---
license: unknown
task_categories:
- object-detection
tags:
- yolo
- medical
- ppe
- cppe-5
size_categories:
- 1K<n<10K
dataset_info:
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: yolo_annotations
    list: string
  - name: classes
    list:
      class_label:
        names:
          '0': Coverall
          '1': Face_Shield
          '2': Gloves
          '3': Goggles
          '4': Mask
  splits:
  - name: train
    num_bytes: 149801040.0
    num_examples: 1000
  - name: validation
    num_bytes: 2437297.0
    num_examples: 29
  download_size: 151255948
  dataset_size: 152238337.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
---

# MPPE Dataset YOLO v1

Formato YOLO derivado de CPPE-5, convertido para fine-tuning de modelos Ultralytics YOLO26.

## Estructura

- images/train/  imagenes de entrenamiento (.jpg)
- images/val/    imagenes de validacion (.jpg)
- labels/train/  anotaciones YOLO (.txt)
- labels/val/    anotaciones YOLO (.txt)
- data.yaml      config para Ultralytics
- classes.txt    nombres de clase

## Clases

| id | nombre |
|---|---|
| 0 | Coverall |
| 1 | Face_Shield |
| 2 | Gloves |
| 3 | Goggles |
| 4 | Mask |

## Formato de anotacion

Cada linea en los .txt sigue el estandar YOLO:
class x_center y_center width height (normalizado 0-1).

## Uso

from ultralytics import YOLO
model = YOLO("yolo26s.pt")
model.train(data="data.yaml", epochs=100, imgsz=640)

## Fuente original

Dataset original CPPE-5 (Coverall, Face_Shield, Gloves, Goggles, Mask), deteccion de equipo de proteccion medica. Convertido de formato COCO-style (HF Datasets) a formato YOLO.