Closed disktnk closed 5 years ago
Merging #210 into master will increase coverage by
0.28%
. The diff coverage is94.64%
.
@@ Coverage Diff @@
## master #210 +/- ##
==========================================
+ Coverage 89.87% 90.15% +0.28%
==========================================
Files 24 24
Lines 1471 1513 +42
==========================================
+ Hits 1322 1364 +42
Misses 149 149
Impacted Files | Coverage Δ | |
---|---|---|
onnx_chainer/export.py | 92.83% <94.64%> (+1.35%) |
:arrow_up: |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update a75c88b...8cb854f. Read the comment docs.
/test
Successfully created a job for commit 8cb854f:
fixes #187
Example graph when set
input_shapes=('b',)+x.shape[1:]
[NOTE1]
Updated shape information in
value_info
are set byonnx.shape_inference
module. There are some unsupported node by the module, like Concat. then set empty (Netron show0
when shape is empty)Some runtime (ONNXRuntime, MXNet) could run correctly even if set
0
[NOTE2]
Currently the below code is not supported, shape value of reshaped is fixed and not set as dynamic parameter.