README.md

metadata

license: apache-2.0
task_categories:
  - visual-question-answering
tags:
  - visual-reasoning
  - multimodal
  - latent-reasoning
configs:
  - config_name: scene_graph
    data_files:
      - split: train
        path: data/scene_graph/*.parquet
  - config_name: edge
    data_files:
      - split: train
        path: data/edge/*.parquet
  - config_name: depth
    data_files:
      - split: train
        path: data/depth/*.parquet
  - config_name: segmentation
    data_files:
      - split: train
        path: data/segmentation/*.parquet
  - config_name: bbox_highlight
    data_files:
      - split: train
        path: data/bbox_highlight/*.parquet
  - config_name: bbox_crop
    data_files:
      - split: train
        path: data/bbox_crop/*.parquet
  - config_name: text_cot
    data_files:
      - split: train
        path: data/text_cot/*.parquet
  - config_name: helper_interleaved
    data_files:
      - split: train
        path: data/helper_interleaved/*.parquet
dataset_info:
  - config_name: scene_graph
    features:
      - name: sample_id
        dtype: string
      - name: category
        dtype: string
      - name: source_dataset
        dtype: string
      - name: question
        dtype: string
      - name: answer
        dtype: string
      - name: input_image
        dtype: image
      - name: intermediate_image_1
        dtype: image
      - name: intermediate_image_2
        dtype: image
      - name: intermediate_image_3
        dtype: image
      - name: num_intermediate_steps
        dtype: int64
      - name: messages_json
        dtype: string
    splits:
      - name: train
        num_examples: 3522
  - config_name: edge
    features:
      - name: sample_id
        dtype: string
      - name: category
        dtype: string
      - name: source_dataset
        dtype: string
      - name: question
        dtype: string
      - name: answer
        dtype: string
      - name: input_image
        dtype: image
      - name: intermediate_image_1
        dtype: image
      - name: intermediate_image_2
        dtype: image
      - name: intermediate_image_3
        dtype: image
      - name: num_intermediate_steps
        dtype: int64
      - name: messages_json
        dtype: string
    splits:
      - name: train
        num_examples: 1394
  - config_name: depth
    features:
      - name: sample_id
        dtype: string
      - name: category
        dtype: string
      - name: source_dataset
        dtype: string
      - name: question
        dtype: string
      - name: answer
        dtype: string
      - name: input_image
        dtype: image
      - name: intermediate_image_1
        dtype: image
      - name: intermediate_image_2
        dtype: image
      - name: intermediate_image_3
        dtype: image
      - name: num_intermediate_steps
        dtype: int64
      - name: messages_json
        dtype: string
    splits:
      - name: train
        num_examples: 537
  - config_name: segmentation
    features:
      - name: sample_id
        dtype: string
      - name: category
        dtype: string
      - name: source_dataset
        dtype: string
      - name: question
        dtype: string
      - name: answer
        dtype: string
      - name: input_image
        dtype: image
      - name: intermediate_image_1
        dtype: image
      - name: intermediate_image_2
        dtype: image
      - name: intermediate_image_3
        dtype: image
      - name: num_intermediate_steps
        dtype: int64
      - name: messages_json
        dtype: string
    splits:
      - name: train
        num_examples: 1328
  - config_name: bbox_highlight
    features:
      - name: sample_id
        dtype: string
      - name: category
        dtype: string
      - name: source_dataset
        dtype: string
      - name: question
        dtype: string
      - name: answer
        dtype: string
      - name: input_image
        dtype: image
      - name: intermediate_image_1
        dtype: image
      - name: intermediate_image_2
        dtype: image
      - name: intermediate_image_3
        dtype: image
      - name: num_intermediate_steps
        dtype: int64
      - name: messages_json
        dtype: string
    splits:
      - name: train
        num_examples: 15186
  - config_name: bbox_crop
    features:
      - name: sample_id
        dtype: string
      - name: category
        dtype: string
      - name: source_dataset
        dtype: string
      - name: question
        dtype: string
      - name: answer
        dtype: string
      - name: input_image
        dtype: image
      - name: intermediate_image_1
        dtype: image
      - name: intermediate_image_2
        dtype: image
      - name: intermediate_image_3
        dtype: image
      - name: num_intermediate_steps
        dtype: int64
      - name: messages_json
        dtype: string
    splits:
      - name: train
        num_examples: 15260
  - config_name: text_cot
    features:
      - name: sample_id
        dtype: string
      - name: category
        dtype: string
      - name: source_dataset
        dtype: string
      - name: question
        dtype: string
      - name: answer
        dtype: string
      - name: input_image
        dtype: image
      - name: intermediate_image_1
        dtype: image
      - name: intermediate_image_2
        dtype: image
      - name: intermediate_image_3
        dtype: image
      - name: num_intermediate_steps
        dtype: int64
      - name: messages_json
        dtype: string
    splits:
      - name: train
        num_examples: 27158
  - config_name: helper_interleaved
    features:
      - name: sample_id
        dtype: string
      - name: category
        dtype: string
      - name: source_dataset
        dtype: string
      - name: question
        dtype: string
      - name: answer
        dtype: string
      - name: input_image
        dtype: image
      - name: intermediate_image_1
        dtype: image
      - name: intermediate_image_2
        dtype: image
      - name: intermediate_image_3
        dtype: image
      - name: num_intermediate_steps
        dtype: int64
      - name: messages_json
        dtype: string
    splits:
      - name: train
        num_examples: 37566

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	37566
total	101951