Skip to content

Latest commit

 

History

History
32 lines (23 loc) · 1.01 KB

EVAL.md

File metadata and controls

32 lines (23 loc) · 1.01 KB

Evaluation

Running Evaluation

To evaluate EntitySAM on the VIPSeg dataset, follow these steps:

cd entitysam
export PYTHONPATH=$PYTHONPATH:$(pwd)
python -u eval/eval_vipseg_entity_seg.py --ckpt_dir checkpoints/vit-l/
python -u eval/eval_vipseg_entity_seg.py --ckpt_dir checkpoints/vit-s/ --model_cfg configs/sam2.1_hiera_s.yaml --mask_decoder_depth 4

Computing Metrics

After running the evaluation, compute the metrics using the following commands:

Video Entity Quality (VEQ) Metric

python -u eval/metric/eval_veq.py -i checkpoints/vit-l/inference

Segmentation and Tracking Quality (STQ) Metric

python -u eval/metric/eval_stq_vspw_clsag.py -i checkpoints/vit-l/inference

Evaluation Results

The evaluation will generate results in the checkpoints/vit-l/inference directory. The metrics include:

  • VEQ: Video Entity Quality score for class-agnostic entity segmentation performance
  • STQ: Segmentation and Tracking Quality score for class-agnostic tracking consistency