You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
You'll install and run Google Model Explorer with Arm extensions
37
-
to inspect ExecuTorch deployment artifacts and relate graph structure to runtime data. First, you'll
38
-
set up a clean environment, clone example models with Git LFS, and launch Model Explorer with
39
-
the ExecuTorch, TOSA, and VGF adapters. Then, you'll open `.pte`, `.tosa`, `.vgf`, `.etrecord`, and `.etdp` files
40
-
to examine delegate regions, work outside delegates, and graph fragmentation. You'll
41
-
compare portable-kernel and XNNPACK PTEs for Cortex-A CPU paths and inspect Ethos-U delegation and Cortex-M targets. Finally, you'll correlate model structure with observed behavior using ETRecord and ETDump overlays that connect exported graphs with collected profiling.
36
+
You'll use pre-generated `.pte`, `.tosa`, `.vgf`, `.etrecord`, and `.etdp`
37
+
artifacts to understand how Google Model Explorer and Arm extensions expose
38
+
model export, lowering, backend delegation, and runtime behavior. After
39
+
installing the ExecuTorch extension and the TOSA and VGF adapters, you'll use
40
+
a Cortex-M PTE to learn graph navigation, compare portable and XNNPACK
41
+
Cortex-A programs, and diagnose no, clean, and fragmented Ethos-U delegation.
42
+
You'll then inspect TOSA intermediate graphs, relate TOSA fragments to
43
+
delegation boundaries, and compare a VGF-backed PTE with the standalone VGF
44
+
graph used by workflows with the ML extensions for Vulkan. Finally, you'll
45
+
open ETRecord files and load their matching ETDump data to identify whether
46
+
delegate calls or non-delegated operators dominate runtime.
42
47
faqs:
43
48
- question: Which extensions should I enable to load the artifacts used here?
0 commit comments