Large Language Models
InfLLM-V2: Dense-Sparse Switchable Attention for Seamless Short-to-Long Adaptation
MiniCPM4: Ultra-Efficient LLMs on End Devices