Hardware

:black_small_square: RAM

total        used        free      shared  buff/cache   available
Mem: 115Gb       340Mb        98Gb       1.0Mb        15Gb       113Gb
Swap:   0B          0B          0B

:black_small_square: CPU

Architecture:                    x86_64
CPU op-mode(s):                  32-bit, 64-bit
Byte Order:                      Little Endian
Address sizes:                   46 bits physical, 48 bits
                                  virtual
CPU(s):                          14
On-line CPU(s) list:             0-13
Thread(s) per core:              1
Core(s) per socket:              1
Socket(s):                       14
NUMA node(s):                    1
Vendor ID:                       GenuineIntel
CPU family:                      6
Model:                           61
Model name:                      Intel Core Processor (Bro
                                 adwell, IBRS)
Stepping:                        2
CPU MHz:                         2399.996
BogoMIPS:                        4799.99
Hypervisor vendor:               KVM
Virtualization type:             full
L1d cache:                       448 KiB
L1i cache:                       448 KiB
L2 cache:                        56 MiB
L3 cache:                        224 MiB
NUMA node0 CPU(s):               0-13

:black_small_square: GPU

description: 3D controller
product: GP100GL [Tesla P100 PCIe 16GB]
vendor: NVIDIA Corporation
physical id: 5
bus info: pci@0000:00:05.0
version: a1
width: 64 bits
clock: 33MHz
capabilities: pm msi pciexpress bus_master cap_list
configuration: driver=nvidia latency=0
resources: iomemory:200-1ff iomemory:240-23f irq:11 memory:fd000000-fdffffff memory:2000000000-23ffffffff memory:2400000000-2401ffffff

⚙️ Config

In /opt/work/ are located the Darknet directory and weights
- /opt/work/ have 777 rights
To connect into the server we have to change the SSH port to 2222 with -p 2222

📋 Installed:

Nvidia drivers, CUDA and CUDNN
vim
libopencv-dev
opencv-python
Python3-pip
python3 -m pip install jupyterlab # Each user need to install it in order to get it working individually.
- After you installed it. Add this into .bashrc: export PATH="$HOME/.local/bin:$PATH"
- Highly recommend to put password to access into Jupyterlab. Do this the first time you run the jupyterlab.
nomacs

🗜 YOLO config & scripts:

YOLO configuration & scripts - NAI20SP

🖼 nomacs (Image viewer)

If you want to load an image run (Remember X11 flag):

nomacs image.jpg

🛂 X11forwarding:

When we run the the predictions we get back a picture with the predictions in coloured boxes and written in the terminal.

In order to get windows predicctions over SSH we enabled the X11forwarding.

To get into the SSH with the X11forwarding permissions add the flag -X :

ssh -X name@IP -P 2222

📒 Access into Jupyter-lab:

In case we have multiple users in the server and we will them use jupyterlab, each user need to install jupyterlab and choose a port for each user also as this example:
| Username | port |
| :——————- | :—————: |
| user1 | 8881 |
| user2 | 8882 |

To check if we have Jupyter-lab running you will get back the PID number with this command: pgrep jupyter-lab
- In case we don’t have jupyter running:

cd /opt/work/
jupyter-lab --no-browser --port=8881

Command to run in our computer:

ssh -N -f -L localhostlocalhost:8881 username@server_IP -p 2222

Open browser and type:
```
localhost:8888
```
🛠 Check if there are training on going:
If we get back a nummber means that some user is doing a training.
```
pgrep darknet
```

🗃 Doing a training with tmux

If we do a training connected by SSH and we lost the conection, all will be gone in case the process is not running on background.

Using tmux we can have a session attached working all the time until we remove it. So we can disconnect and reconnect anytime.

Creating a session called “work”:

tmux new -s work

🛠 Now we can run our training …

↪️ To deattach the session press:

Ctrl + b, d

Now we can left the server without any problem 🆒! .

↩️ To attach again into the last session use:

tmux a

🗂 To see the active sessions:

tmux ls

📎 To attach into session by name:

tmux attach-session -t my_session

📥 Download a file from server to our computer by SCP:

With the following command you will download the file yolov3_best.weights into your current directory.

scp -P 2222 user@IP:/opt/work/darknet/data/my_test_data/backup/yolov3_best.weights .