phytodoc
- 0 Collaborators
The project aims to use the Kaggle New Plant Disease dataset to develop a model that can accurately predict plant diseases using ResNet neural network architecture. The trained model will be converted to ONNX format and deployed on Azure Serverless Functions for efficient and cost-effective executio ...learn more
Project status: Published/In Market
Artificial Intelligence, Cloud
Groups
Student Developers for oneAPI
Intel Technologies
Intel CPU
Overview / Usage
The project aims to use the Kaggle New Plant Disease dataset to develop a model that can accurately predict plant diseases using ResNet neural network architecture. The trained model will be converted to ONNX format and deployed on Azure Serverless Functions for efficient and cost-effective execution. This will enable farmers and gardeners to quickly and accurately identify plant diseases, allowing for timely intervention and mitigation of the spread of disease, ultimately leading to increased crop yields and healthier plants.
Methodology / Approach
- Data Collection and Preparation: The first step is to obtain the Kaggle New Plant Disease dataset, which consists of images of diseased and healthy plant leaves belonging to 38 different categories. The dataset needs to be cleaned and preprocessed, which includes resizing the images, converting them to arrays, and splitting the data into training and testing sets.
- Model Training: The ResNet neural network architecture is used to train the model using the prepared dataset. The ResNet model is a deep convolutional neural network that has been proven to perform well on image classification tasks.
- Model Optimization: After training the model, it needs to be optimized to achieve better performance. This includes techniques such as pruning, regularization, and fine-tuning the model's hyperparameters.
- Conversion to ONNX Format: Once the model is trained and optimized, it needs to be converted to the ONNX format, which is a common format for machine learning models that can be used across different platforms.
- Deployment on Azure Serverless Functions: The final step is to deploy the model on Azure Serverless Functions. This involves creating an Azure Function App, uploading the ONNX model, and creating a Python script that can be used to make predictions using the deployed model.
Overall, the methodology involves collecting and preparing the data, training and optimizing the model, converting it to the ONNX format, and deploying it on Azure Serverless Functions for efficient and cost-effective execution.
Technologies Used
Pytorch
Python