Custom Training with YOLOv7 🔥

Some Important links

• Model Inference🤖
• 🚀Training Yolov7 on Kaggle
• Weight and Biases 🐝
• HuggingFace 🤗 Model Repo

Contact Information

• Name - Owais Ahmad
• Phone - +91-9515884381
• Email - owaiskhan9654@gmail.com
• Portfolio - https://owaiskhan9654.github.io/

Objective

To Showcase custom Object Detection on the Given Dataset to train and Infer the Model using newly launched YoloV7.

Data Acquisition

The goal of this task is to train a model that can localize and classify each instance of Person and Car as accurately as possible.

• Link to the Downloadable Dataset

from IPython.display import Markdown, display

display(Markdown("../input/Car-Person-v2-Roboflow/README.roboflow.txt"))

Custom Training with YOLOv7 🔥

In this Notebook, I have processed the images with RoboFlow because in COCO formatted dataset was having different dimensions of image and Also data set was not splitted into different Format. To train a custom YOLOv7 model we need to recognize the objects in the dataset. To do so I have taken the following steps:

• Export the dataset to YOLOv7
• Train YOLOv7 to recognize the objects in our dataset
• Evaluate our YOLOv7 model's performance
• Run test inference to view performance of YOLOv7 model at work

📦 YOLOv7

https://raw.githubusercontent.com/Owaiskhan9654/Yolo-V7-Custom-Dataset-Train-on-Kaggle/main/car-person-2.PNG" width=800>

Image Credit - jinfagang

Step 1: Install Requirements

!git clone https://github.com/WongKinYiu/yolov7 # Downloading YOLOv7 repository and installing requirements
%cd yolov7
!pip install -qr requirements.txt
!pip install -q roboflow

Downloading YOLOV7 starting checkpoint

!wget "https://github.com/WongKinYiu/yolov7/releases/download/v0.1/yolov7.pt"

import os
import glob
import wandb
import torch
from roboflow import Roboflow
from kaggle_secrets import UserSecretsClient
from IPython.display import Image, clear_output, display # to display images



print(f"Setup complete. Using torch {torch._version_} ({torch.cuda.get_device_properties(0).name if torch.cuda.is_available() else 'CPU'})")

https://camo.githubusercontent.com/dd842f7b0be57140e68b2ab9cb007992acd131c48284eaf6b1aca758bfea358b/68747470733a2f2f692e696d6775722e636f6d2f52557469567a482e706e67">

I will be integrating W&B for visualizations and logging artifacts and comparisons of different models!

YOLOv7-Car-Person-Custom

try:
  user_secrets = UserSecretsClient()
  wandb_api_key = user_secrets.get_secret("wandb_api")
  wandb.login(key=wandb_api_key)
  anonymous = None
except:
  wandb.login(anonymous='must')
  print('To use your W&B account,
Go to Add-ons -> Secrets and provide your W&B access token. Use the Label name as WANDB. 
Get your W&B access token from here: https://wandb.ai/authorize')
  
  
  
wandb.init(project="YOLOvR",name=f"7. YOLOv7-Car-Person-Custom-Run-7")

Step 2: Assemble Our Dataset

https://uploads-ssl.webflow.com/5f6bc60e665f54545a1e52a5/615627e5824c9c6195abfda9_computer-vision-cycle.png" alt="">

In order to train our custom model, we need to assemble a dataset of representative images with bounding box annotations around the objects that we want to detect. And we need our dataset to be in YOLOv7 format.

In Roboflow, We can choose between two paths:

• Convert an existing Coco dataset to YOLOv7 format. In Roboflow it supports over 30 formats object detection formats for conversion.
• Uploading only these raw images and annotate them in Roboflow with Roboflow Annotate.

Version v2 Aug 12, 2022 Looks like this.

https://raw.githubusercontent.com/Owaiskhan9654/Yolo-V7-Custom-Dataset-Train-on-Kaggle/main/Roboflow.PNG" alt="">

user_secrets = UserSecretsClient()
roboflow_api_key = user_secrets.get_secret("roboflow_api")

rf = Roboflow(api_key=roboflow_api_key)
project = rf.workspace("owais-ahmad").project("custom-yolov7-on-kaggle-on-custom-dataset-rakiq")
dataset = project.version(2).download("yolov7")

Step 3: Training Custom pretrained YOLOv7 model

Here, I am able to pass a number of arguments: - img: define input image size - batch: determine

KITTI Dataset For Training YOLOv7

## Overview

KITTI Dataset For Training YOLOv7 is a dataset for object detection tasks - it contains Vehicles Pedestrians annotations for 7,481 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

Train Test Split For Freiburg Dataset In YOLOv7 Format

## Overview

Train Test Split For Freiburg Dataset In YOLOv7 Format is a dataset for object detection tasks - it contains Groceries annotations for 8,879 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

Yolo V7 Train

## Overview

Yolo V7 Train is a dataset for object detection tasks - it contains Lego annotations for 1,823 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

The MegaWeeds dataset has been used for pre-training YOLO from scratch. The weights obtained with the code below is MW_weights.pt !python train.py --device 0 --batch-size 8 --epochs 300 --img 864 864 --multiscale –data data/custom_data.yaml --hyp data/hyp.scratch.custom.yaml --cfg cfg/training/yolov7-custom.yaml --weights “ ” --name save_name

This is part of the dataset concerning the YOLO training of satellite images (Barchan dunes). In this dataset, you find the "YOLOv8 train" folder that contains the structure and images obtained from HiRISE, CTX global mosaic, Google Earth Pro, and Copernicus. We saved the images with the HiView, Google Earth Pro, and Copernicus software to train a CNN with images of barchan dunes, the "Train Results" folder that contains the figures and weights of YOLO detection of barchan dunes, the "Earth detections" that contains some barchan dune detections on different locations of Earth, the "Mars detections" that contains some barchan dune detections on different locations of Mars, and the "Code Files" that contains the scripts to detect barchan dunes, train a YOLOv8, convert masks to polygons and plot resulting YOLO parameters.

Yolov7 Seg Test

## Overview

Yolov7 Seg Test is a dataset for instance segmentation tasks - it contains Person 07zm annotations for 338 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

This repository contains datasets to train, validate and test deep learning models to detect microfossil fish teeth and denticles called "ichthyoliths". All the dataset contains images of glass slides prepared from deep-sea sediment obtained from Pacific Ocean, and annotation label files formatted to YOLO. 01_original_all The dataset contains 12219 images and 6945 label files. 6945 images include at least one ichthyolith and 5274 images include no ichthyolith. These images and label files were randomly split into three subset, "train" that contains 9740 images and 5551 label files, "val" that contains 1235 images and 695 label files and "test" that contains 1244 images and 699 label files. All the images were selected manually. 02_original_selected This dataset is generated from 01_original_all by removing images without ichthyoliths. The dataset contains 6945 images that include at least one ichthyolith and 6945 label files. The dataset contains three subset, "train" that contains 5551 images and label files, "val" that contains 695 images and label files and "test" that contains 699 images and label files. 03_extended_all This dataset is generated from 01_original_all by adding 4463 images detected by deep learning models. The dataset contains 16682 images and 9473 label files. 9473 images include at least one ichthyolith and 7209 images include no ichthyolith. These images and label files were split into three subset, "train" that contains 13332 images and 7594 label files, "val" that contains 1690 images and 947 label files and "test" that contains 1660 images and 932 label files. Label files were checked manually. 04_extended_selected This dataset is generated from 03_extended_all by removing images without ichthyoliths. The dataset contains 9473 images that include at least one ichthyolith and 9473 label files. The dataset contains three subset, "train" that contains 7594 images and label files, "val" that contains 947 images and label files and "test" that contains 932 images and label files.

Custom Multi-Altitude Aerial Vehicles Dataset:

Created for publishing results for ICUAS 2023 paper "How High can you Detect? Improved accuracy and efficiency at varying altitudes for Aerial Vehicle Detection", following the abstract of the paper.

Abstract—Object detection in aerial images is a challenging task mainly because of two factors, the objects of interest being really small, e.g. people or vehicles, making them indistinguishable from the background; and the features of objects being quite different at various altitudes. Especially, when utilizing Unmanned Aerial Vehicles (UAVs) to capture footage, the need for increased altitude to capture a larger field of view is quite high. In this paper, we investigate how to find the best solution for detecting vehicles in various altitudes, while utilizing a single CNN model. The conditions for choosing the best solution are the following; higher accuracy for most of the altitudes and real-time processing ( > 20 Frames per second (FPS) ) on an Nvidia Jetson Xavier NX embedded device. We collected footage of moving vehicles from altitudes of 50-500 meters with a 50-meter interval, including a roundabout and rooftop objects as noise for high altitude challenges. Then, a YoloV7 model was trained on each dataset of each altitude along with a dataset including all the images from all the altitudes. Finally, by conducting several training and evaluation experiments and image resizes we have chosen the best method of training objects on multiple altitudes to be the mixup dataset with all the altitudes, trained on a higher image size resolution, and then performing the detection using a smaller image resize to reduce the inference performance. The main results

The creation of a custom dataset was necessary for altitude evaluation as no other datasets were available. To fulfill the requirements, the footage was captured using a small UAV hovering above a roundabout near the University of Cyprus campus, where several structures and buildings with solar panels and water tanks were visible at varying altitudes. The data were captured during a sunny day, ensuring bright and shadowless images. Images were extracted from the footage, and all data were annotated with a single class labeled as 'Car'. The dataset covered altitudes ranging from 50 to 500 meters with a 50-meter step, and all images were kept at their original high resolution of 3840x2160, presenting challenges for object detection. The data were split into 3 sets for training, validation, and testing, with the number of vehicles increasing as altitude increased, which was expected due to the larger field of view of the camera. Each folder consists of an aerial vehicle dataset captured at the corresponding altitude. For each altitude, the dataset annotations are generated in YOLO, COCO, and VOC formats. The dataset consists of the following images and detection objects:

    Data
    Subset
    Images
    Cars


    50m
    Train
    130
    269


    50m
    Test
    32
    66


    50m
    Valid
    33
    73


    100m
    Train
    246
    937


    100m
    Test
    61
    226


    100m
    Valid
    62
    250


    150m
    Train
    244
    1691


    150m
    Test
    61
    453


    150m
    Valid
    61
    426


    200m
    Train
    246
    1753


    200m
    Test
    61
    445


    200m
    Valid
    62
    424


    250m
    Train
    245
    3326


    250m
    Test
    61
    821


    250m
    Valid
    61
    823


    300m
    Train
    246
    6250


    300m
    Test
    61
    1553


    300m
    Valid
    62
    1585


    350m
    Train
    246
    10741


    350m
    Test
    61
    2591


    350m
    Valid
    62
    2687


    400m
    Train
    245
    20072


    400m
    Test
    61
    4974


    400m
    Valid
    61
    4924


    450m
    Train
    246
    31794


    450m
    Test
    61
    7887


    450m
    Valid
    61
    7880


    500m
    Train
    270
    49782


    500m
    Test
    67
    12426


    500m
    Valid
    68
    12541


    mix_alt
    Train
    2364
    126615


    mix_alt
    Test
    587
    31442


    mix_alt
    Valid
    593
    31613

It is advised to further enhance the dataset so that random augmentations are probabilistically applied to each image prior to adding it to the batch for training. Specifically, there are a number of possible transformations such as geometric (rotations, translations, horizontal axis mirroring, cropping, and zooming), as well as image manipulations (illumination changes, color shifting, blurring, sharpening, and shadowing).

Unmanned aerial vehicles (UAVs) have become increasingly popular in recent years for both commercial and recreational purposes. Regrettably, the security of people and infrastructure is also clearly threatened by this increased demand. To address the current security challenge, much research has been carried out and several innovations have been made. Many faults still exist, however, including type or range detection failures and the mistaken identification of other airborne objects (for example, birds). A standard dataset that contains photos of drones and birds and on which the model might be trained for greater accuracy is needed to conduct experiments in this field. The supplied dataset is crucial since it will help train the model, giving it the ability to learn more accurately and make better decisions. The dataset that is being presented is comprised of a diverse range of images of birds and drones in motion. Pexel website's images and videos have been used to construct the dataset. Images were obtained from the frames of the recordings that were acquired, after which they were segmented and augmented with a range of circumstances. This would improve the machine-learning model's detection accuracy while increasing dataset training. The dataset has been formatted according to the YOLOv7 PyTorch specification. The test, train, and valid folders are contained within the given dataset. These folders each feature a plaintext file that corresponds to an associated image. Relevant metadata regarding the discovered object is described in the plaintext file. Images and labels are the two subfolders that constitute the folders. The collection consists of 20,925 images of birds and drones. The images have a 640 x 640 pixel resolution and are stored in JPEG format.

Cane Train Licence Plate

## Overview

Cane Train Licence Plate is a dataset for object detection tasks - it contains Plate annotations for 1,931 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

This repository contains the code and resources for the project titled "Detection of Areas with Human Vulnerability Using Public Satellite Images and Deep Learning". The goal of this project is to identify regions where individuals are living under precarious conditions and facing neglected basic needs, a situation often seen in Brazil. This concept is referred to as "human vulnerability" and is exemplified by families living in inadequate shelters or on the streets in both urban and rural areas.

Focusing on the Federal District of Brazil as the research area, this project aims to develop two novel public datasets consisting of satellite images. The datasets contain imagery captured at 50m and 100m scales, covering regions of human vulnerability, traditional areas, and improperly disposed waste sites.

The project also leverages these datasets for training deep learning models, including YOLOv7 and other state-of-the-art models, to perform image segmentation. A comparative analysis is conducted between the models using two training strategies: training from scratch with random weight initialization and fine-tuning using pre-trained weights through transfer learning.

• Two new satellite image datasets focusing on human vulnerability and improperly disposed waste sites, available in public domains.
• Comparison of image segmentation models, including YOLOv7 and Segmentation Models, with performance metrics.
• Best F1-scores: 0.55 for YOLOv7 and 0.64 for Segmentation Models.

This repository provides the code, models, and data pipelines used for training, evaluation, and performance comparison of these deep learning models.

@TECHREPORT {TechReport-Julia-Laura-HumanVulnerability-2024,
  author   = "Julia Passos Pontes, Laura Maciel Neves Franco, Flavio De Barros Vidal",
  title    = "Detecção de Áreas com Atividades de Vulnerabilidade Humana utilizando Imagens Públicas de Satélites e Aprendizagem Profunda",
  institution = "University of Brasilia",
  year    = "2024",
  type    = "Undergraduate Thesis",
  address   = "Computer Science Department - University of Brasilia - Asa Norte - Brasilia - DF, Brazil",
  month    = "aug",
  note    = "People living in precarious conditions and with their basic needs neglected is an unfortunate reality in Brazil. This scenario will be approached in this work according to the concept of \"human vulnerability\" and can be exemplified through families who live in inadequate shelters, without basic structures and on the streets of urban or rural centers. Therefore, assuming the Federal District as the research scope, this project proposes to develop two new databases to be made available publicly, considering the map scales of 50m and 100m, and composed by satellite images of human vulnerability areas,
regions treated as traditional and waste disposed inadequately. Furthermore, using these image bases, trainings were done with the YOLOv7 model and other deep learning models for image segmentation. By adopting an exploratory approach, this work compares the results of different image segmentation models and training strategies, using random weight initialization
(from scratch) and pre-trained weights (transfer learning). Thus, the present work was able to reach maximum F1
score values of 0.55 for YOLOv7 and 0.64 for other segmentation models."
}

This project is licensed under the MIT License - see the LICENSE file for details.

The datset was created for fly and mosquito detection using Yolov8. This file contains 1764 image file and each image file contains text file. This file caries labeling each image

The datasets contain RGB photos of Scots pine seedlings of three populations from two different ecotypes originating in the Czech Republic:Plasy - lowland ecotype,Trebon - lowland ecotype,Decin - upland ecotype.These photos were taken in three different periods (September 10th 2021, October 23rd 2021, January 22nd 2022).File dataset_for_YOLOv7_training.zip contains image data with annotations for training YOLOv7 segmentation model (training and validation sets)The dataset also contains a table with information on individual Scots pine seedlings:affiliation to parent tree (mum)affiliation to population (site)row and column in which the seedling was grown (row, col)affiliation to the planter in which the seedling was grown (box)mean RGB values of pine seedling in three different periods (B_september, G_september, R_september B_october, G_october, R_october, B_january, G_january, R_january)mean HSV values of pine seedling in three different periods (H_september, S_september, V_september, H_october, S_october, V_october, H_january, S_january, V_january)

Val Dataset Creation For COCO + Landing Pad Image Dataset

## Overview

Val Dataset Creation For COCO + Landing Pad Image Dataset is a dataset for object detection tasks - it contains COCO And LandingPads annotations for 1,852 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

The Safety Helmet and Reflective Jacket dataset contains 10,500 images that have been annotated with bounding boxes for two vital object classes: safety_helmet and reflective_jacket. The main objective behind this dataset is to facilitate the training of an object detection model using the YOLOv7 architecture to accurately identify and locate safety equipment within a diverse array of settings and environments. To ensure effective model development and evaluation, the dataset has been divided into train, test, and val subsets, maintaining a balanced ratio of 70% for training and 15% each for testing and validation, resulting in a comprehensive 100% split.

BD Step2 V1_3 Train

## Overview

BD Step2 V1_3 Train is a dataset for instance segmentation tasks - it contains Coal Rock annotations for 755 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

This dataset containe over 10,000 images for (Person, Car, Bus, Motorcycle) I the dataset have been collected from Open-Image dataset. I used this dataset to train YOLOv7 and able to get mAP@50 76, Precision 71, and F1-Score 74, with 80 Epoches with local GPU Nivida RTX 3080 Laptop GPU.

Effectiveness of the basic YOLOv7 model before data curation. The basic YOLOv7 model was fine-tuned and evaluated using the P1.5 dataset (before it underwent thorough data curation). This model was employed to subjectively inspect FP and FN predictions originating from the training, validation, and test sets. Both precision and recall values were calculated using an IoU threshold of 0.5, along with the confidence score corresponding to the highest F1-score.