hybrid-gpt-moe-284m-a114m

Description

hybrid-gpt-moe-284m-a114m is an open-weight text-generation model with roughly 284M parameters published by garagelm on the Hugging Face Hub. It has 481 downloads. Runs with transformers.

Categories

• Coding

Tags

• moe_gpt
• text-generation
• mixture-of-experts
• moe
• hybrid-attention
• sliding-window
• small-model
• research

Capabilities

• Text generation
• Library: transformers
• 284M parameters