Entry point: OK Transformer website
A collection of self-contained notebooks on machine learning theory, engineering, and operations. I try to cover topics that frequently come up as building blocks for applications or further theory. I also explore areas where I want to clarify my understanding or delve into details that I personally find interesting or intriguing.
The notebooks should ideally run end-to-end with reproducible results between runs. Exact output values may change due to external dependencies such as hardware and changing dataset versions, but the conclusions should still generally hold. Please open an issue if you find that this is not the case (as I often do)!
git clone git@github.com:particle1331/ok-transformer.git && cd ok-transformer
pip install --user tox
tox -e build
The notebooks can be found in docs/nb.
A virtual environment for running the notebooks can be
created using pdm
(use this as jupyter kernel):
pip install -U pdm
pdm install
The following libraries (specified in pdm.lock) will be installed:
╭────────────────────────────────────┬────────────────╮
│ fastapi │ 0.111.0 │
│ Flask │ 3.0.3 │
│ keras │ 2.15.0 │
│ lightning │ 2.3.0 │
│ matplotlib │ 3.9.0 │
│ mlflow │ 2.13.2 │
│ numpy │ 1.26.4 │
│ optuna │ 3.6.1 │
│ pandas │ 2.2.2 │
│ scikit-learn │ 1.5.0 │
│ scipy │ 1.13.1 │
│ seaborn │ 0.13.2 │
│ tensorflow │ 2.15.1 │
│ tensorflow-datasets │ 4.9.6 │
│ tensorflow-estimator │ 2.15.0 │
│ torch │ 2.2.2 │
│ torchaudio │ 2.2.2 │
│ torchinfo │ 1.8.0 │
│ torchmetrics │ 1.4.0.post0 │
│ torchvision │ 0.17.2 │
│ uvicorn │ 0.30.1 │
│ xgboost │ 2.0.3 │
╰────────────────────────────────────┴────────────────╯
GPU 0: Tesla P100-PCIE-16GB
CPU: Intel(R) Xeon(R) CPU @ 2.00GHz
Core: 1
Threads per core: 2
L3 cache: 38.5 MiB
Memory: 15 Gb