Dataset Viewer
The dataset could not be loaded because the splits use different data file formats, which is not supported. Read more about the splits configuration. Click for more details.
Couldn't infer the same data file format for all splits. Got {NamedSplit('train'): ('text', {}), NamedSplit('validation'): ('text', {}), NamedSplit('test'): ('imagefolder', {})}
Error code: FileFormatMismatchBetweenSplitsError
Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
DefectForge β foreign-body conveyor dataset (synthetic)
Synthetic object-detection dataset produced end-to-end by DefectForge (a molab/NVIDIA port of the open-source aionVIS agent swarm, MIT, by 0xAthkot) for one task: A food-packaging plant needs to detect foreign objects - metal fragments, broken rubber seal pieces, plastic shards and loose screws - lying on or next to sealed packaged goods moving on a conveyor belt, before the packages leave the production line.
- Classes: metal_fragment, rubber_seal_fragment, plastic_shard, loose_screw
- Images generated with SDXL (stabilityai/stable-diffusion-xl-base-1.0), labelled by YOLOE open-vocabulary segmentation, verified by the aionVIS Critic (geometric box re-derivation from masks) plus a Qwen3.8 VLM semantic spot-check of every accepted label. Zero human labels.
- Compiled split: 37 train / 7 val images; 94 boxes accepted, 344 rejected/regenerated by the Critic.
raw/holds the generated images as produced (uploaded incrementally during the run);train_dataset/is the compiled YOLO dataset;seeded_test/is a composited proxy test set (see the model card β it is NOT field data).
All images are synthetic. Performance on this data does not by itself establish real-world performance.
- Downloads last month
- -