kda-graft-qwen2.5-0.5b-linear-attention-3to1

Description

kda-graft-qwen2.5-0.5b-linear-attention-3to1 is an open-weight text-generation model with roughly 0.5B parameters published by Rob1234567 on the Hugging Face Hub. It has 712 downloads. Runs with pytorch.

Community Metrics

Categories

• Productivity

Tags

• qwen2
• kda
• kimi-delta-attention
• linear-attention
• hybrid-attention
• weight-surgery
• model-grafting
• long-context

Capabilities

• Text generation
• Conversational
• Library: pytorch
• 0.5B parameters

Information