« BackReusing Computation in Text-to-Image Diffusion for Efficient Image Generationarxiv.orgSubmitted by gidellav 4 days ago
  • operator-name 5 hours ago

    This looks similar to https://github.com/chengzeyi/ParaAttention or https://github.com/ali-vilab/TeaCache.

    It’s a shame they don’t compare against or mention them.

    • transformi 9 hours ago

      Interesting approach! Remind me the early insights that neurons in DNN that capture similar concepts.