diff --git a/dataset_quality/debug_viz/pe_rel_viz_100/sampled_items.json b/dataset_quality/debug_viz/pe_rel_viz_100/sampled_items.json new file mode 100644 index 0000000000000000000000000000000000000000..a7098c674fac2edd46f4ebdf82d7a551f2990cf7 --- /dev/null +++ b/dataset_quality/debug_viz/pe_rel_viz_100/sampled_items.json @@ -0,0 +1,3024 @@ +[ + { + "video": "/weka-train/mm-olmo/video_datasets/PE-Video/extended/482308242.mp4", + "mask_json": "/weka-train/royg/vsg-data/pe-extend/sam3_verification/postprocessed/mask/000258_482308242_rle.json", + "obj_list": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35 + ], + "has_relationships": true, + "conversations": [ + { + "from": "human", + "value": "Output the Video Scene Graph from the video and object trajectories:\n