Відео

Hi! Did you try with udp?

@@universityofthefuture2077 yes. The udp technically should be faster, but with unstable internet it bring very big video artifacts. Tcp makes it much clearer. My usage is for moving device, so it will use mobile network. Anyway, i will be trying with bare tcp for now, maybe ill reach the desired latency. If not, rtsp or webrtc.

Noted, thanks!🫡

You need drivers on your host machine

Thanks;) What do you mean, “over the network” -> as your personal GPU station or using providers like AWS, Google etc?

@@universityofthefuture2077 I have some old GPUs running somewhere remote (from old bitcoin mining days haha) and i am looking for a way to use them on my computer (or VPS like you mentioned AWS) without being physically connected to my motherboard, lets say for training AI model. Did i answer your question? haha

Hyym, cant you just connect to this remote via ssh?

@@universityofthefuture2077 I dont know how, i searched youtube but no video explained it good for me. But this video was just perfect. Just an idea for next video i thought.

I am not😛

check out video description for source code link:)

Thanks! Sure, it’s public on my github:)

Never trust strangers on the internet..😆sorry my mistake, I added it minute ago, thanks:)

Thanks! You can easily export yolo to tflite format (details here: docs.ultralytics.com/integrations/tflite/#deployment-options-in-tflite). I’ve never done this, so I am not sure if ESP can handle it, but probably yes. Hover if your goal is just to make it run, it would be much easier to use rpi zero for example:)

Hi, now its late response, hah, but this snippet it’s quite unreadable, please provide some github repo

@@universityofthefuture2077 i already solve the problem but thanks 😉

@@universityofthefuture2077 guys can you help me please? a have a this problem to. Thanks

@@rodrigocorte5289 can you help me please? a have a this problem to. Thanks

@@rodrigocorte5289 can you help me please?i hame i same problem

Hi, what do you mean? You can add simple alarm with for example playsound library.

@@universityofthefuture2077 Hello and thanks for the quick response!! It's not that simple, I put an if to play sound when it detects fire but it doesn't work. I tried with this code : from ultralytics import YOLO from playsound import playsound # Função para detecção de fogo e alerta sonoro def detect_fire_and_alert(): # Inicializa o modelo YOLO model = YOLO('best.pt') # Realiza a detecção contínua de fogo while True: # Realiza a predição results = model.predict(source=0, imgsz=320, conf=0.6, show=True) # Verifica se há detecções de fogo for detection in results.pred: if 'fire' in detection['names']: # Se o fogo for detectado, reproduz um som de alerta playsound("alarm_sound.mp3") break # Interrompe o loop quando o fogo é detectado # Chamada da função para detecção de fogo e alerta sonoro detect_fire_and_alert() I can´t do it, could you help me with this

@@universityofthefuture2077 Hi bro, thanks for responding so quickly It's not that simple, I've already tried with this code and I'm not succeeding. from ultralytics import YOLO from playsound import playsound # Function for fire detection and sound alert def detect_fire_and_alert(): # Initialize the YOLO model model = YOLO('best.pt') # Performs continuous fire detection whileTrue: # Perform the prediction results = model.predict(source=0, imgsz=320, conf=0.6, show=True) # Checks for fire detections for detection in results.pred: if 'fire' in detection['names']: # If fire is detected, plays an alert sound playsound("alarm_sound.mp3") break # Breaks the loop when fire is detected # Function call for fire detection and audible alert detect_fire_and_alert() If you could help I would be very grateful

@@universityofthefuture2077 Hi bro, thanks for responding so quickly. This is not that simple. I've tried several times and still haven't managed to get an alarm to sound when it detects the fire. I used this code and I can't do it, do you have any suggestions? from ultralytics import YOLO from playsound import playsound # Function for fire detection and sound alert def detect_fire_and_alert(): # Initialize the YOLO model model = YOLO('best.pt') # Performs continuous fire detection whileTrue: # Perform the prediction results = model.predict(source=0, imgsz=320, conf=0.6, show=True) # Checks for fire detections for detection in results.pred: if 'fire' in detection['names']: # If fire is detected, plays an alert sound playsound("alarm_sound.mp3") break # Breaks the loop when fire is detected # Function call for fire detection and audible alert detect_fire_and_alert() I would be very grateful if you could help me. Bro i really need your help!!

@@universityofthefuture2077 Hi bro, thanks for responding so quickly!! I've tried several times and I can't get an alarm to sound when fire is detected. I used this code, but it doesn't work. Do you have any suggestions? # Function for fire detection and sound alert def detect_fire_and_alert(): # Initialize the YOLO model model = YOLO('best.pt') # Performs continuous fire detection whileTrue: # Perform the prediction results = model.predict(source=0, imgsz=320, conf=0.6, show=True) # Checks for fire detections for detection in results.pred: if 'fire' in detection['names']: # If fire is detected, plays an alert sound playsound("alarm_sound.mp3") break # Breaks the loop when fire is detected # Function call for fire detection and audible alert detect_fire_and_alert() I would be very grateful if you could help me. I need your help!!

Prediction is quite fast even without gpu and real-time video streaming on Colab requires some additional set up:)

What do you mean :D ?

Have you install all necessary packages?

Yes but, while importing yolo from ultralytics , it is showing a yellow underline below that I think its logical error

Maybe it’s different version of the package?

do you write all codes ??

can you show the full error?

done @@universityofthefuture2077

You can zip it, then download option should appear when you click three dots

In my case it was in PROJECT/runs/detect/train/weights/best.pt, but if you are using different version of the ultralytics package, the path can be slightly different.

Really depends on your goal. You should keep track of loss chart, if the curve stops going down, it's sign that probably further training is not going to make it better.

Hi, which part of the code fails?

try updating your paths to the image folders in data.yaml: test: ./test/images train: ./train/images val: ./valid/images

After you run the script, the folder should be created automatically in the project root directory

@@universityofthefuture2077 alright thank youu, I will try it

I followed the same instructions but the webcam doesn't run for me. The webcam is activated and all. What can be the problem?

@@universityofthefuture2077 where to find the project root directory?

przynajmniej nie wyrzucił na ulice

github.com/mcdominik/simple-scripts/blob/master/cow.py

universe.roboflow.com/-jwzpw/continuous_fire