VN-Noxa-v1-7b-Beta-Low

VN-Noxa v1 7b (Beta-Low)

VN-Noxa v1 7b (Beta-Low) is an ultra-lightweight variant of the open-source Large Language Model (LLM) lineup developed and maintained by Faizz, the Founder of NovaVest.

The LOW version is compressed using 2-bit quantization (Q2_K). The primary focus of this variant is maximum inference speed (high tokens-per-second generation) and minimal hardware footprint, making it incredibly friendly to run on low-end hardware or machines with highly restricted memory.

🚀 Key Features

• Ultra-Fast Inference: Provides lightning-fast responses for core trading logic and day-to-day coding queries.
• Hardware Friendly: Engineered specifically to bypass memory bottlenecks on older laptops or budget setups with tight RAM configurations.
• Professional Context: Retains VN-Noxa's signature well-structured, systematic, professional, and educational communication style despite the high compression rate.

📊 Model Details

• Developer: Faizz Bangkit Adhi Pratama (Founder of NovaVest)
• Base Architecture: Qwen 2.5 Coder 7B
• Fine-Tuning Method: QLoRA (Unsloth)
• File Format: Single .gguf file
• File Size: ~2 GB - 2.5 GB
• Minimum Hardware Requirements: At least 4GB - 8GB RAM Laptop/PC (Extremely lightweight to run locally via LM Studio / Ollama).

🛠️ Usage (Offline Inference)

1. Download the super-lightweight .gguf file from this repository.
2. Load the file into a local LLM application such as LM Studio or Ollama.
3. Use the following chat template format for optimal performance:

<|im_start|>system
Anda adalah VN-Noxa v1 7b, asisten AI profesional untuk quantitative trading dan coding.<|im_end|>
<|im_start|>user
[Your Question Here]<|im_end|>
<|im_start|>assistant