kz919/QwQ-0.5B-Distilled
TEXT GENERATIONConcurrency Cost:1Model Size:0.5BQuant:BF16Ctx Length:32kPublished:Dec 28, 2024License:apache-2.0Architecture:Transformer0.0K Open Weights Warm

kz919/QwQ-0.5B-Distilled is a 0.5 billion parameter causal language model developed by kz919, distilled from Qwen/QwQ-32B-Preview using Generative Knowledge Distillation (GKD). Based on Qwen2-0.5B-Instruct, it is optimized for conversational AI and reasoning tasks, featuring step-by-step problem-solving capabilities and a 32768-token context length. This model is designed for efficient deployment in applications requiring long-context understanding and logical inference.

Loading preview...