webnn: Implement MLBuffer on Core ML
Unlike compute(), the implementation of dispatch() synchronously
executes the graph with the given MLBuffers as inputs and outputs.
This papers over the lack of proper resource sharing (as seen in the
TFLite backend), which is left to follow-up CLs
Bug: 333392274
Change-Id: I3aaddc9fefa26b26b7c1927a139ceefd829b74e7
Cq-Include-Trybots: luci.chromium.try:mac14-blink-rel,mac14.arm64-blink-rel
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5750342
Reviewed-by: Phillis Tang phillis@chromium.org
Reviewed-by: Reilly Grant reillyg@chromium.org
Commit-Queue: Austin Sullivan asully@chromium.org
Cr-Commit-Position: refs/heads/main@{#1336289}