kobybibas

# Having a requirements.txt file has follows
torch==1.5.0
numpy==1.18.1

# Add channels. Last added is with the highest priorety
conda config --add channels pytorch
conda config --add channels conda-forge
conda config --add channels anaconda

# Install pip for fallback

kobybibas

# In .ssh/config file:
Host aws
    HostName **********.compute.amazonaws.com
    User ubuntu
    IdentityFile ******.pem
    IdentitiesOnly yes
    LocalForward 6006 0.0.0.0:600
    
# In aws server
tensorboard --logdir .

kobybibas

To remove a submodule you need to:

• Delete the relevant section from the .gitmodules file.
• Stage the .gitmodules changes git add .gitmodules
• Delete the relevant section from .git/config.
• Run git rm --cached path_to_submodule (no trailing slash).
• Run rm -rf .git/modules/path_to_submodule (no trailing slash).
• Commit git commit -m "Removed submodule "
• Delete the now untracked submodule files rm -rf path_to_submodule

kobybibas

Summry of the following guide https://nvlabs.github.io/eccv2020-mixed-precision-tutorial/files/szymon_migacz-pytorch-performance-tuning-guide.pdf

1. In the dataset class

pin_memory=True

2. Enable for device specific CNN acceleration

kobybibas

# Crop such that the output is heightxheight. Then scale to 640:640. finally videeo compression ("28", higher is more compression)
ffmpeg -i intput -filter:v "crop=ih:ih,scale=640:-2" -crf 28  output.mp4

# Split a video to 280 sec segments
ffmpeg -i input.mp4 -acodec copy -f segment -segment_time 280 -vcodec copy -reset_timestamps 1 -map 0 output_%d.mp4

# Crop with side by side merge (https://stackoverflow.com/questions/35349935/ffmpeg-crop-with-side-by-side-merge)
ffmpeg -i in0.mp4 -i in1.mp4 -filter_complex "[0:v]crop=iw/2:ih:0:0[left];  [1:v]crop=iw/2:ih:ow:0[right]; [left][right]hstack" output.mp4

kobybibas

conda install -y pip
conda install -y -c conda-forge matplotlib
conda install -y -c anaconda numpy
conda install -y pandas
yes | pip install hydra-core --upgrade
yes | pip install hydra-joblib-launcher --upgrade
conda install -y -c anaconda scipy
conda install -y scikit-learn
conda install -y -c conda-forge jupyterlab
conda install -y tqdm

kobybibas

1. Download and run: https://www.nvidia.com/Download/Find.aspx
2. Reboot
3. Install cuda driver from: https://developer.nvidia.com/cuda-downloads
4. Reboot
5. Make sure nvidia-smi equals nvcc --version

kobybibas

Instructions are from
https://wiki.blender.org/wiki/Building_Blender/Other/BlenderAsPyModule Tested with Blender 2.93 on macOS 10.15.7

1. Install xcode command line

sudo xcode-select -s /Applications/Xcode.app/Contents/Developer

2. Install python 3.7
   https://www.python.org/ftp/python/3.7.9/python-3.7.9-macosx10.9.pkg

kobybibas

# To see what volumes are avilable
lsblk 

# c5
sudo mount /dev/nvme1n1 /data

# p2
sudo mount /dev/xvdf /data

kobybibas

// launch.json: 
{
    // Use IntelliSense to learn about possible attributes.
    // Hover to view descriptions of existing attributes.
    // For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387
    "version": "0.2.0",
    "configurations": [
        {
            "name": "Python: Current File",
            "type": "python",