---
dataset_info:
  features:
  - name: text
    dtype: string
  - name: label
    dtype: int64
  - name: intensity
    dtype: int64
  splits:
  - name: original
    num_bytes: 4586
    num_examples: 30
  - name: augmented
    num_bytes: 187640
    num_examples: 1200
  download_size: 32454
  dataset_size: 192226
configs:
- config_name: default
  data_files:
  - split: original
    path: data/original-*
  - split: augmented
    path: data/augmented-*
---

# Yash Gym Machines Text Dataset

## Dataset Summary
This dataset contains descriptive text about **gym machines**, with both binary and numerical targets.  
It includes:
- `original`: 100+ manually authored samples (~200 characters each)
- `augmented`: 1000+ synthetic samples created using NLTK-based word-swapping augmentation.

---

## Intended Use
Educational dataset for text classification and augmentation tasks.  
Not suitable for medical, fitness, or training prescriptions.

---

## Features
- `text` (string): ~200-character description of a gym machine
- `label` (binary): 0 = Lower Body, 1 = Upper Body
- `intensity` (int): Muscle intensity score from 1 (light) to 5 (heavy)

---

## Preprocessing & Augmentation
- Original samples: Self-authored text descriptions
- Augmentation: Random word-swapping using NLTK tokenization
- Splits:
  - `original` → 100+ samples
  - `augmented` → 1200+ samples

---

## Exploratory Data Analysis (EDA)
- Checking dataset size
- Checking label distribution
- Checking intensity distribution

---

## Ethical Notes
- Dataset contains no personal or sensitive data.
- Labels (upper vs lower body, intensity) are subjective.
- Augmentation may produce awkward grammar but preserves label meaning.
- For educational purposes only.

---

## AI Usage Disclosure
- Generative AI was **not used** to create the original dataset.
- ChatGPT (OpenAI GPT-5) was used to draft this README.
"""