Multi-Class Image Classification Model for Tomato Leaf Disease Using TensorFlow Take 2

Template Credit: Adapted from a template made available by Dr. Jason Brownlee of Machine Learning Mastery.

SUMMARY: This project aims to construct a predictive model using a TensorFlow convolutional neural network (CNN) and document the end-to-end steps using a template. The Tomato Leaf Disease Dataset is a multi-class classification situation where we attempt to predict one of several (more than two) possible outcomes.

INTRODUCTION: The dataset contains over 20,000 images of tomato leaves with ten diseases and one healthy class. The research team collected these images to investigate the possibilities of developing a lightweight model that can predict tomato leaf disease on a mobile app.

ANALYSIS: The EfficientNetV2M model’s performance achieved an accuracy score of 96.74% after five epochs using the training dataset. When we applied the model to the validation dataset, the model achieved an accuracy score of 95.96%.

CONCLUSION: In this iteration, the TensorFlow EfficientNetV2M CNN model appeared suitable for modeling this dataset.

Dataset ML Model: Multi-Class classification with numerical features

Dataset Used: Tomato Leaves Dataset

Dataset Reference: https://www.kaggle.com/datasets/ashishmotwani/tomato

One source of potential performance benchmarks: https://www.kaggle.com/datasets/ashishmotwani/tomato/code

The HTML formatted report can be found here on GitHub.

Multi-Class Image Classification Model for Tomato Leaf Disease Using TensorFlow Take 1

Template Credit: Adapted from a template made available by Dr. Jason Brownlee of Machine Learning Mastery.

SUMMARY: This project aims to construct a predictive model using a TensorFlow convolutional neural network (CNN) and document the end-to-end steps using a template. The Tomato Leaf Disease Dataset is a multi-class classification situation where we attempt to predict one of several (more than two) possible outcomes.

INTRODUCTION: The dataset contains over 20,000 images of tomato leaves with ten diseases and one healthy class. The research team collected these images to investigate the possibilities of developing a lightweight model that can predict tomato leaf disease on a mobile app.

ANALYSIS: The Xception model’s performance achieved an accuracy score of 97.97% after five epochs using the training dataset. When we applied the model to the validation dataset, the model achieved an accuracy score of 97.74%.

CONCLUSION: In this iteration, the TensorFlow Xception CNN model appeared suitable for modeling this dataset.

Dataset ML Model: Multi-Class classification with numerical features

Dataset Used: Tomato Leaves Dataset

Dataset Reference: https://www.kaggle.com/datasets/ashishmotwani/tomato

One source of potential performance benchmarks: https://www.kaggle.com/datasets/ashishmotwani/tomato/code

The HTML formatted report can be found here on GitHub.

Michael Bungay Stanier on Do More Great Work, Part 6

In the book, Do More Great Work.: Stop the Busywork, and Start the Work that Matters, Michael Bungay Stanier shares his inspiration and techniques to help us do more work that matters.

These are some of my favorite takeaways from reading the book.

Map 5 What’s Calling You

In this chapter, Michael provided two maps that can be used to spot opportunities for Great Work. One map is a more general one of opportunities for Great Work in the whole landscape of our life. The other is more specific to potential opportunities at work.

Next, scan the chosen map and identify areas where there might be opportunities to do more Great Work.

Next, we add more details to the map, or we can customize the map by changing labels or replacing something that works better for us.

After the map is customized to reflect our life, we circle five areas where we think there might be opportunities to do more Great Work. Consider the following:

  • Is there a major project we have wanted to do for some time?
  • What part of the map do we naturally gravitate to?
  • Where could we enhance our work, upgrade our effort, and change it from Good to Great?
  • Where might we begin something new or spark something different?

Also, consider the following:

  • What did we learn by focusing on the big picture? What surprised us?
  • What obvious opportunities for Great Work would we have been almost too blind to see?
  • What might be possible in the less obvious places?
  • What do we know about ourselves that we had not fully realized before?

Finally, Seth Godin offers this secret to doing Great Work.

“My advice for creating Great Work is disarmingly simple: Don’t settle.”

“If you honestly believe that Great Work matters, then the issue is settled. You can and should start today. Identify where you’re settling, and stop.”

三思而後行

(從我一個尊敬的作家,賽斯·高汀

有些問題需要來重新審視。第二個,第三個,甚至第四個想法都是有成效的,因為我們最初的衝動可能無法反映出我們為理解情況的細微差別所做的最大努力。

但是許多問題只是產生了更多的想法,而沒有生產性的產出。當我們面對一些不太可能有一個簡單或富有成效的前進方式的事情時,會很容易陷入一種精神錯亂的想像解決方案。

這真正的藝術是理解我們面臨什麼樣的問題,並且對我們所處的情況投入適量的思考(不是更少,也絕對不是更多)。花費周期對問題進行分類可能比把時間浪費在不值得我們努力的問題上會更有成效。

Binary-Class Image Classification Model for Car vs. Bike Classification Using TensorFlow Take 5

Template Credit: Adapted from a template made available by Dr. Jason Brownlee of Machine Learning Mastery.

SUMMARY: This project aims to construct a predictive model using a TensorFlow convolutional neural network (CNN) and document the end-to-end steps using a template. The Car vs. Bike Classification Dataset is a binary classification situation where we attempt to predict one of two possible outcomes.

INTRODUCTION: The dataset contains 2,000 images of cars and bikes. The research team collected these images to investigate the machine learning model’s ability to understand and distinguish the basic structure of cars and bikes. The research team also made sure that all types of bikes and cars were included for a high degree of variety of cars and bikes.

ANALYSIS: The InceptionV3 model’s performance achieved an accuracy score of 98.56% after five epochs using the training dataset. When we applied the model to the validation dataset, the model achieved an accuracy score of 99.17%.

CONCLUSION: In this iteration, the TensorFlow InceptionV3 CNN model appeared suitable for modeling this dataset.

Dataset ML Model: Binary classification with numerical features

Dataset Used: Car vs. Bike Classification Dataset

Dataset Reference: https://www.kaggle.com/datasets/utkarshsaxenadn/car-vs-bike-classification-dataset

One source of potential performance benchmarks: https://www.kaggle.com/datasets/utkarshsaxenadn/car-vs-bike-classification-dataset/code

The HTML formatted report can be found here on GitHub.

Binary-Class Image Classification Model for Car vs. Bike Classification Using TensorFlow Take 4

Template Credit: Adapted from a template made available by Dr. Jason Brownlee of Machine Learning Mastery.

SUMMARY: This project aims to construct a predictive model using a TensorFlow convolutional neural network (CNN) and document the end-to-end steps using a template. The Car vs. Bike Classification Dataset is a binary classification situation where we attempt to predict one of two possible outcomes.

INTRODUCTION: The dataset contains 2,000 images of cars and bikes. The research team collected these images to investigate the machine learning model’s ability to understand and distinguish the basic structure of cars and bikes. The research team also made sure that all types of bikes and cars were included for a high degree of variety of cars and bikes.

ANALYSIS: The NASNetMobile model’s performance achieved an accuracy score of 98.53% after five epochs using the training dataset. When we applied the model to the validation dataset, the model achieved an accuracy score of 95.33%.

CONCLUSION: In this iteration, the TensorFlow NASNetMobile CNN model appeared suitable for modeling this dataset.

Dataset ML Model: Binary classification with numerical features

Dataset Used: Car vs. Bike Classification Dataset

Dataset Reference: https://www.kaggle.com/datasets/utkarshsaxenadn/car-vs-bike-classification-dataset

One source of potential performance benchmarks: https://www.kaggle.com/datasets/utkarshsaxenadn/car-vs-bike-classification-dataset/code

The HTML formatted report can be found here on GitHub.

Binary-Class Image Classification Model for Car vs. Bike Classification Using TensorFlow Take 3

Template Credit: Adapted from a template made available by Dr. Jason Brownlee of Machine Learning Mastery.

SUMMARY: This project aims to construct a predictive model using a TensorFlow convolutional neural network (CNN) and document the end-to-end steps using a template. The Car vs. Bike Classification Dataset is a binary classification situation where we attempt to predict one of two possible outcomes.

INTRODUCTION: The dataset contains 2,000 images of cars and bikes. The research team collected these images to investigate the machine learning model’s ability to understand and distinguish the basic structure of cars and bikes. The research team also made sure that all types of bikes and cars were included for a high degree of variety of cars and bikes.

ANALYSIS: The ResNet50V2 model’s performance achieved an accuracy score of 98.35% after five epochs using the training dataset. When we applied the model to the validation dataset, the model achieved an accuracy score of 95.33%.

CONCLUSION: In this iteration, the TensorFlow ResNet50V2 CNN model appeared suitable for modeling this dataset.

Dataset ML Model: Binary classification with numerical features

Dataset Used: Car vs. Bike Classification Dataset

Dataset Reference: https://www.kaggle.com/datasets/utkarshsaxenadn/car-vs-bike-classification-dataset

One source of potential performance benchmarks: https://www.kaggle.com/datasets/utkarshsaxenadn/car-vs-bike-classification-dataset/code

The HTML formatted report can be found here on GitHub.

Binary-Class Image Classification Model for Car vs. Bike Classification Using TensorFlow Take 2

Template Credit: Adapted from a template made available by Dr. Jason Brownlee of Machine Learning Mastery.

SUMMARY: This project aims to construct a predictive model using a TensorFlow convolutional neural network (CNN) and document the end-to-end steps using a template. The Car vs. Bike Classification Dataset is a binary classification situation where we attempt to predict one of two possible outcomes.

INTRODUCTION: The dataset contains 2,000 images of cars and bikes. The research team collected these images to investigate the machine learning model’s ability to understand and distinguish the basic structure of cars and bikes. The research team also made sure that all types of bikes and cars were included for a high degree of variety of cars and bikes.

ANALYSIS: The EfficientNetV2M model’s performance achieved an accuracy score of 98.47% after five epochs using the training dataset. When we applied the model to the validation dataset, the model achieved an accuracy score of 95.00%.

CONCLUSION: In this iteration, the TensorFlow EfficientNetV2M CNN model appeared suitable for modeling this dataset.

Dataset ML Model: Binary classification with numerical features

Dataset Used: Car vs. Bike Classification Dataset

Dataset Reference: https://www.kaggle.com/datasets/utkarshsaxenadn/car-vs-bike-classification-dataset

One source of potential performance benchmarks: https://www.kaggle.com/datasets/utkarshsaxenadn/car-vs-bike-classification-dataset/code

The HTML formatted report can be found here on GitHub.