Getting Started¶

Lab 01 · Miniconda Installation & Environment Setup walks through installing Miniconda from scratch. If you already have it, this is the whole setup.

Create the environment¶

conda create -n envNeural python=3.12 -y

conda activate envNeural && pip install torch torchvision --index-url https://download.pytorch.org/whl/cpu && pip install numpy matplotlib scikit-learn scipy pillow jupyter ipykernel

Extras for Labs 17-19¶

Lab 17's deployment section needs onnx, onnxruntime and onnxscript; Labs 18-19's retrieval labs need transformers:

conda activate envNeural && pip install onnx onnxruntime onnxscript transformers

CPU vs. GPU¶

Every lab runs on CPU. Lab 12 · Transfer Learning and Lab 16 · GANs are noticeably faster on a GPU — if you have one, install the CUDA build of PyTorch instead of the whl/cpu index above.

Datasets and pretrained weights download automatically on first run, into ./data and the Hugging Face cache respectively.

Running the labs¶

Clone the repository and open any notebook with Jupyter:

git clone https://github.com/IinjyI/Hands-On-Neural-Networks.git
cd Hands-On-Neural-Networks
conda activate envNeural
jupyter lab

Every notebook on this site already carries its own output, so you can read a lab end to end without running it first — running it yourself is where the experiment sections at the end of each lab come in.