All pull requests

Pull requests

• Train Text tokenizers on stored columns streamed from the index (stack 3/3)

  #307·Vedaant-J opened Nov 11, 2026Contributor·

  ·
  4

• Fit table-column encoders on table rows up to an encoder cutoff (stack 2/3)

  #306·Vedaant-J opened Nov 11, 2026Contributor·

  ·
  5

• Encode sampled raw features without combining Arrow chunks

  #305·haoweiyu-kumo opened Nov 11, 2026Collaborator·3 tasks done·

  ·
  8

• Test ADLS model persistence across Python processes

  #304·anle2024 opened Nov 11, 2026Collaborator·3 tasks done·

  ·
  3

• Tighten nightly and weekly result checks

  #303·anle2024 opened Nov 11, 2026Collaborator·

  ·
  3

• Fix native training-table output type mismatches

  #301·haoweiyu-kumo opened Nov 10, 2026Collaborator·

  ·
  4

• Delete the unused streaming prediction-table path

  #300·Vedaant-J opened Nov 10, 2026Contributor·

  ·
  4

• Add kumo.settings for the shared scratch directory (stack 1/3)

  #299·Vedaant-J opened Nov 10, 2026Contributor·

  ·
  8

• Remove unused production encoder-stat fitting

  #298·haoweiyu-kumo opened Nov 10, 2026Collaborator·3 tasks done·

  ·
  2

• Apply a link prediction's candidate-table filter at prediction, hosted and local

  #297·Vedaant-J opened Nov 10, 2026Contributor·

  ·
  1

  3

• Preserve epoch metric history in model bundles

  #295·anle2024 opened Nov 10, 2026Collaborator·

  ·
  3

• Read scoped Unity Catalog Volume Parquet sources

  #293·nivle1661 opened Nov 10, 2026Collaborator·

  ·
  4

  

• Add SageMaker notebook lifecycle and native UI validation

  #292·nivle1661 opened Nov 10, 2026Collaborator·

  ·
  5

  

• Validate configurable tracking timeouts with shared defaults

  #291·nivle1661 opened Nov 10, 2026Collaborator·3 tasks done·

  ·
  9

  

• Expose opt-in batch prediction XAI through the Python API

  #289·danielmawhirter opened Nov 10, 2026Collaborator·2/2·3 tasks done·

  ·
  2

• Generate batch prediction cohorts and sampled graphs

  #288·danielmawhirter opened Nov 10, 2026Collaborator·1/2·3 tasks done·

  ·
  2

• Add opt-in Buffered sampling I/O alongside mmap

  #285·danielmawhirter opened Nov 10, 2026Collaborator·4 tasks done·

  ·
  3

• Gate, ownership and post-merge alerts for kumo-pytorch

  #284·lambyqq opened Nov 10, 2026Contributor·3 tasks done·

  ·
  3

• Precompute scalar features and targets into enriched anchors

  #283·Victsm opened Nov 10, 2026Collaborator·

  ·
  3

• Keep anchor and history loading in native model batches

  #282·Victsm opened Nov 10, 2026Collaborator·

  ·
  3

• Prepare shared native row execution once for anchor batches

  #281·Victsm opened Nov 10, 2026Collaborator·

  ·
  3

• Lower scalar PQL selection into shared row-aggregate IR

  #280·Victsm opened Nov 10, 2026Collaborator·

  ·
  3

• Reuse TT numeric reductions in temporal row queries

  #279·Victsm opened Nov 10, 2026Collaborator·

  ·
  3

• Add Databricks notebook lifecycle and graph UI validation

  #270·nivle1661 opened Nov 9, 2026Collaborator·

  ·
  5

  

• Execute local Engine notebooks on CPU with captured outputs

  #269·9Olive opened Nov 9, 2026Contributor·3 tasks done·

  ·
  5