Thank you so much for this explaination! Really helped me for my Final Year Project at University
@АльбертИванов-ц4х2 жыл бұрын
hi! thanks for sharing! i have succesfully trained yolox_s on windows with GTX 1060. but when i checked the results i saw classes person and bycicle (should be fire and smoke) 😂 bound boxex are correct. i corrected the CLASSES before the start train. but may be somewhere else it should be done ?
@manirajan_2 жыл бұрын
Hi, Please check coco_classes.py to have only these 2 classes. If still not working, you can define coco classes in tools/demo.py before calling Predictor COCO_CLASSES = ( "Fire", "Smoke" )
@АльбертИванов-ц4х2 жыл бұрын
@@manirajan_ thanks! got it working. just have 2 installations of yolox on different drives and import was from another one )
@harshwardhankurale3109 ай бұрын
Error installing torch torchvision torchaudio ….” could not find a version that satisfies the requirements”……please help..which versions to use?
@syaikhulawwali14214 ай бұрын
i got this error message: "RuntimeError: CUDA error: device-side assert triggered Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions." Do you have idea how to solve it , sir?
@MsModestMan2 жыл бұрын
Thank you for the video! Tell me, can you share a link to the program code of the telegram bot that sends photos of the fire? Or make a tutorial video on how to create such a bot with yolox.
@manirajan_2 жыл бұрын
Hi, telepot is used for telegram bot. below is api document. telepot.readthedocs.io/en/latest/
@cuocsongcuayen96162 жыл бұрын
Tks so much
@MsModestMan2 жыл бұрын
thanks for the help!
@kesavan1157 Жыл бұрын
how to train seperately for train and val sir?
@mahmodkohel59592 жыл бұрын
Hey sir, how i can export the model with h5 file format
@suleymangunes8022 жыл бұрын
Could you share your json, train data?
@manirajan_2 жыл бұрын
Hi, Please find uploaded dataset in kaggle and link is www.kaggle.com/datasets/deeplearn1/fire-and-smoke-bbox-coco-dateset
@suleymangunes8022 жыл бұрын
@@manirajan_ Thank you sir.
@abubakarsaleem51672 жыл бұрын
Sir can u Please share,demo.py file
@manirajan_2 жыл бұрын
Hi, Please find the file in below path and update with your telegram token and receiver id github.com/manirajanvn/yolox_custom_training
@abubakarsaleem51672 жыл бұрын
Sir can u please guide mei where i can find token and receiver id in telegram?
@FalahgsGate2 жыл бұрын
I was training custom dataset by yolox and get the final model in latest_ckpt.pth.tar format how to convert it to ncnn for android application could u please help me and thanks for help
@manirajan_2 жыл бұрын
Hi, I also have to try, meanwhile below link might help you yolox.readthedocs.io/en/latest/demo/ncnn_android_readme.html
@ghassenhajjem72332 жыл бұрын
I met this while I train my model yolox.core.launch:90 - An error has been caught in function 'launch', Also I have keyerrors