metadata
configs:
- config_name: default
data_files:
- split: train
path: pokemon_gpt4o_*.parquet
dataset_info:
features:
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
- name: images
list:
dtype: image
license: apache-2.0
task_categories:
- text-generation
- question-answering
- visual-question-answering
language:
- en
- zh
tags:
- llama-factory
- multimodal
size_categories:
- 1K<n<10K
Borrowed from: https://huggingface.co/datasets/jugg1024/pokemon-gpt4o-captions
You can use it in LLaMA Factory by specifying dataset: pokemon_cap.