How long should training take place? Mine has been going on for at least 30 minutes. I’m analyzing a 1-hr long video. Is this why it’s that long?
Training and video analysis are two separate steps. Just to be clear, training uses only the labels you placed on the images. For timing and other key info, I recommend checking out the Nature Protocols paper: https://rdcu.be/bHpHN
in short, training time can be from 3-12 hours on a GPU (depends on your GPU, your image sizes, and the batch size), or 10 days on a CPU (go for the GPU )… you can use free GPUs on colab, see much more in the code repo.
Thanks, Mackenzie. Also, what does iteration mean? What if I increase or decrease my iterations?
Also, where can I find the GPU in my Python/Anaconda software?