3 pointsby jchandra3 hours ago2 comments
  • vivahir2152 hours ago
    Interesting Approach. Curious about the latency tradeoff: OLS + SVD are much heavier than Top-K.Have you benchmarked end-to-end inference latency?
  • jchandra2 hours ago
    [dead]