Find your next starting point.

Model-provider interface and proxy tooling for inspecting request normalization, usage accounting and backend differences.

C/C++ language-model inference implementation for studying quantization and execution tradeoffs across supported hardware.

Experiment and model lifecycle tooling for preserving parameters, artifacts and evaluation lineage across runs.

Additional iterator operations for grouping, windowing, combinatorial enumeration, and stream processing.

Conversion of Jupyter notebooks to other document formats, with optional execution workflows.

Reading, writing, and validating Jupyter notebook document structures.

Graph construction, traversal, structural analysis, and network algorithms in Python.

Neural-network inspection and intervention interfaces for tracing internal activations and testing counterfactual computations.

Just-in-time compilation of supported Python numerical functions using LLVM.

Language-model research training repository for studying configuration transparency and reproducibility of model-building workflows.

Gradient transformation and optimization tools for testing optimizer state and update equations.

Model optimization integrations for investigating the accuracy and latency consequences of a chosen deployment transformation.