[Example] LLM (unmasked attention) optimization using allo

[EzraReiss]

Description

These folders contain the work Juhyoung and Ezra worked on over the summer. We designed an optimized MHA kernel utilizing allo scheduling primitives.

Problems

To add the llm work we've done

Proposed Solutions

Created optimized attention kernel for FPGA architectures

Examples

Its just allo examples that can be used in the future or integrated with other transformer architecture kernels