Spaces:
Sleeping
Sleeping
Commit History
-
Pin the Space runtime to the environment that reproduces the paper
cisdeliThe Space installed ultralytics/torch unpinned, and on the identical bundled image it reported 671.73 in2 of projected area against 737.40 in2 locally, with 6 tubers shifting out of OV. Detections were identical, so the drift is in mask post-processing, which is version-dependent. Pin the app's runtime to the locked local versions and log them at startup so a future drift is visible in the Space logs.4a09df2 -
Match the demo's segmentation parameters to the batch pipeline
cisdeliThe app passed retina_masks=True while scripts/run_yolo_inference.py uses the Ultralytics defaults. Detections are identical either way, but full-resolution masks move polygon boundaries: on a controlled crop, total projected area shifted 2.6% and 14% of tubers crossed a USDA size threshold, so the demo reported a different class composition than the published per-tuber CSV.71b2c50
-
Bind Gradio to HF port
cisdeli1f2a93b
-
feat: gradio demo
cisdeli044af6e