media-type-1.0

v2.78 Export Bundle

Selected endpoint for the v2 line after the clean-core retraining and final merge sweep.

Recommended checkpoint: best_model.safetensors

Model:

• architecture: tf_efficientnet_b0
• classes: 6
• input size: 336
• resize: int(input_size * 1.143) then center crop input_size
• normalize mean: [0.485, 0.456, 0.406]
• normalize std: [0.229, 0.224, 0.225]

Class order:

• illustration
• painting_physical
• real_photo
• rendered_2d
• rendered_3d
• screenshot

Holdout metrics:

• harmonic F1: 0.912398
• accuracy: 0.914089
• min recall: 0.840000
• bottom-2 mean recall: 0.840000

Per-class recall:

• illustration: 0.975610
• painting_physical: 0.960000
• real_photo: 0.960000
• rendered_2d: 0.840000
• rendered_3d: 0.840000
• screenshot: 0.920000

Merge provenance:

• anchor: v2.70_clean_core_v263_finish_weighted8_lr5e5
• donor: v2.29_res244_4plus4
• alpha: 0.10

Known weak areas:

• The remaining weakest classes are the digital seam pair rendered_2d and rendered_3d, but both are now at 0.84 recall on the holdout.
• The most persistent residual OOD confusions are:
  • rendered_2d -> illustration
  • rendered_3d -> real_photo
  • rendered_2d -> rendered_3d
• Receipt-like documents still tend to route to screenshot, which is usually acceptable for OCR-style handling.
• Scanned book pages are still not reliably separated from illustration; use an additional document head or router if page/document handling matters.