I had been stuck for hours trying to install TensorRT and the lack of clear support from NVIDIA or any forum or Stack Overflow seemed impossible until I came across your video. Thank you so much for this
@Smi1eM3 Жыл бұрын
No problem.
@IsiUbilla3 ай бұрын
You just save my mental health, thank you a lot!
@yuval_kahan9 ай бұрын
wow wow wow so simple but so hard to understand by myself thank you so much!!!! THANK YOU !
@ShittheswaranSelvakumar Жыл бұрын
can you please make a video on yolov5 with pytroch on TensorRT ? is that possible?
@matthewbowman624711 ай бұрын
Hey I was wondering if you changed your discord was trying to add you but it wasn't coming up? wanted to talk about circumventing detection dual pc vs bluetooth etc
@g83.076 ай бұрын
idk how to use my best.engine with YOLOV8 plz help
@luizconrado Жыл бұрын
Is there any chance you could make a couple more videos showing how to do the same work using YOLO V8 together with PyTorch and TensorRT?
@Smi1eM3 Жыл бұрын
I will see if I have the time.
@luizconrado Жыл бұрын
@@Smi1eM3 That would be very nice. I was able to get amazing results with YOLOv8, however, I am having hard time moving my YOLOv8 model into PyTorch and TensorRT. . Is there anything I could do to help you get these videos ready?
@Smi1eM3 Жыл бұрын
@@luizconrado YoloV8 is still in its early stages and has less support at the moment. Yolov5 with tensorrt will beat base Yolov8 performance until Yolov8 gets support for tensorrt.
@aishwaryaravi5249 Жыл бұрын
Yes please, i hope you can make a video on this
@AJ_JAM_M006 Жыл бұрын
None of the python wheels are working. Do you know what might be the error??
@yeetyeet447910 ай бұрын
People who have passed the arduino step, what mouse did you use? My glorious model o- doesn’t work and I can’t find a mouse parser.
@easycodelessons31928 ай бұрын
recently bought g102 ligtsync for this project, works well for me
@Cat-gd8ju4 ай бұрын
use a kmbox
@EVOasalto10 ай бұрын
Do you know anything related to this error? AssertionError: input size torch.Size([1, 3, 416, 736]) not equal to max model size (1, 3, 736, 736). I exported its with img size 736 like my model
@Smi1eM310 ай бұрын
export the model in a different size. make it 416,736 look at he other comments.
@burhanahmed3525 Жыл бұрын
hello sir. i really like this tutorial very much. a small issue im facing here. i followed all the instructions u said. tensorrt is successfully installed in my windows 10 but when i convert .pt to engine format using the same env from which i installed tensorrt. it is showing me that " there is no module called tensorrt ". although i set all the paths in environment variables but still facing this issue. Sir do you have idea with this issue or can you share tips maybe it works for me. i will be really thank full to you
@vrai00811 ай бұрын
same error. did you manage to solve it?
@vrai00811 ай бұрын
Solved it. I was using wrong file as path while copying 10:20 I was installing the wrong file in python folder
@user-wc8lz9pk9e10 ай бұрын
unpickling stack underflow. Cache may be out of date, try `force_reload=True`, this comes up when i put .engine file path in the code its runs alright with .onnx but not with .engine, any fixes?
@sambokalo15928 ай бұрын
Did u fix it? I have that now
@muhammadviandyrasyah93706 ай бұрын
me too
@adityadas1765 Жыл бұрын
where to change the loading model from yolov to tensor?
@adityadas1765 Жыл бұрын
like how to switch like you did..
@Smi1eM3 Жыл бұрын
@@adityadas1765 Just switch the file path to the tensorrt version
@rallyworld3417 Жыл бұрын
very interesting .. when i create best.engine file with this method and same config like u .. mine .engine file is 2x bigger then best.pt ???
@Smi1eM3 Жыл бұрын
Yeah it’s generally larger I believe
@rallyworld3417 Жыл бұрын
@@Smi1eM3 but .engine file is smaller then .pt file in your own repo? Bruh
@Smi1eM3 Жыл бұрын
@@rallyworld3417 half precision
@Boltron4 ай бұрын
i found the step of adding the path to the folder on the desktop kinda weird becuase we placed all the files into cuda folders which are already in the path. edit: thanks for the instructions for the tenssorRT installation . but people like you making cheats for video game are the problem with FPS games today . why not just get good. and use your talents of coding for something else.
@Smi1eM33 ай бұрын
Just an educational course 😄
@YassineIsco Жыл бұрын
At the end is it actually possible to make it a standalone with all the "improvements" included. So even a friend with nothing installed but Arduino can actually run & use this decently at 60fps?
@Smi1eM3 Жыл бұрын
no
@davia.sampaio8633 Жыл бұрын
What is your hardware?
@Smi1eM3 Жыл бұрын
Laptop 3060 is my gpu
@davia.sampaio8633 Жыл бұрын
@@Smi1eM3 do you think it's possible to get 40 fps in a gt1030 with a i5 9000k?
@Smi1eM3 Жыл бұрын
@@davia.sampaio8633 No the graphics card is not good enough
@dzizuru-sama17 күн бұрын
8:15
@moxxaaf9197 Жыл бұрын
what if i only have amd vga ?
@ahmadullaheshan1644 ай бұрын
when i did this my fps dropped from 50 to 25
@ahmadullaheshan1645 ай бұрын
hey i will pay u 300 dollars if u can fix mine to word with a coral ai usb accelerator
@Smi1eM35 ай бұрын
Are you connecting the accelerator to an external device to run the model or your own computer?
@ahmadullaheshan1645 ай бұрын
@@Smi1eM3 ok so i connected my accelerator to my pc i converted pytorch to tensorflow lite like my best.pt i set up the accelerator tested it with the method google provided on there website and it worked now i just need to modify the code a little to make it work with the USB accelerator instead of my gpu
@ahmadullaheshan1645 ай бұрын
@@Smi1eM3so you think u can edit the code for me that it will work with coral ai
@sambokalo15929 ай бұрын
Hello Sir, i want to thank you for this amazing Course. I followed through it and everything is working. I just have one problem, when i switch to best.engine it gives me this error( Cache may be out of date, try `force_reload=True`). Is there a way to fix it? Anyways thx again