This is the first release of luminal, a deep learning framework focused on static computation and aggressive optimization.

This release marks the first functional version. Currently examples/llama can run the full llama 1 7B and generate valid text, though it runs very slow.

Future work will speed this up significantly.