Is this your dataset? Claim it with the Hugging Face account that owns it.

nanovisionx/
RAEv2-data

Name: RAEv2-data
Creator: nanovisionx

text-to-image

unconditional-image-generation

other

diffusion

representation-autoencoder

rae

Views

No views yet

RAEv2 Data

Pre-processed datasets and pretrained encoders for RAEv2: Improved Baselines with Representation Autoencoders. All rights to the original owners; per-subset attribution below.

Repo Structure

RAEv2-data/ |-- imagenet-256/ # ImageNet-1k at 256x256 (Arrow) |-- blip3o-256/ # BLIP3o captioned images (WDS) |-- render-text-256/ # Rendered-text images (WDS) |-- scale-rae-256/ # Synthetic FLUX images (WDS) |-- recon-256/ # Robot… See the full description on the dataset page: https://huggingface.co/datasets/nanovisionx/RAEv2-data.