Fixing ONNX test #3253
Fixing ONNX test #3253
Conversation
Score column being named "Score0". The ONNX model will rename the output columns by design, therefore a different class with the ColumnName of "Score0" is needed. This fixes the test pipeline to address this issue. Fixes dotnet#2981
Codecov Report
@@ Coverage Diff @@
## master #3253 +/- ##
==========================================
+ Coverage 72.69% 72.69% +<.01%
==========================================
Files 807 807
Lines 145172 145171 -1
Branches 16225 16225
==========================================
Hits 105537 105537
+ Misses 35220 35219 -1
Partials 4415 4415
|
occur if the input type was not a variable vector or vector type. This is not needed as we do support converting basic types to equivalent ONNX tensor type. Therefore removing the check in GetOutputSchema fixes this problem.
| @@ -564,8 +564,6 @@ public override SchemaShape GetOutputSchema(SchemaShape inputSchema) | |||
| var input = Transformer.Inputs[i]; | |||
| if (!inputSchema.TryFindColumn(input, out var col)) | |||
| throw Host.ExceptSchemaMismatch(nameof(inputSchema), "input", input); | |||
| if (!(col.Kind == SchemaShape.Column.VectorKind.VariableVector || col.Kind == SchemaShape.Column.VectorKind.Vector)) | |||
| throw Host.ExceptSchemaMismatch(nameof(inputSchema), "input", input, "vector", col.GetTypeString()); | |||
|
|
|||
There was a problem hiding this comment.
if input col.Kind == VariableVector, how is that case handled? If not handled, should throw exception.
There was a problem hiding this comment.
so prior to this change, we allowed VariableVector in that it would not throw -- this change now allows for vector and simple types. I would prefer to have a test that uses VariableVector, but I have yet to find something that provides a VariableVector output AND can be saved to ONNX.
There was a problem hiding this comment.
@jignparm from our conversation, you mentioned that we should not have allowed VariableVector to begin with and should fix as part of this change. I have updated the code to reflect that, but we should have a test that confirms if this works correctly. I have created the following issue for tracking #3375
|
LGTM @singlis |
ghost
locked as resolved and limited conversation to collaborators
The test pipeline for consuming an ONNX model would fail due to the Score column being named "Score0". The ONNX model will rename the output columns by design, therefore a different class with the ColumnName of "Score0" is needed. This fixes the test pipeline to address this issue.
Fixes #2981