qingwuuu/ngld-grape-leaf-vlm-w-img-without-diff-ref-v4

---
dataset_info:
features:
- name: sample_id
dtype: string
- name: image_id
dtype: string
- name: split
dtype: string
- name: task
dtype: string
- name: prompt
dtype: string
- name: answer
dtype: string
- name: image
dtype: image
- name: teacher_full_json
dtype: string
- name: teacher_normalized_json
dtype: string
- name: raw_disease_label
dtype: string
- name: normalized_disease_label
dtype: string
- name: data_source
dtype: string
- name: orig_image_name
dtype: string
- name: orig_json_name
dtype: string
- name: teacher_path
dtype: string
- name: answer_v4
dtype: string
- name: prompt_v4
dtype: string
splits:
- name: train
num_bytes: 2120859044
num_examples: 5838
- name: val
num_bytes: 280700444
num_examples: 730
- name: test
num_bytes: 265130415
num_examples: 730
download_size: 2626888513
dataset_size: 2666689903
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: val
path: data/val-*
- split: test
path: data/test-*
---
## Dataset Source and Credits

This dataset is derived from the **Niphad Grape Leaf Disease Dataset (NGLD)** published on Mendeley Data.

**Original dataset:**
- Title: Niphad Grape Leaf Disease Dataset (NGLD)
- Authors: Madhuri Dharrao, Deepak Dharrao, Rakesh Sonawane
- Institution: Symbiosis Institute of Technology, Symbiosis International University
- DOI: https://doi.org/10.17632/8nnd2ypcv3.1
- License: CC BY 4.0

The original dataset contains high-quality images of table grape leaves categorized into multiple disease conditions.
In this work, the images were reorganized, annotated, and extended with instruction-style supervision (captioning, diagnosis, and reasoning) to support multimodal vision-language model training.

All credit for the original images belongs to the original authors.