Liveanimalcams

v7 is created by AlexeyAB and WonKinYiu the authors of YOLOv4. Who took over YOLOv3 from the original author Joseph Redmon, after Redmon quit the CV industry due to ethical concerns.

It mentions prizes, but I don’t see them listed. What are they?

ROS2 - currently following some youtube channels and their wiki. Trying to make my own robot to change the world, or rather clean it up

my problem with the coral usb that I got was that Google's own AutoML service did not fully support it. Which sucks since they own it...

my startup tracks and finds items in videos. I'm sure I can spin up a new model for you. my email is in my profile

My problem is I've had a huge problem with getting tensorflow running on anything but a raspberry pi :(

The eco system around the pi is really amazing and I feel like its more alive. Which makes sense since its a learning platform.

I think it would be good to try something different. First I would get them good at typing code. This way they will not have to think about typing and thinking, they can just think. Second I would figure out a way to get them used to the concepts of programing, like for loops, variables, ect.. by doing other tasks. For example run a soccer drill like a for loop so they are learning secondly. I think this might help with some kids giving up early since its hard. It should be a little easier since they will have some experience with the concepts since they have run the drill before. Hope this makes sense, good luck!

I'm running the front and backend of the consumer site on Heroku. The meat of pipeline is hosted on a DigitalOcean High CPU Droplet. I use ffmpeg to extract images from the provided videos. I store everything in Google Cloud Storage and create references to each photo in Firestore. I use Firebase to power for the image verifying/labeling app I built. Its a simple app that presents the viewer with the image and the label that it was given. If its not correct they enter the correct label. I use a cloud function to move the images into an exportable format for autoML once a new image threshold has been hit. Testing is me using it and seeing if it is correctly identifying the objects.

I'm not a professional but I built a pipeline for Makers Part List - It involves ingesting a video URL, converting the video into images, then storing the images in google storage. Once stored I trigger the model to classify the image. The images are then displayed to annotators who verify/relabel the images. Once I get enough new images the system creates a .csv and uploads it to googles autoML where it retrains my model.

My bottlenecks now are splitting the videos into images as its a very CPU intensive process. Implementing a queue here is my best choice I think.

The point of autoML from what I've gathered is to make it as easy as can be to get a model working in production. AutoML at least AutoML Vision is using transfer learning to retrain X number of layers from their algorithm (the one google uses is escaping me right now). The number of layers it has to retrain is the value they offer, it tries and tries optimizing for accuracy.

I've had good results with it, but you do have to do things their way and its not always well documented. If you want more control you should create your own model and host it on google app engine, otherwise AutoML is what it is, no way to customize or tune it other than changing the training data you give it.

When I attended Hack Reactor they did tell us not to push them. However since they didn't teach us git (they expected us to know it) many still pushed them up. You would know because they'd get an email from some random company/person letting them know that they found their secret keys and that they should enroll/buy their services if they don't know what they're doing. Luckily no one from my class got hosed, but others in the past had.

I've been working on solar powered Peltier AC for my home. This past summer it was able to keep my living room (250sqft) at a comfortable 72F - 75F even when the outside temp was over 95F. It is crazy power hungry, using 12v 5A, but with solar its fine. The hard part as people have said is the heat dissipation and dealing with high ambient temperatures.

My newest design incorporates a few new upgrades. 1) geothermal cooling to dissipate the heat. As the ground temperature stays consistent at ~65F. Allowing me to not worry about the ambient temperature anymore 2) I moved to a 24V 6amp peltier which has a larger delta from the hot to cold side. 3) I now use a 24v ~7.6amp solar panel to power it. 4) I added a 15 Amp high-power motor driver so I can switch between producing cold and hot air. So now it will be viable in all weather conditions.

A core idea is to keep costs cheap which is why I have not added a battery to the mix. Although it would be nice as some nights are quite hot. I got the 24v panel used at $50 and the peltiers are super cheap (3 for ~15). The motor-controller/h-bridge was ~$40, but I didn't want to cheap out on that though and start a fire.

Its quiet unless I turn the fans up and doesn't cost anything to run after initial setup costs, due to the solar. My future plans for making it even better are to use multi-stage peltiers and try to leverage the hot side to produce more power. Also to add raspberry pi so I can control it remotely.