WorldParser-0.5B-1903-16bit Overview

WorldParser-0.5B-1903-16bit is a compact 0.5 billion parameter instruction-tuned model developed by fevohh. It is based on the Qwen2.5 architecture and was finetuned from unsloth/Qwen2.5-0.5B-Instruct-unsloth-bnb-4bit. A key aspect of its development is the utilization of Unsloth and Huggingface's TRL library, which facilitated a 2x faster training process.

Key Characteristics

• Architecture: Qwen2.5-based causal language model.
• Parameter Count: 0.5 billion parameters, making it suitable for resource-constrained environments.
• Training Efficiency: Benefited from Unsloth for accelerated finetuning.
• License: Distributed under the Apache-2.0 license.

Good For

• Applications requiring a lightweight yet capable instruction-following model.
• Scenarios where efficient deployment and faster inference are critical due to its smaller size.
• Developers interested in models finetuned with Unsloth for optimized training.