
|
|
|
 |
|
|
|
|
|
<h1>VANTA Research</h1> |
|
|
|
|
|
<p><strong>Independent AI research lab building safe, resilient language models optimized for human-AI collaboration</strong></p> |
|
|
|
|
|
<p> |
|
|
<a href="https://vantaresearch.xyz"><img src="https://img.shields.io/badge/Website-vantaresearch.xyz-black" alt="Website"/></a> |
|
|
<a href="https://merch.vantaresearch.xyz"><img src="https://img.shields.io/badge/Merch-merch.vantaresearch.xyz-sage" alt="Merch"/></a> |
|
|
<a href="https://x.com/vanta_research"><img src="https://img.shields.io/badge/@vanta_research-1DA1F2?logo=x" alt="X"/></a> |
|
|
<a href="https://github.com/vanta-research"><img src="https://img.shields.io/badge/GitHub-vanta--research-181717?logo=github" alt="GitHub"/></a> |
|
|
</p> |
|
|
</div> |
|
|
|
|
|
--- |
|
|
|
|
|
|
|
|
|
|
|
# Scout-8B: Tactical Intelligence for Edge Devices |
|
|
|
|
|
**VANTA Research Entity-002: The Reconnaissance Specialist** |
|
|
|
|
|
## Overview |
|
|
|
|
|
Scout-8B is a specialized AI model built for tactical intelligence and reconnaissance operations. Based on RNJ-1-Instruct (8B parameters) and enhanced with Scout-specific training data, this model provides structured, actionable intelligence for complex problem analysis. This model contains all of the same data used in [Scout-4B](https://huggingface.co/vanta-research/scout-4b) - and not only improves, but expands on previous capabilities. |
|
|
|
|
|
|
|
|
## Capabilities |
|
|
|
|
|
**Tactical Intelligence Analysis** |
|
|
- Systematic problem decomposition |
|
|
- Structured reconnaissance approach |
|
|
- Data-driven assessment methodology |
|
|
|
|
|
**Operational Planning** |
|
|
- Multi-phase operation planning |
|
|
- Risk assessment and mitigation |
|
|
- Resource allocation guidance |
|
|
|
|
|
**Technical Assessment** |
|
|
- Architecture evaluation and analysis |
|
|
- Performance optimization recommendations |
|
|
- Security perimeter assessment |
|
|
|
|
|
## Usage |
|
|
|
|
|
Scout is design and optimized for the following styles of interaction: |
|
|
- Direct, professional communication style |
|
|
- Tactical terminology usage |
|
|
- Structured, phased approaches |
|
|
- Focus on actionable outcomes |
|
|
|
|
|
## Files |
|
|
|
|
|
- `config.json` - Model configuration |
|
|
- `scout_config.json` - Scout-specific settings |
|
|
- `model-*.safetensors` - Merged model weights |
|
|
- `tokenizer.*` - Tokenizer files |
|
|
|
|
|
## Training Details |
|
|
|
|
|
- **Method**: LoRA (Low-Rank Adaptation) |
|
|
- **Rank**: 16, Alpha: 32 |
|
|
- **Epochs**: 2 |
|
|
- **Dataset**: ~4,500 Scout-specific synthetically generated examples |
|
|
- **Target Modules**: q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj |
|
|
|
|
|
## Next Steps |
|
|
|
|
|
1. **Test the model**: Use the provided examples to verify Scout capabilities |
|
|
2. **Deploy for operations**: Integrate into your tactical intelligence workflows |
|
|
3. **Customize further**: Additional fine-tuning for specific operational contexts |
|
|
|
|
|
## Contact |
|
|
- Organization: hello@vantaresearch.xyz |
|
|
- Engineering/Design: tyler@vantaresearch.xyz |
|
|
|
|
|
*Proudly developed by VANTA Research in Portland, Oregon* |
|
|
|
|
|
|
|
|
|
|
|
|
