Step_By_Step_Pytorch_Tutorials

Set Up Enviroment and Install Required Packages

1. Install miniconda

2. Create and activate a conda enviroment

   conda create -n pytorch_env && source activate pytorch_env
   

3. Install Pytorch and verify installation success

   conda install pytorch torchvision -c pytorch
   python -c 'import torch; print(torch.rand(2,3));print(torch.cuda.is_available())'
   

4. Install tensorboard and pytorch tensorboard utils (optional)

   pip install tensorboard
   pip install tb-nightly
   pip install -r requirements.txt
   

   Verify the pytorch tensorboard python tensorboard_test_script.py Run tensorboard --logdir=runs and type http://localhost:6006/ in your browser to access the Tensorboard

5. Install PyTorchViz (optional)

   pip install torchviz
   

6. Install Jupyter and Jupyter conda extension

   pip install jupyter
   conda install nb_conda
   

   Run jupyter notebook to activate conda enviroment and assess the Jupyter notebook

7. Clone the tutorial repository

git clone https://github.com/Shaoli-Huang/Step_by_Step_Pytorch_Tutorials.git