fix the xxx.cudnn.so.8 missing error

[IsaacMarsh]

The base image and requirements were done well and tested!

docker exec -ti d7d592e84c1e /bin/bash
root@d7d592e84c1e:/app# nvidia-smi
Fri Dec 6 02:18:38 2024
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 555.42.06 Driver Version: 555.42.06 CUDA Version: 12.5 |
|-----------------------------------------+------------------------+----------------------+
| GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|=========================================+========================+======================|
| 0 NVIDIA GeForce RTX 3070 Off | 00000000:03:00.0 Off | N/A |
| 0% 35C P8 17W / 240W | 350MiB / 8192MiB | 0% Default |
| | | N/A |
+-----------------------------------------+------------------------+----------------------+

+-----------------------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=========================================================================================|
+-----------------------------------------------------------------------------------------+

Note that the host has cuda 12.5 and the container has cudnn 8 files

[MahmoudAshraf97]

Hi, if this suits q project you are running, that's okay, but this is a solution to a non-existent problem, installing the requirements withou pinning requires cudnn9, and we should not pin requirements unnecessarily

[IsaacMarsh]

Sorry for the inconvinience, but I doubt if anyone else had tested the docker file thoroughly

[MahmoudAshraf97]

Maybe the docker file isn't the most popular thing in this repo, so a CI test will be helpful if you can add it