Pretty wide list of users. Is he curious how this ML tool compares to others. OpenXLA (with already semi popular IREE Intermediate-Representation) had a pretty

Pretty wide list of users. Is he curious how this ML tool compares to others. OpenXLA (with already semi popular IREE Intermediate-Representation) had a pretty elegant launch recent; very understandable pieces/architecture. It'd be lovely to see similar explanation from TVM, to see a compare/contrast. One of the wins of OpenXLA is how retargetable it is. TVM shows both AMD & Xilinx as users, presumably using their respective hardwares. How that happens with TVM would be an interesting specific point to compare on. TVM also has IR, or IRs plural? How do they compare?