3 ms·

Thanks! Do you by chance know if it is practical to convert huge .h5 Keras models to TF.js layers model? (~1GB). I have some state-of-art computer vision models

by bitL 8y ago

Thanks! Do you by chance know if it is practical to convert huge .h5 Keras models to TF.js layers model? (~1GB). I have some state-of-art computer vision models and it would be great if they could be used in a browser with a WebCam, if they can fit into memory and be performant for inferencing.

zawerf 8y ago
Mobile has similar constraints as web (but to a lesser extent) so you will find more resources there. Look into quantizing the weights (e.g., reduce from 32bit per parameter to 8 bits) to reduce model size: https://www.tensorflow.org/mobile/optimizing#model_size https://www.tensorflow.org/mobile/optimizing#model_size