i love that we are able to able to watch out for our little friends. there are so many trolls out there. anything we can do to protect them and keep them safe I am all for it. they need our help because we are their worst enemies. great job. thanks for the video!
Great video! What is the benefit of embedding the model inside the device, instead of doing all the inference in the cloud? Is the on-device inference used to select which photos will be uploaded?
Usually there are increased costs with doing processing in the cloud. If you have a mesh network of 50 cameras and they ping the cloud every hour, you could build a big bill by the end of the month. Hardware like the Coral TPU Dev Board and Coral TPU Accelerator allow for TensorFlow to be ran locally with minimal costs other than onboard electricity. We are exploring TPU hardware as well.
You can, but several of the common machine learning libraries (in Python at least) for training models are not as well supported on Windows unfortunately. Pytorch for instance doesn't support multiprocessing on Windows. So it might be a bit more work.
i love that we are able to able to watch out for our little friends. there are so many trolls out there. anything we can do to protect them and keep them safe I am all for it. they need our help because we are their worst enemies. great job. thanks for the video!
Great cause and amazing application of AI to photographs
Glad you think so!
Very interesting! Great to see that we're using such cutting edge technologies to take care of our planet and all of its inhabitants
Great video! What is the benefit of embedding the model inside the device, instead of doing all the inference in the cloud? Is the on-device inference used to select which photos will be uploaded?
It helps pre-process images locally. So when you import images they are already classified.
Usually there are increased costs with doing processing in the cloud. If you have a mesh network of 50 cameras and they ping the cloud every hour, you could build a big bill by the end of the month.
Hardware like the Coral TPU Dev Board and Coral TPU Accelerator allow for TensorFlow to be ran locally with minimal costs other than onboard electricity. We are exploring TPU hardware as well.
Great.😀 Please never choice the easier Way🙏
Learn howOw You will Go 😀
Can we use windows computer? Cause its say use linux computer. Thank you
You can, but several of the common machine learning libraries (in Python at least) for training models are not as well supported on Windows unfortunately. Pytorch for instance doesn't support multiprocessing on Windows. So it might be a bit more work.
@@LarsWestergren I see. It's same when I use matlab GNU octave and that's not available for some library in windows. Thank you🙏
Awesome!
Thanks!
Mmmm... interesting but those animals are not in your own backyard, rather you are trespassing into their ancestral habitat.
Molto utile