model inference

3 papers

• Diffusion on Demand: Selective Caching and Modulation for Efficient Generation
• Preventing Shortcuts in Adapter Training via Providing the Shortcuts
• STree: Speculative Tree Decoding for Hybrid State Space Models