This is an automated email from the ASF dual-hosted git repository.
lukhut pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/tvm.git
The following commit(s) were added to refs/heads/main by this push:
new d83cd217a5 [microNPU][ETHOSU] Fix ConcatRewriter args processing
(#16003)
d83cd217a5 is described below
commit d83cd217a59e0dfbb83012138de0b29d2c66fafa
Author: Aleksei-grovety <113356454+aleksei-grov...@users.noreply.github.com>
AuthorDate: Tue Oct 31 17:15:47 2023 +0400
[microNPU][ETHOSU] Fix ConcatRewriter args processing (#16003)
In ConcatRewriter the case was not considered when the concatenation
argument is TupleGetItem.
---
python/tvm/relay/backend/contrib/ethosu/legalize.py | 2 +-
tests/python/contrib/test_ethosu/test_codegen.py | 16 ++++++++++++++++
2 files changed, 17 insertions(+), 1 deletion(-)
diff --git a/python/tvm/relay/backend/contrib/ethosu/legalize.py
b/python/tvm/relay/backend/contrib/ethosu/legalize.py
index 2806ef8a46..7ed69e1e9b 100644
--- a/python/tvm/relay/backend/contrib/ethosu/legalize.py
+++ b/python/tvm/relay/backend/contrib/ethosu/legalize.py
@@ -1184,7 +1184,7 @@ class ConcatRewriter(DFPatternCallback):
# Find the tensors that are inputs to the concat and the scales and
zero points
concat_args = list()
for arg in post.args:
- if isinstance(arg, tvm.relay.expr.Call):
+ if isinstance(arg, (tvm.relay.expr.Call,
tvm.relay.expr.TupleGetItem)):
concat_args.append(arg)
axis = post.op.body.attrs.axis
diff --git a/tests/python/contrib/test_ethosu/test_codegen.py
b/tests/python/contrib/test_ethosu/test_codegen.py
index 66809a775f..f69c114cab 100644
--- a/tests/python/contrib/test_ethosu/test_codegen.py
+++ b/tests/python/contrib/test_ethosu/test_codegen.py
@@ -1170,6 +1170,22 @@ def test_tflite_concat(shapes, axis, accel_type):
infra.compare_tvm_with_tflite(concat_func, shapes, accel_type,
enable_cascader=False)
+def test_tflite_unstack_concat():
+ np.random.seed(0)
+ shapes = [(2, 4, 16)]
+ axis = 1
+ accel_type = "ethos-u55-256"
+
+ @tf.function
+ def concat_func(input):
+ inputs = tf.unstack(input)
+ inputs.reverse()
+ op = tf.concat(inputs, axis)
+ return op
+
+ infra.compare_tvm_with_tflite(concat_func, shapes, accel_type,
enable_cascader=False)
+
+
def test_tflite_concat_with_reused_args():
np.random.seed(0)
shapes = [(1, 1, 24, 1), (1, 1, 24, 1), (1, 1, 10, 1), (1, 1, 68, 1)]
