lhutton1 commented on code in PR #10915:
URL: https://github.com/apache/tvm/pull/10915#discussion_r875634397
##########
tests/python/frontend/tflite/test_forward.py:
##########
@@ -909,6 +939,24 @@ def representative_data_gen():
)
+def test_forward_quantized_convolution():
Review Comment:
Interesting there were no tests using `_test_tflite2_quantized_convolution`
previously... probably a good idea to make use of
`_test_tflite2_quantized_depthwise_convolution` in the future as well
##########
tests/python/frontend/tflite/test_forward.py:
##########
@@ -884,23 +912,25 @@ def _test_tflite2_quantized_convolution(
data_in = tf.keras.layers.Input(shape=data.shape[1:])
conv = tf.keras.layers.Conv2D(
- filters=kernel_shape[3],
+ filters=filters,
kernel_size=(kernel_shape[0], kernel_shape[1]),
- strides=strides,
+ activation=tf.nn.relu,
padding=padding,
- data_format=data_format,
Review Comment:
looks like `data_format` should still be an argument, the tests didn't pick
this up so we should add a test for "NHWC" as well
##########
tests/python/frontend/tflite/test_forward.py:
##########
@@ -884,23 +912,25 @@ def _test_tflite2_quantized_convolution(
data_in = tf.keras.layers.Input(shape=data.shape[1:])
conv = tf.keras.layers.Conv2D(
- filters=kernel_shape[3],
+ filters=filters,
kernel_size=(kernel_shape[0], kernel_shape[1]),
- strides=strides,
+ activation=tf.nn.relu,
padding=padding,
- data_format=data_format,
- activation="relu",
Review Comment:
any reason to remove the activation testing?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
