Hii everyone ,I am interested in studying the FOMO algorithm and I studied FOMO Self-Attention. How can I get mathematical information on the FOMO algorithm? Also how can i calculate center points
I recommend downloading the .tflite model from one of your projects that has a trained FOMO model (or you can clone this example FOMO project). From there, upload the .tflite model to netron.app to view the model architecture, ops, quantization center points, etc.
2 Likes
Hii thank you for your interest @shawn_edgeimpulse please do you mean this library ( Tensai Flow library)
I do not understand the question. Can you provide a link to the library you are talking about? What are you trying to do with that library?
Hi @shawn_edgeimpulse
I am very sorry, but I did not understand the meaning of tflite model. I thought I could find it in libraries. Can you help me find tflite model?
This mathematics is very important to me because I study this algorithm as part of my research and project 
A .tflite (or .lite) file is a compressed library file (in the “FlatBuffer” format) containing the model architecture and weights for a TensorFlow Lite file. You can read about how the TensorFlow Lite conversion process works here: https://www.tensorflow.org/lite/models/convert
In your Edge Impulse project, go to your Dashboard and click on the TensorFlow Lite model file to download it (I recommend starting with the float32 model, as it will be easier to understand).
Go to netron.app. Click “Open Model…” and upload the file that you just downloaded. You should see a model architecture graph that you can navigate and click on the various nodes to learn more about the various operations.
You can then look up each operation in the TensorFlow documentation. For example, you can read about the Conv2D operation here.
1 Like
