[GitHub] vandanavk commented on a change in pull request #12399: ONNX export: Add Crop, Deconvolution and fix the default stride of Pooling to 1

[GitBox <gi...@apache.org>]

vandanavk commented on a change in pull request #12399: ONNX export: Add Crop, Deconvolution and fix the default stride of Pooling to 1
URL: https://github.com/apache/incubator-mxnet/pull/12399#discussion_r245162200
 
 

 ##########
 File path: python/mxnet/contrib/onnx/mx2onnx/_op_translations.py
 ##########
 @@ -219,6 +219,72 @@ def convert_convolution(node, **kwargs):
     return [conv_node]
 
 
+@mx_op.register("Deconvolution")
+def convert_deconvolution(node, **kwargs):
+    """Map MXNet's deconvolution operator attributes to onnx's ConvTranspose operator
+    and return the created node.
+    """
+    name, inputs, attrs = get_inputs(node, kwargs)
+
+    kernel_dims = list(parse_helper(attrs, "kernel"))
+    stride_dims = list(parse_helper(attrs, "stride", [1, 1]))
+    pad_dims = list(parse_helper(attrs, "pad", [0, 0]))
+    num_group = int(attrs.get("num_group", 1))
+    dilations = list(parse_helper(attrs, "dilate", [1, 1]))
+    adj_dims = list(parse_helper(attrs, "adj", [0, 0]))
+
+    pad_dims = pad_dims + pad_dims
+
+    deconv_node = onnx.helper.make_node(
+        "ConvTranspose",
+        inputs=inputs,
+        outputs=[name],
+        kernel_shape=kernel_dims,
+        strides=stride_dims,
+        dilations=dilations,
+        output_padding=adj_dims,
+        pads=pad_dims,
+        group=num_group,
+        name=name
+    )
+
+    return [deconv_node]
+
+
+@mx_op.register("Crop")
+def convert_crop(node, **kwargs):
+    """Map MXNet's crop operator attributes to onnx's Crop operator
+    and return the created node.
+    """
+    name, inputs, attrs = get_inputs(node, kwargs)
+
+    num_inputs = len(inputs)
+
+    proc_nodes = kwargs["proc_nodes"]
+    input_node = proc_nodes[kwargs["index_lookup"][inputs[0][0]]].name
+
+    x, y = list(parse_helper(attrs, "offset"))
+    h, w = list(parse_helper(attrs, "h_w", [0, 0]))
+    if num_inputs > 1:
+        h, w = kwargs["out_shape"][-2:]
+    border = [x, y, x + w, y + h]
+
+    crop_node = onnx.helper.make_node(
+        "Crop",
+        inputs=[input_node],
+        outputs=[name],
+        border=border,
+        scale=[1, 1],
+        name=name
+    )
+
+    logging.warning(
+        "Using an experimental ONNX operator: Crop. " \
+        "Its definition can change.")
+
+    return [crop_node]
+
 
 Review comment:
   @ptrendx export tests usually perform an import-export-import on an ONNX model and then compare inference results.
   
   if you are planning to add an ONNX import for crop, then adding the test would be straightforward - just add a test_case in test_node.py.
   
   if you are adding export alone for crop, then in test_node.py, create a new test case list (same format as the existing maybe), add a new function test_export() in class TestNode(unittest.TestCase), export to onnx format. not sure how to test inference in this case though.
   

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services