DeepstackGPU issue with Nvidia Tesla K10

[marrobHD]

I tried DeepstackAI GPU on Ubuntu 20.04 and Windows Server 2019. DeepstackGPU does not want on either operating system. DeepstackCPU works like a charm.

I used this python script:

import requests

image_data = open("family-and-dog.jpg","rb").read()

response = requests.post("http://localhost:80/v1/vision/detection",files={"image":image_data}).json()

for object in response["predictions"]:
    print(object["label"])

print(response)

C:\Users\Administrator>nvidia-smi
Sat Mar 20 20:57:53 2021
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 461.33       Driver Version: 461.33       CUDA Version: 11.2     |
|-------------------------------+----------------------+----------------------+
| GPU  Name            TCC/WDDM | Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  Tesla K10           TCC  | 00000000:1B:00.0 N/A |                  N/A |
| 30%   34C    P8    N/A /  N/A |      9MiB /  4035MiB |     N/A      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+
PS C:\Users\Administrator> deepstack --VISION-DETECTION True --PORT 80
DeepStack: Version 2021.02.1
/v1/vision/detection
---------------------------------------
---------------------------------------
v1/backup
---------------------------------------
v1/restore
[GIN] 2021/03/20 - 20:57:29 |�[97;41m 500 �[0m|          1m0s |   192.168.1.182 |�[97;46m POST    �[0m /v1/vision/detection
[GIN] 2021/03/20 - 20:57:56 |�[97;41m 500 �[0m|          1m0s |   192.168.1.182 |�[97;46m POST    �[0m /v1/vision/detection
[GIN] 2021/03/20 - 21:01:44 |�[97;41m 500 �[0m|          1m0s |   192.168.1.182 |�[97;46m POST    �[0m /v1/vision/detection

[gillonba]

It's been a while for me and I haven't used it on Windows, but don't you have to set a command line parameter to use the GPU?

[marrobHD]

On the docs about DeepstackGPU for windows, there isn't nothing about special arguments. Probably they aren't needed because DeepstackGPU works without docker.
On linux it is a different thing.
The --gpus all parameter allows deepquestai/deepstack:gpu to access all gpus on the system. It's needed because docker doesn't allow all containers to access all resources.

[marrobHD]

@johnolafenwa Any updates on this?