This blog post was originally published at Deeplite’s website. It is reprinted here with the permission of Deeplite.
Are you looking to enable machine vision while saving on digitalization and manufacturing costs? Implementing AI in smart manufacturing can be transformational to your business, but it can also be challenging to achieve!
At Deeplite, we are very familiar with these challenges. To enhance the accessibility and applicability of AI real-world scenarios in smart manufacturing, Deeplite partnered with a leading brand in IoT intelligent systems. As a result of this collaboration, we provided an optimized model zoo tailored for well-known machine vision deployments and other use cases in smart manufacturing, such as person detection for equipment safety.
We used standard frameworks such as PyTorch, TensorFlow, Keras, and ONNX. We also optimized the model zoo to enable developers to easily and repeatably create ideal AI-powered solutions for their applications and devices! The zoo includes a set of 30+ pre-trained deep learning models for various classification, detection and segmentation-based applications such as person detection, automated optical inspection, defect detection, mask detection and more. Here are some examples:
- MobileNets (classification)
- ResNets (classification)
- VGG (classification)
- Single Shot Detection models
- Yolo Detection models
- UNet Segmentation models
You can view the results and download the full case study in PDF here.
