abc-grasp-eval / README.md

Commit History

13 days ago

  1. Replace old benchmark with agent and world-model evaluation galleries

    Verified signature
    9d71edb

  1. Add configurable model selector and model-specific result downloads

    Verified signature
    159070f
  2. Preserve success probability in accuracy and calibration views

    Verified signature
    c2669bd
  3. Show classification accuracy by probability and confidence

    Verified signature
    998ce41
  4. Add success probability calibration plot with bin counts

    Verified signature
    97598a5
  5. Use neutral early-sample timing labels

    Verified signature
    a8acadd
  6. Simplify summary to accuracy by difficulty and recorded outcome

    Verified signature
    fd1df4f
  7. Show accuracy by difficulty and sample type with filters

    Verified signature
    cc089f4
  8. Upload folder using huggingface_hub

    Verified signature
    2f26975
  9. initial commit

    Verified signature
    09ce7ec