[tvm] branch main updated: [Bugfix][Relay][Keras] Fix UpSampling2D about the wrong assertion about size (#15082)

[echuraev]

This is an automated email from the ASF dual-hosted git repository.

echuraev 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 081cc2ef64 [Bugfix][Relay][Keras] Fix UpSampling2D about the wrong 
assertion about size (#15082)
081cc2ef64 is described below

commit 081cc2ef64c866f6dc2e33dbf6912395d09ca749
Author: Qingchao Shen <qingchaos...@outlook.com>
AuthorDate: Wed Jun 14 17:13:19 2023 +0800

    [Bugfix][Relay][Keras] Fix UpSampling2D about the wrong assertion about 
size (#15082)
    
    * fix wrong assertion about unsample in keras.py
    
    * Update test_forward.py
    
    * Update test_forward.py
---
 python/tvm/relay/frontend/keras.py          | 4 +---
 tests/python/frontend/keras/test_forward.py | 5 +++++
 2 files changed, 6 insertions(+), 3 deletions(-)

diff --git a/python/tvm/relay/frontend/keras.py 
b/python/tvm/relay/frontend/keras.py
index 16192617fe..d963a5d160 100644
--- a/python/tvm/relay/frontend/keras.py
+++ b/python/tvm/relay/frontend/keras.py
@@ -767,10 +767,8 @@ def _convert_upsample(
         params["scale_h"] = h
     elif upsample_type == "UpSampling2D":
         h, w = keras_layer.size
-        if h != w:
-            raise tvm.error.OpAttributeInvalid("Height must equal width for 
operator Upsample.")
         params["scale_h"] = h
-        params["scale_w"] = h
+        params["scale_w"] = w
 
         if hasattr(keras_layer, "interpolation"):
             interpolation = keras_layer.interpolation
diff --git a/tests/python/frontend/keras/test_forward.py 
b/tests/python/frontend/keras/test_forward.py
index debd50b37a..45935f87f4 100644
--- a/tests/python/frontend/keras/test_forward.py
+++ b/tests/python/frontend/keras/test_forward.py
@@ -389,6 +389,11 @@ class TestKeras:
         x = keras_mod.layers.UpSampling2D(size=(3, 3), 
interpolation=interpolation)(data)
         keras_model = keras_mod.models.Model(data, x)
         verify_keras_frontend(keras_model)
+        # Height and width are not equal for the attribute size
+        data = keras_mod.layers.Input(shape=(2, 1, 3))
+        x = keras_mod.layers.UpSampling2D(size=(1, 2), 
interpolation=interpolation)(data)
+        keras_model = keras_mod.models.Model(data, x)
+        verify_keras_frontend(keras_model)
 
     def test_forward_reshape(self, keras_mod):
         """test_forward_reshape"""