Image Classification Model Maker Using Google's Framework -- 2
Budget: $250 – $750 CAD
Automated Google Image Classification Application
1. Produce an image classification model builder application that will produce a model with similar or superior performance levels.
2. The application will be operating on a local graphic workstation and must leverage the workstations Nvidia’s RTX GPU’s.
3. We expect the performance results of any model produce by the this application to have equivalent or superior performance result to a model produce by Googles Teachable Machine: https://teachablemachine.withgoogle.com/train
4. After the model has been completed, the model will be saved along with the derived hyperparameters and jpg of Googles GUI performance and text analysis.
5. All of the images provided in this application are 16 bit mono PNG’s
6. Client will provide providing between 10,000 to 100,000 images for each of 9 classes when running the application..
7. After the model has been produced, inference each of the supplied ‘SOURCE_CLASS’ TEST IMAGES
8. Application should built using: https://www.tensorflow.org/lite/models/modify/model_maker/image_classification
OR
https://www.tensorflow.org/tutorials/keras/classification
Developer will be required to:
I. provided source code and executable in bitbucket along with detailed documentation. Source code must be well structured and contain clear detailed remarks throughout describing each action.
II. Sign an NDA supplied by client and IP agreement transferring exclusive usage of the source code and documentation associated with the supplied project/application.
III. Complete application within 1 week from assignment and demonstrate that the performance of the application is equivalent to exceeds the performance/accuracy of a model produced with Googles Teachable Machine Application using the input classification images.
IV. Developer will be required to demonstrate the performance of the application on the customers workstation using Team Viewer and provide update via Telegram each day demonstrating their progress.
Please read complete attachment if you are intersted in working on this project.
1. Produce an image classification model builder application that will produce a model with similar or superior performance levels.
2. The application will be operating on a local graphic workstation and must leverage the workstations Nvidia’s RTX GPU’s.
3. We expect the performance results of any model produce by the this application to have equivalent or superior performance result to a model produce by Googles Teachable Machine: https://teachablemachine.withgoogle.com/train
4. After the model has been completed, the model will be saved along with the derived hyperparameters and jpg of Googles GUI performance and text analysis.
5. All of the images provided in this application are 16 bit mono PNG’s
6. Client will provide providing between 10,000 to 100,000 images for each of 9 classes when running the application..
7. After the model has been produced, inference each of the supplied ‘SOURCE_CLASS’ TEST IMAGES
8. Application should built using: https://www.tensorflow.org/lite/models/modify/model_maker/image_classification
OR
https://www.tensorflow.org/tutorials/keras/classification
Developer will be required to:
I. provided source code and executable in bitbucket along with detailed documentation. Source code must be well structured and contain clear detailed remarks throughout describing each action.
II. Sign an NDA supplied by client and IP agreement transferring exclusive usage of the source code and documentation associated with the supplied project/application.
III. Complete application within 1 week from assignment and demonstrate that the performance of the application is equivalent to exceeds the performance/accuracy of a model produced with Googles Teachable Machine Application using the input classification images.
IV. Developer will be required to demonstrate the performance of the application on the customers workstation using Team Viewer and provide update via Telegram each day demonstrating their progress.
Please read complete attachment if you are intersted in working on this project.