README.md

---
pretty_name: Fruit Ripeness Dataset
task_categories:
- image-to-text
dataset_info:
  features:
  - name: id
    dtype: string
  - name: fruit_type
    dtype: string
  - name: image
    dtype: image
  - name: growth_stage
    dtype: string
  - name: recommendation
    dtype: string
  - name: consumer_score
    dtype: int32
  - name: local_path
    dtype: string
  splits:
  - name: Apple
  - name: Banana
  - name: DragonFruit
  - name: Grape
  - name: Guava
  - name: Kiwi
  - name: Lychee
  - name: Mango
  - name: Orange
  - name: Papaya
  - name: Peach
  - name: pear
  - name: Pomegranate
  - name: Pomelo
  - name: Strawberry
  - name: Tomato
configs:
- config_name: default
  data_files:
  - split: Apple   
    path:  label/Apple_dataset.parquet
  - split: Banana   
    path: label/Banana_dataset.parquet
  - split: DragonFruit   
    path: label/DragonFruit_dataset.parquet
  - split: Grape   
    path: label/Grape_dataset.parquet
  - split: Guava   
    path: label/Guava_dataset.parquet
  - split: Kiwi   
    path: label/Kiwi_dataset.parquet
  - split: Lychee   
    path: label/Lychee_dataset.parquet
  - split: Mango   
    path: label/Mango_dataset.parquet
  - split: Orange   
    path: label/Orange_dataset.parquet
  - split: Papaya   
    path: label/Papaya_dataset.parquet
  - split: Peach   
    path: label/Peach_dataset.parquet
  - split: pear   
    path: label/pear_dataset.parquet
  - split: Pomegranate   
    path: label/Pomegranate_dataset.parquet
  - split: Pomelo   
    path: label/Pomelo_dataset.parquet
  - split: Strawberry   
    path: label/Strawberry_dataset.parquet
  - split: Tomato   
    path: label/Tomato_dataset.parquet
  
---

# 🥭 FruitBench: A Multimodal Benchmark for Fruit Growth Understanding

**Paper**: *FruitBench: A Multimodal Benchmark for Comprehensive Fruit Growth Understanding in Real-World Agriculture*  
**Conference**: NeurIPS 2025 (submitted)  
**Authors**: Jihao Li*, Jincheng Hu*, Pengyu Fu*, Ming Liu, et al.

---
## 📌 Dataset Summary

**FruitBench** is the first large-scale multimodal benchmark designed to evaluate vision-language models on real-world agricultural understanding. It focuses on **fruit growth modeling**, supporting:

- 🍎 Fruit Type Classification  
- 🌱 Growth Stage Recognition (`unripe`, `pest-damaged`, `mature`, `rotten`)  
- 🌾 Agricultural Action Recommendation  (`keep for further growth`, `picking it`, `try to recover it`, `discard it`)
- 🍽️ Consumer Score Prediction (1–100)

The dataset contains **3,200 high-quality expert-annotated images** covering **16 fruit categories**, each across **4 growth stages**.

<p align="center">
<img src="fig2.png" alt="Dataset sample" width="80%"/>
</p>

---
## 🔍 Tasks

<p align="center">
<img src="fig1.png" alt="Task Overview" width="80%"/>
</p>

1. **Type Classification**  
2. **Growth Stage Identification**  
3. **Action Recommendation**  
4. **Consumer Score Prediction**
   
All tasks are evaluated under both **zero-shot** and **one-shot** settings using multimodal large language models (MLLMs).

---
## 📂 Data Structure

The dataset is organized as follow:
```
FruitBench/
├── Data/
│ ├── Apple/
│ │ ├── Mature/
      ├──0001.png
      ├──0002.png
      ├──0003.png
      ├──...
      └──0050.png
│ │ ├── Unripe/
│ │ ├── Rotten/
│ │ └── Pest-damage/
│ ├── Banana/
│ ├── Mango/
│ └── ...
├── label/
│ ├── Apple_dataset.parquet
│ ├── Banana_dataset.parquet
│ ├── Mango_dataset.parquet
│ └── ...
├── json/
│ ├── Apple.json
│ ├── Banana.json
│ └── ...
```

## Evaluation
We evaluate a total of **15 multimodal models** of different types and sizes, covering diverse model architectures, parameter scales, and vision-language capabilities. The evaluated models include:

- CogVLM2-Llama3-Chat  
- DeepSeek-VL-Chat  
- DeepSeek-VL2  
- InternVL2_5  
- Janus-Pro  
- Mantis-siglip-llama3  
- Mantis-Idefics2  
- MiniCPM-Llama3-V2_5  
- MiniCPM-o-2.6  
- mPLUG-OWL3  
- Qwen2.5-VL-Instruct  
- Yi-VL  
*(15 models in total, with various types and sizes)*

## ⚙️ Environment Setup

We provide both `conda` and `pip` setup options (Python 3.11 recommended).

### ✅ Option A: Conda (Recommended)

```bash
conda env create -f environment.yml
conda activate fruitbench
```
### ✅ Option B: pip

```bash
pip install -r requirements.txt
```
---
## 🚀 Usage

### 1. Clone the Repository

```bash
git lfs install
git clone https://huggingface.co/datasets/TJIET/FruitBench

```
### 3. Evaluate Models
As an example, the evaluation command for **CogVLM2-Llama3-Chat** is:
```bash
python scripts/CogVLM2-0-shot.py
```
## 📊 Benchmark Details
- ✅ 3,200 annotated fruit images
- 📦 16 fruit types: strawberry, tomato, guava, dragon fruit, orange, pear, lychee, mango, kiwi, papaya, apple, grape, pomegranate, peach, banana, pomelo
- 🌱 4 growth stages: unripe, pest-damaged, mature, rotten
- 🧑‍🌾 Expert action labels: keep for growth / pick it / recover / discard
- 🎯 Consumer scores: average of 30 human ratings (range: 1–100)

---