MiMo-V2.6：將強化學習擴展至自我改進
MiMo-V2.6: Scaling Reinforcement Learning Towards Self-Improvement
arXiv
110
Liang-Hsun Huang 許願
MiniMax-M1：以閃電注意力高效地擴展測試時運算
MiniMax-M1: Scaling Test-Time Compute Efficiently with Lightning Attention
arXiv
0
Liang-Hsun Huang 許願
MiniCPM：以可擴展的訓練策略揭示小型語言模型的潛力
MiniCPM: Unveiling the Potential of Small Language Models with Scalable Training Strategies
arXiv
23
Liang-Hsun Huang 許願
訓練長上下文視覺語言模型：以 128K 之外的泛化能力為目標的有效訓練方案
Training Long-Context Vision-Language Models Effectively with Generalization Beyond 128K Context
arXiv
7
Teds Lin 許願
LongAct：利用內在激活模式實現長上下文強化學習
LongAct Harnessing Intrinsic Activation Patterns for Long-Context Reinforcement Learning
arXiv
63
Teds Lin 許願
上下文並行：可擴展的百萬 Token 推理
Context Parallelism for Scalable Million-Token Inference
arXiv
87
kerg kerg 許願
就地測試時訓練
In-Place Test-Time Training
arXiv
2
Teds Lin 許願