@vivekpawar4847
Thank you very much sir for such awesome tutorials I trained model to recognize ironman and its working completely fine.....🖖 Learned alot...!
@behnoudshafizadeh3281
thanks mr mark jay,,,this is the best course for begginers and professional person.
@Robots-and-androids
Hi Mark. YOu certainly got YOLO nailed. I am attempting to implement it on my own PC. I have a robot and I want it to be able to identify the objects in front of it. -Thomas
@go64bit
Finally the most awaited video!
@SunShinekr
Thank you very so much. I trained model to detection number plate and its working fine. Hope you make more awesome tutorials!
@sureshganesan8746
Mark, Thanks for a lovely video. I could train the custom object. As I understand, Yolo is classifying 80 objects and hence the final layer is (80+5)*5 parameters of 19x19 neurons. What needs to be done, if I need to include Fidget Spinner as 81st object for detection rather than isolated detection
@sebasman1000
have been fooling around with tensorflow object detection for quite some time now, and this tutorial is an absolute breeze, thanks a million. also for everybody getting the "expected x bytes, found y" error. look at : https://github.com/thtrieu/darkflow/issues/338
@MuhammadIhsan-wd5if
Hello, hope this comment finds you well. It was 6y old video but, I wonder that you read it and answer my question. Now I working on robot that must has ability to detect an object. The challenge is I must use a low budget component and device. Which is I plan to use nvidia gtx 1050 as a gpu. My question is, what gpu that you use, and how much fps that you hot when you run it on real time. Thank you
@amniamadour9010
you are the best ! thank you so much
@yundahong940
Thank you Mark ,it works perfectly and it's really help
@akshayd2304
Thank you , really great series (y)
@ngochieuthai5436
Thank you Mark! What a great tutorial! However, I came up with 2 question. 1/ If my intention was to draw bounding box, there are two ways of approaching. The first one, of course, is by YOLO. My second though is to build a simple CNN( LeNet for example) in order to classify the image into classes with confidence score. Then i'll use the pyramid model and sliding windows to raise ROI for classification. Any ROI who scores higher than a threshold for any classes will be put inside a rectangle. Which one do you think is a better approach for a 2 or 3 class problem ? 2/ Are we doing transfer learning while retraining YOLO weights?
@donfung8665
Hey Mark I'm curious to know if you know the theory/reason behind why the bounding box jitters when the object in the video is static? 15:50-16:00
@andrezx10r
Hello, excellent tutorial! Thank you!!! But I have a question ... After the complete training, do I test the network with my test data with the same call via the checkpoint, just like the video? And is there any specific outgoing message that indicates that the network has been trained correctly, without error? Can you indicate a relatively good loss value to have a recognition of the object in the test?
@tahsinsiad4515
Sir, I want to add a new class to the existing pre-trained model. How can I do that?
@tobechukwunwatu848
Thank you !! very much sir for this tutorial. Awesome !!!
@StealthWhale80
when I run the video.py the network loads, but no video feed pops up and I get this:
Traceback (most recent call last):
File "video.py", line 22, in <module>
results = tfnet.return_predict(frame)
File "/mnt/c/Users/abell/Desktop/darkflow-master/darkflow/net/flow.py", line 78, in return_predict
'Image is not a np.ndarray'
AssertionError: Image is not a np.ndarray
@rubencg195
Awesome tutorials!
@farkasistvanfi
Thank you very much! It's great! I have done the 8 parts.
@srikanthvelpuri2973