xesutr/aom-news-4b
xesutr/aom-news-4b is a 4.3 billion parameter language model fine-tuned from Google Gemma 4B by xesutr. This model is specifically optimized for structured news analysis, generating a concise summary, sentiment score, and numerical score for financial and cryptocurrency articles. It is designed for efficient local inference on CPU and low-resource environments, making it suitable for edge computing applications.
Loading preview...
Model Overview
xesutr/aom-news-4b is a specialized 4.3 billion parameter language model, fine-tuned from Google Gemma 4B, designed for structured news analysis. Its core function is to process financial and cryptocurrency articles and output a concise summary, a sentiment score, and a numerical score. This model is particularly optimized for CPU and low-resource environments, making it highly efficient for local inference and edge computing.
Key Capabilities
- Structured News Analysis: Generates a predefined output format including a summary, sentiment (e.g., POSITIVE), and a numerical score (e.g., 0.9) for news articles.
- Financial & Cryptocurrency Focus: Specifically trained to analyze content within these domains.
- Resource-Efficient: Optimized for deployment on CPUs and devices with limited resources, offering high inference speed.
Important Limitations
- Specialized Task Only: This model is aggressively fine-tuned for its specific structural analysis task and is not suitable for general conversation. It will attempt to force all inputs into its defined output format.
Available Formats
The model is provided in two primary formats to support diverse deployment needs:
- Safetensors: Ideal for further fine-tuning or high-performance GPU deployments where maximum accuracy is critical.
- GGUF (Quantized): Recommended for Ollama, Llama.cpp, and CPU-constrained environments due to its smaller size and optimized inference speed on standard hardware.