GPIC
Datasets
All datasets matching “GPIC”gpic
GPIC: A Giant Permissive Image Corpus for Visual Generation
Keshigeyan Chandrasegaran*1,
Kyle Sargent*1,
Suchir Agarwal1,
Michael Jang1,
Michael Poli1,2,
Juan Carlos Niebles1,4,
Justin Johnson3,
Jiajun Wu1,
Li Fei-Fei1
1 Stanford University
2 Radical Numerics
3 University of Michigan
4 Salesforce… See the full description on the dataset page: https://huggingface.co/datasets/stanford-vision-lab/gpic.GPIC-Camera
GPIC-Camera
Per-image camera parameter annotations for the GPIC dataset
(train / test / val; train = 8,000 shards, test = 1,000,000 images, val = 200,000 images), captioned by the Puffin-World model. More captioned datasets are provided in our Puffin-16M website.
The collage above visualizes the camera maps on sample images — each
pair shows the up field (green arrows: the projected gravity-up direction)
and the latitude field (colored contours: angle above/below the horizon).… See the full description on the dataset page: https://huggingface.co/datasets/KangLiao/GPIC-Camera.gpic-bcc-sam3-qwen38-27b
GPIC Bidirectional Concept Correspondence Data
This release was generated by ConCor Training Data Generation. Each training example connects a text mask—a set of caption character spans—to an image mask made from one or more segmented instances. Disjoint co-referring spans can therefore share the same correspondence.
The three training configs intentionally match the caption-row format used by UWGZQ/ConCor-1-Data. Our richer pipeline records and complete per-image dispositions… See the full description on the dataset page: https://huggingface.co/datasets/suryadv/gpic-bcc-sam3-qwen38-27b.CLIP-GPIC-embeddings
CLIP-GPIC-embeddings
Vision Transformer embeddings for split:TEST (1M) of stanford-vision-lab/gpic
Mainly for use with cross-attention read/no-read bridge experiments from my github.
Features: pre-indexed image TAR member byte offset and size -> HTTP Range requests to fetch only selected image bytes.
Meaning: Won't require downloading 1M images for retrieval, will just fetch matches from remote shard.
The MIT license applies to the embedding-bank files, metadata… See the full description on the dataset page: https://huggingface.co/datasets/zer0int/CLIP-GPIC-embeddings.gpic_latents