metadata
dataset_info:
features:
- name: image
dtype: image
- name: embedding
sequence:
sequence: float16
- name: tag
dtype: string
- name: depth_map
dtype: image
- name: focallength
dtype: float64
- name: masks
sequence:
sequence:
sequence: uint8
- name: bboxes_or_points
sequence:
sequence: float64
- name: captions
sequence: string
- name: pointclouds
sequence: string
- name: is_canonicalized
dtype: bool
- name: prompts
sequence: 'null'
- name: truncated_prompts
sequence: 'null'
- name: messages
sequence: 'null'
splits:
- name: train
num_bytes: 8131318
num_examples: 8
download_size: 5505577
dataset_size: 8131318
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
tags:
- vqasynth
- remyx