datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
ulvr_subset
ULVR stage-0 subsets (latent + source)
Curated, nested subsets of the Unified Visual Latent Reasoning (ULVR) stage-0
training data. Each subset folder is self-contained and ships both:
latent/ — pre-computed teacher latents, identical schema to
RuoliuYang/step0-all
source/ — the matching source samples (images + question/answer +
messages), identical schema to
RuoliuYang/ULVR_v2_clean
Latents and source rows are joinable by sample_id (within a category).
Folder… See the full description on the dataset page: https://huggingface.co/datasets/RuoliuYang/ulvr_subset.ULVR_v2_clean
ULVR_v2_clean
Universal Latent Visual Reasoning training data, cleaned. 8 categories (subsets); each has train + validation splits.
Every sample: input image + question -> assistant produces <abs_vis_token> + intermediate visual step(s) + \boxed{answer}.
subset
train
validation
text_cot
333,911
3,533
bbox_highlight
229,237
2,558
bbox_crop
229,237
2,558
depth
40,000
25
edge
40,000
14
segmentation
40,000
326
helper_interleaved
340,210
3,544
scene_graph
40… See the full description on the dataset page: https://huggingface.co/datasets/RuoliuYang/ULVR_v2_clean.ULVR-filtered
ULVR-filtered
Filtered subset of RuoliuYang/ULVR_v2_clean: the 101,951 training samples that Qwen2.5-VL-7B-Instruct answered incorrectly given only input_image, but correctly once the intermediate_image_* were also provided (judged by Qwen3-VL-32B-Instruct). Same schema / subsets / train-split structure as the source.
subset
rows
scene_graph
3522
edge
1394
depth
537
segmentation
1328
bbox_highlight
15186
bbox_crop
15260
text_cot
27158
helper_interleaved… See the full description on the dataset page: https://huggingface.co/datasets/williamium/ULVR-filtered.ULVR_all
Full Monet training data (compressed)
Loose files under images/ on this repo are incomplete (Hub directory file limit). Download and extract:
Archive
Contents
archives/monet_train.tar.zst
no_text/train.jsonl
archives/images_input.tar.zst
images/input/
archives/images_intermediate.tar.zst
images/intermediate/
See archives/EXTRACT.md. Example:
huggingface-cli download RuoliuYang/ULVR_all --repo-type dataset --include "archives/*" --local-dir ./ULVR_all
cd… See the full description on the dataset page: https://huggingface.co/datasets/RuoliuYang/ULVR_all.ULVR-mirage-size40kulvr-v1-60k-lvr-sftYxtravolXraevon7AevrynthAstrymorQuivryonUlvrynthVynthrae