MiniCPM Language Models
12 repos
Compact, efficient language models designed for conversational AI and text generation across various scales and training approaches. The cluster centers on MiniCPM variants—including 1B and 2B parameter models trained with supervised fine-tuning (SFT), direct preference optimization (DPO), and extended context windows (128k tokens)—along with related inference and deployment implementations. These models prioritize compatibility with standard transformer endpoints while maintaining conversational quality at reduced computational cost.
text-generation ·3,760
conversational ·3,760
transformers ·3,760
endpoints_compatible ·3,122
minicpm ·2,990
llama ·2,701
edge-ai ·2,701
on-device ·2,701
minicpm5 ·2,701
long-context ·2,701
