Machine Learning Akisamb • 6 months ago • 100%

Qwen1.5-MoE-A2.7B: A Small MoE Model with only 2.7B Activated Parameters yet Matching the Performance of State-of-the-Art 7B models

Comments 0