omarabid59/yolo_google-cloud

Google Cloud Machine Learning for backend YOLO v3 object detection inference with a browser based WebRTC connection

yolo_google-cloud's Issues

Not able to connect to the http server in the final step

After following each step, during the final step in order to launch the website, I'm not able to connect. It just says connection refused.

What is causing this and what can be done to solve it? Is there something that need to be changed in the Google Cloud platform?

Thanks

Uncaught ReferenceError: wcSendVideoCanvas is not defined

Uncaught ReferenceError: wcSendVideoCanvas is not defined
at loadWebcam (init_screen.js:29)
at initScreen (init_screen.js:39)
at main.js:4

Browser just shows Object Detection On the Cloud Demo! and nothing else

Application is broken

Error occured sending or receiving data on ML client. 'tuple' object has no attribute 'get'

It seems that it fails at this line
https://github.com/omarabid59/YOLO_Google-Cloud/blob/master/mlserver/MODULE_DATA.py#L20

CUDA Error: no CUDA-capable device is detected

I'm running the code as instructed in the documentation to this GitHub repo, and I'm getting the following error logged in nohup.out file after runnning nohup python3 ~/YOLO_Google-Cloud/mlserver/mlserverclient.py &

layer filters size input output
0 python3: ./src/cuda.c:36: check_error: Assertion `0' failed.
CUDA Error: no CUDA-capable device is detected

Any idea what might be wrong?

Object detection not working

I've managed to setup everything right, and the website with a video steam work just fine. But I'm not detecting any object in the video stream (I'm in the camera view, so a person should have been detected).

Here is what I get on the server side in my GCP SSH Terminal:

INFO:ice:Connection(0) ICE failed
DEBUG:ice:controlled - checking -> failed
DEBUG:dtls:client - State.NEW -> State.CONNECTING
DEBUG:dtls:client x DTLS handshake failed (connection error)
DEBUG:dtls:client - State.CONNECTING -> State.FAILED
ICE connection state is failed
DEBUG:ice:controlled - failed -> closed
ICE connection state is closed

Regarding my YOLO model

I have trained YOLOv3 model on my own custom dataset so by just replacing my cfg , name and data files will this work out without any issues???

Error occured sending or receiving data on ML client.

I am able to run both script but once i open url in browser it's break due to some json loads error.
What i have found through troubleshooting. there is issue with MODEL_DATA.py in updateData function. I checked variable it doesn't contain dictionary.
Here is log of mlserverclient.py
Loading weights from /home/gb-pc08/Project/aiortc/examples/server/YOLO_Google-Cloud/mlserver/model/yolov3-tiny.weights...Done! Starting YOLO Predictor Thread Starting ZeroMQ DataHandler DATA Connected! MESSAGE Connected! Exception in thread ZeroMQ DataHandler: Traceback (most recent call last): File "/usr/lib/python3.6/threading.py", line 916, in _bootstrap_inner self.run() File "/home/gb-pc08/Project/aiortc/examples/server/YOLO_Google-Cloud/mlserver/ZeroMQ.py", line 68, in run self.update(self.name) File "/home/gb-pc08/Project/aiortc/examples/server/YOLO_Google-Cloud/mlserver/ZeroMQ.py", line 76, in update self.moduleData.updateData(data) File "/home/gb-pc08/Project/aiortc/examples/server/YOLO_Google-Cloud/mlserver/MODULE_DATA.py", line 43, in updateData data = json.loads(message) File "/usr/lib/python3.6/json/__init__.py", line 354, in loads return _default_decoder.decode(s) File "/usr/lib/python3.6/json/decoder.py", line 339, in decode obj, end = self.raw_decode(s, idx=_w(s, 0).end()) File "/usr/lib/python3.6/json/decoder.py", line 357, in raw_decode raise JSONDecodeError("Expecting value", s, err.value) from None json.decoder.JSONDecodeError: Expecting value: line 1 column 1 (char 0)

Messages flushing in terminal after running both steps, but no image is showing in the website