SAM2 Household Fine-tuned Checkpoint

Fine-tuned facebook/sam2.1-hiera-small for household object masks with bbox prompts.

  • Base model: facebook/sam2.1-hiera-small
  • Dataset source: nielsr/coco-panoptic-val2017
  • Training samples: 3000
  • Validation samples: 256
  • Test samples: 256
  • Steps: 1000
  • Batch size: 16
  • Best validation IoU: 0.804804
  • Best validation Dice: 0.891847
  • Best threshold: 0.45
  • W&B: https://wandb.ai/sshibinthomass-tu-dortmund/household-object-finetuning/runs/mbtqh4nt
  • Modal app: sam2-household-modal-train
  • Hub repo: sshibinthomass/sam2-household-train-20260709-100604
Downloads last month
-
Safetensors
Model size
38.5M params
Tensor type
F32
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for sshibinthomass/sam2-household-train-20260709-100604

Finetuned
(17)
this model

Dataset used to train sshibinthomass/sam2-household-train-20260709-100604