Taobao's MNN Team Open-Sources MNN-Pocket-0.3B, a 173MB On-Device Model Built for Phone Apps
Alibaba's MNN team at Taobao open-sourced MNN-Pocket-0.3B, a 173MB INT4 on-device model it says decodes about 350 tokens per second offline on an iPhone 18 Pro.
Sources
- T2Taobao's MNN Team Open-Sources MNN-Pocket-0.3B, a 173MB On-Device Model Built for Phone AppsChina: SCMP Tech / TechNode / Pandaily / KrASIA / Sixth Tone