Python Cheatsheet

Huggingface

1. Enable large files
   Verify?

   git lfs version
   

   Install?

   git lfs install
   

   Enabling

   huggingface-cli lfs-enable-largefiles .
   

Docker

1. Copy file(s)

   docker cp b4a34cd339de://app/saves/Llama-7B/lora/train_2024-11-16-12-13-34/training_loss.png D:\
   

2. show container

   docker ps -a
   

3. building and composing docker image

   docker-compose build --no-cache && docker compose up -d --force-recreate && docker compose exec llamafactory bash
   

4. docker-compose logs -f

docker-compose logs -f

Conda

Enlist installed packages in the Env

1. listing packages

   conda list
   

2. To list specific pkg

   conda list pkg_name
   

Conda Activate/Deactivate Env

1. Open Anaconda Powershell Prompt
2. Run

   Conda init bash
   

3. To activate

   conda activate env_name
   

4. To deactivate

   conda deactivate
   

Conda Create Environment - CMD

1. To create

   conda create -n yourenvname python=x.x anaconda
   

2. To activate

   source activate yourenvname
   

3. pure pip deactivate

   source deactivate
   

4. To install

   conda install pkg_name
   

Conda Create Environment - Manual

1. Go to Anaconda Navigator and Create An Env
2. Go to Pycharm Setting and Search Interpreter
3. Choose Create new Interpreter
4. Select your Env
5. on Main page, choose edit config
6. select your path

Conda change the python version

1. Install python in conda

   conda install python=3.10
   

PIP Cheatsheet

1. pip list

   pip list
   

2. pip uninstall pkg_name

   pip uninstall pkg_name
   

3. pip show/list 【one specific installed package detail 】

   pip show pkg_name
   

pip create packages file

1. pip freeze > requirements.txt

   pip freeze > requirements.txt
   

2. pip install -r requirements.txt

   pip install -r requirements.txt
   

Nvidia

1. Verify num of GPU, their RAM , temp

   nvidia-smi
   

2. Verify version

   !nvcc --version
   

Git

1. Git show commits

   git log --oneline
   

2. git rebase/merge commit before push

   git reset HEAD~3
   

3. git merge after push

   git rebase -I HEAD~3
   

follow the window and replace pick with squash and leave the first pick work as it is

4. git delete all current changes before commit

   git restore file_name
   

5. Git stop file tracking

git rm -r --cached file_name

Installation from Github

1. install mscoco python api !pip install git+https://github.com/cocodataset/cocoapi.git#subdirectory=PythonAPI

Cuda

Verify if cuda is available

python -c "import torch; print(torch.cuda.is_available())" 

Resources

1. Python machine learning helper functions
2. Sample notebook for training
3. Hardware Requriments for deep Learning
4. Use Windows specific commands
5. Use this for database like mysql configuration and troubleshooting
6. Use this for server side deployment troubleshooting installation accessing
7. Use this for nlp llm technical terminologies for review purpose
8. Use this University Information
9. Django Deployment using python buildin env
10. Django Deployment using uwsgi
11. Laravel Deployment With Livewire