CodeProject AI on Github

[amdscooter]

Do a full system back up and go for it. Nothing to lose if you do it that way. I've finished work for Xmas so I will be doing this soon.

Did a full disk mirror b4 I tried 2.9.5... good advise.

[amdscooter]

Updated to cuDNN 9.6 without issue.

[TimG]

Updated to cuDNN 9.6 without issue.

Is there a CPAI batch file for that, or did you just use the exe file ? I'm getting ready

[TimG]

OK, looking good. For me it went like this:

14 Dec 2024 Installing CPAI 2.9.5 over 2.6.5
1. Download CPai 2.9.5 from https://codeproject.github.io/
2. Download latest NVIDIA driver for GTX1650.
3. Download latest CUDA driver.
4. Download latest cuDNN script.
5. Disable BI5 service. Uninstall CPAI (Add/remove).
6. Delete folder C:\Program Files\CodeProject
7. Delete folder C:\ProgramData\CodeProject
8. Remove 11.7 CUDA from add/remove.
9. Remove all old Nvidia drivers from Add/Remove. Reboot.
10. Install new Nvidia driver 566.36 (Express mode). Reboot.
11. Install NEW CUDA toolkit 12.6.3_windows.exe (Express mode). Kept failing. Found it due to Nsight. Unticked in custom install and installed separately. Reboot.
12. Run NEW cuDNN 9.6 install script as admin.
13. If no script follow these instructions https://docs.nvidia.com/deeplearning/cu ... ndows.html
14. Install CPAI 2.9.5. Let it complete installation - see web page for progress.
10. Reboot then re-enable and start BI5 service.

Yolov5.NET is OK, but ALPR keeps stopping.

2p9p5.png (255.94 KiB) Viewed 1411 times

[MikeLud]

OK, looking good. For me it went like this:

14 Dec 2024 Installing CPAI 2.9.5 over 2.6.5
1. Download CPai 2.9.5 from https://codeproject.github.io/
2. Download latest NVIDIA driver for GTX1650.
3. Download latest CUDA driver.
4. Download latest cuDNN script.
5. Disable BI5 service. Uninstall CPAI (Add/remove).
6. Delete folder C:\Program Files\CodeProject
7. Delete folder C:\ProgramData\CodeProject
8. Remove 11.7 CUDA from add/remove.
9. Remove all old Nvidia drivers from Add/Remove. Reboot.
10. Install new Nvidia driver 566.36 (Express mode). Reboot.
11. Install NEW CUDA toolkit 12.6.3_windows.exe (Express mode). Kept failing. Found it due to Nsight. Unticked in custom install and installed separately. Reboot.
12. Run NEW cuDNN 9.6 install script as admin.
13. If no script follow these instructions https://docs.nvidia.com/deeplearning/cu ... ndows.html
14. Install CPAI 2.9.5. Let it complete installation - see web page for progress.
10. Reboot then re-enable and start BI5 service.

Yolov5.NET is OK, but ALPR keeps stopping.


2p9p5.png

Due to the age of your GPU most likely it will not work with CUDA 12.6, use CUDA 11.8

[TimG]

Thank you for that MikeLud. The ALPR is still turning off, so I will look into that.

Is there a list of what works with each gpu anywhere ? I was hoping just to be able to use the latest versions of everything.

[TimG]

OK, here we go. ALPR is definitely hosed, so changes are afoot:

16 Dec 2024. CUDA 11.8 to replace CUDA 12.6
1. Download CUDA 11.8 driver.
2. Disable BI5 service. Uninstall CPAI (Add/remove).
3. Delete folder C:\Program Files\CodeProject
4. Delete folder C:\ProgramData\CodeProject
5. Remove 12.6 CUDA from add/remove.
6. Install NEW CUDA toolkit 11.8_windows.exe (Express mode).
7. Run NEW cuDNN 9.6 install script as admin.
8. If no script follow these instructions https://docs.nvidia.com/deeplearning/cu ... ndows.html
9. Install CPAI 2.9.5. Let it complete installation - see web page for progress.
10. Reboot then re-start BI5 service.

So far so good. ALPR isn't stopping, but I'm waiting for a car to trigger it

[TimG]

And now working OK. Thanks to MikeLud

cpai.png (196.62 KiB) Viewed 1203 times

[amdscooter]

Thank you for that MikeLud. The ALPR is still turning off, so I will look into that.

Is there a list of what works with each gpu anywhere ? I was hoping just to be able to use the latest versions of everything.

Glad you got it worked out re: previous post.

I have CUDA 12.6.3 downloaded but am holding off installing. A list of supported GPU's by generation would be useful. I'm running a 2060 Super but will stand pat on 11.8.89 for now.

[MikeLud]

Below is the list of supported GPU's by generation for the ALPR module

Screenshot 2024-12-16 174608.jpg (241.11 KiB) Viewed 1151 times