0.2

What's Changed

Everything.

And also:

• Simple Pooling implementation by @TheSeamau5 in #4
• Conv1D module by @TheSeamau5 in #7
• Conv2d by @TheSeamau5 in #8
• Add support for Mistral by @TheSeamau5 in #9
• Shell script to download mistral that actually works by @TheSeamau5 in #10
• Fix swish by @TheSeamau5 in #11
• Minor improvement to f16 matmul, Longer prompt and token generation for testing by @TheSeamau5 in #12
• Small Improvements to main by @TheSeamau5 in #17
• Q8 Weight Quantization by @jafioti in #24

Full Changelog: 0.1...0.2

0.1

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.