Source Code and Data

Read the scripts in your browser, or download the full practice bundle to run locally.

Download practice bundle ZIP

The bundle contains the tutorial, Python code, tests, requirement lists and candidate data; no model weights, training outputs or local environments. The data still needs review.

Download the GPU training notebook · How to use it

The notebook reads data/dataset.json from the bundle directly and must be used together with the project code. The local GPU upload bundle from chapter 3 also contains the tokenizer; the general source bundle above contains no model cache.

File	Read
book_review.sh	Open
data/dataset.json	Open
lab/__init__.py	Open
lab/audit.py	Open
lab/common.py	Open
lab/data.py	Open
lab/download-lock.json	Open
lab/download.py	Open
lab/evaluate.py	Open
lab/infer.py	Open
lab/merge.py	Open
lab/prepare.py	Open
lab/train.py	Open
lab/training.py	Open
pyproject.toml	Open
requirements-book-lock.txt	Open
requirements-book.txt	Open
requirements-cpu.txt	Open
requirements-data.txt	Open
requirements-download.txt	Open
requirements-notebook.txt	Open
requirements-train.txt	Open
tests/test_artifacts.py	Open
tests/test_book.py	Open
tests/test_bundle.py	Open
tests/test_data.py	Open
tests/test_download.py	Open
tests/test_eval.py	Open
tests/test_gpu_zip.py	Open
tests/test_notebook.py	Open
tests/test_prepare.py	Open
tests/test_render.py	Open
tests/test_training.py	Open
tests/test_tutorial.py	Open
tools/__init__.py	Open
tools/build_gpu_zip.py	Open