MLBEDSW-8240: Document reference comparison point
- Updated TensorFlow Support section
Change-Id: Ic2551f44e7dfa996a5dcc8840d480b7985415a0a
Signed-off-by: Tim Hall <tim.hall@arm.com>
diff --git a/README.md b/README.md
index 4b29bad..35c8d3a 100644
--- a/README.md
+++ b/README.md
@@ -45,9 +45,11 @@
(EXPERIMENTAL).
## TensorFlow Support
-Vela is tested by comparing the numerical behaviour of the optimised operators
-against that of the corresponding TensorFlow Lite reference kernels. The
-following list indicates which version is used for comparison:
+Vela is tested by comparing the bit exact numerical behaviour of the Ethos-U
+optimised operators against that of the corresponding TensorFlow Lite reference
+kernels (or TensorFlow Lite for Microcontrollers reference kernels in the case
+of the UNIDIRECTIONAL_SEQUENCE_LSTM operator). The following list indicates
+which version is used for comparison:
* Vela 3.10.0 to current supports TensorFlow 2.14
* Vela 3.9.0 supports TensorFlow 2.12