Add tvm cast fo weight of embedding op from float32 to bfloat16 and a…

…dd mlir hacks to push the compile to the end. Now embedding hangs in ttnn runtime, which is expected from tenstorrent/tt-mlir#1404
tenstorrent · Dec 10, 2024 · 19bb39e · 19bb39e
1 parent da804a9
commit 19bb39e
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 2 deletions.
diff --git a/third_party/tt-mlir b/third_party/tt-mlir
diff --git a/third_party/tvm b/third_party/tvm