---
dataset_info:
  features:
  - name: messages
    list:
    - name: content
      list:
      - name: index
        dtype: int64
      - name: text
        dtype: string
      - name: type
        dtype: string
    - name: role
      dtype: string
  - name: images
    list: image
  splits:
  - name: train
    num_bytes: 2436536154.535
    num_examples: 25235
  - name: test
    num_bytes: 272317950.4
    num_examples: 2804
  download_size: 2689790918
  dataset_size: 2708854104.935
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
tags:
- remyx
- synthetic
license: apache-2.0
task_categories:
- visual-question-answering
pretty_name: OpenSpaces
size_categories:
- 10K<n<100K
---




# VQASynth_spacellava
![image/png](https://cdn-uploads.huggingface.co/production/uploads/647777304ae93470ffc28913/t_hxwRDUgf-HywL89rJX4.png)

Uses the [VQASynth](https://github.com/remyxai/VQASynth) pipeline to synthesize spatialVQA samples, mixed with general VQA samples used to fine-tune LLaVA-v1.5-13b.