Watch LIVE (recorded) as we create a TensorFlow container to handle an audio processing task - writing the classification script, the Dockerfile, handling the VM quota increase, all to get ready for the next step where we get to deploy it to classify some bird audio data!
Check out the app at [ Ссылка ]!
In this video as part of our Sound Event Detection training process, we're creating a python script with a pre-trained Keras CNN model and a spectrogram function to reclassify all of the training data, in an effort to filter it before the next step in the training.
This is the UNABRIDGED version where mistakes are made, code is written on-the-fly, and sometimes we pause (but keep recording) to look things up. It's sometimes nice to watch at 1.5-2x speed, or check out one of the other (abridged) videos.
Ещё видео!