Skip to content
View oshaulz's full-sized avatar
πŸš€
πŸš€
  • Canadian Severe Storms Laboratory
  • London, Ontario
  • 03:25 (UTC -04:00)
  • in/oshaulz

Organizations

@Canadian-Severe-Storms-Laboratory @frc10554

Block or report oshaulz

Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse

Pinned Loading

  1. glen glen Public

    Embedded document database for Nim with built-in spatial, time-series, and numeric primitives. Designed for low-latency in-process workloads from edge devices to single-node analytics

    Nim 2

  2. XORFlow. Some datasets have mutually... XORFlow. Some datasets have mutually-exclusive feature groups, if only one group is ever "active" per sample a MSE model will slightly activate both groups. That looks fine on average but is wrong for every individual sample. XORFlow fixes this by forcing exclusivity after your model runs.
    1
    #!/usr/bin/env python3
    2
    """xorflow
    3
    4
    Some datasets have two (or more) groups of features that don't turn on
    5
    together. A normal model trained with plain MSE will hedge and light up