Kwker

Quickstarts

Each quickstart takes one workflow from install to a result you can check, in about five minutes. Every change keeps your results: exactly the same, or within a precision you choose. Pick yours:

  • QuickstartPyTorch

    Speed up a PyTorch model on the CPU without changing its code - kwker.torch_ops.install() or torch.compile(backend="kwker") - and check that the results stay the same.

  • QuickstartLanguage models

    Generate text with a Hugging Face language model on the CPU with KwkDecoder, keep model.generate() with install(), or serve an OpenAI-compatible API.

  • QuickstartEmbeddings

    Compute text embeddings and reranker scores faster on the CPU with KwkEncoder - BERT, RoBERTa, XLM-R, MPNet, ModernBERT, NomicBERT, EuroBERT and their embedders.

  • QuickstartImage models

    Classify images faster on the CPU with KwkCNN - ResNet, MobileNet, EfficientNet and similar models - in float32 or calibrated int8.

  • QuickstartDataFrames

    Group, sort and take the top rows of pandas, Polars and pyarrow tables faster with kwker.frame - you get back the same table type, with the same results.

  • QuickstartDuckDB

    Run ORDER BY ... LIMIT, sorts and GROUP BY on DuckDB results faster with kwker.duck - a DuckDB relation in, a DuckDB relation (or a pyarrow Table) out.

  • QuickstartNumPy

    Make an existing NumPy program's sort, argsort, unique and searchsorted calls run on Kwker with one line, and get NumPy's exact results.

  • QuickstartKwker Core, in any language

    Sort, select and rank arrays with Kwker Core in Python, Rust, C, C++, JavaScript, Go, Java or C#, every example runnable in your browser.

All of them start with the same install:

ShellOn your machine.
pip install kwker
python -m kwker doctor            # your CPU, the engine Kwker picked, any warnings

The PyTorch features ship in the Linux x86-64 wheel built for your torch release.

Next steps