Overview

The mwhyd2262/Alita-V4-Full-Merged is a 14.8 billion parameter instruction-tuned language model based on the Qwen2.5 architecture. It was developed by mwhyd2262 and finetuned from the unsloth/qwen2.5-14b-instruct-bnb-4bit model. A notable aspect of its development is the use of Unsloth and Huggingface's TRL library, which facilitated a 2x faster training process.

Key Capabilities

• Instruction Following: Designed to accurately follow a wide range of user instructions.
• Qwen2.5 Architecture: Benefits from the robust capabilities and performance characteristics of the Qwen2.5 base model.
• Efficient Training: Utilized Unsloth for accelerated finetuning, indicating potential for efficient deployment or further customization.

Good For

• General-purpose instruction-following applications.
• Tasks requiring a capable 14.8B parameter model with a 32K context length.
• Developers interested in models finetuned with Unsloth for performance and efficiency.