# Jupyter Notebook
* On GPU node
`jupyter lab --no-browser --port 7077`

* From local machine
`ssh -t -t [e-mail] -L 7077:localhost:8088 ssh n0002 -L 8088:localhost:7077`
# Conda
Create `.condarc` in $HOME directory and add the following

auto_activate_base: false
envs_dirs:
  - /path/to/custom/envs/directory
pkgs_dirs:
  - /research/arcs/am2455/softwares/conda/pkgs
channels:
  - conda-forge

If you want to automate package installations without user prompts, you can use the `-y` option like this:
`conda install -y package_name`

## Write Conda installation log
`conda install -y package_name > install_log.txt 2>&1`
## PyTorch

### Install sklearnex
`conda install -c anaconda scikit-learn-intelex`
`python -m sklearnex CBert_SVM.py`

### Documentation of sklearnex
https://intel.github.io/scikit-learn-intelex/

precision    recall  f1-score   support

           0       0.80      0.71      0.75      4631
           1       0.92      0.95      0.93     16424

    accuracy                           0.90     21055
   macro avg       0.86      0.83      0.84     21055
weighted avg       0.89      0.90      0.89     21055

## Set the CUDA_HOME environment variable with this command.

`conda env config vars set CUDA_HOME="/path/to/environment" -n env`

Please note that you should replace `/path/to/environment` with the actual path of your Conda environment, and replace `env` with the name of your Conda environment.

#
`conda create -n tf-gpu tensorflow-gpu`

# Create executable from Pyton script
`pyinstaller --onefile sq.py`