pytorch/caffe2/python/onnx/tests/onnx_backend_test.py
Will Feng cdead5ace1 Enable CircleCI for Linux jobs (#12389)
Summary:
Changes in this PR:
1. Intermediate Docker image is shared from build stage to test stage through ECR, in order to fix the Caffe2 flaky CUDA tests.
2. There are ~7 Caffe2 operator tests that are only flaky in `caffe2_py2_gcc4_8_ubuntu14_04_test` on CPU. Disabling those tests on that config only, which is okay to do because we are still running those tests in other test jobs.

After this PR is merged, CircleCI will be running on master automatically, and will be running on PRs if the author rebased their PR onto the newest master (which we will ask all the authors to do when we switch off Jenkins for Linux).
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12389

Differential Revision: D10224267

Pulled By: yf225

fbshipit-source-id: dd1a90a425c3d13b870d3d328cb301eee2e6e2cd
2018-10-08 17:09:37 -07:00

70 lines
3.1 KiB
Python

## @package onnx
# Module caffe2.python.onnx.tests.onnx_backend_test
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
from __future__ import unicode_literals
import os
import unittest
import onnx.backend.test
import caffe2.python.onnx.backend as c2
# This is a pytest magic variable to load extra plugins
pytest_plugins = 'onnx.backend.test.report',
backend_test = onnx.backend.test.BackendTest(c2, __name__)
backend_test.exclude(r'(test_hardsigmoid' # Does not support Hardsigmoid.
'|test_hardmax' # Does not support Hardmax.
'|test_cast.*FLOAT16.*' # Does not support Cast on Float16.
'|test_depthtospace.*' # Does not support DepthToSpace.
'|test_reduce_l1.*' # Does not support ReduceL1.
'|test_reduce_l2.*' # Does not support ReduceL2.
'|test_reduce_log_sum.*' # Does not support ReduceLogSum.
'|test_reduce_prod.*' # Does not support ReduceProd.
'|test_reduce_sum_square.*' # Does not support ReduceSumSquare
'|test_tile.*' # Tile's Caffe2 implementation needs some tweak
'|test_lstm.*' # Seems LSTM case has some problem
'|test_simple_rnn.*' # Seems simple RNN case has some problem
'|test_gru.*' # Seems GRU case has some problem
'|test_prelu.*' # PRelu is not compliant with ONNX yet
'|test_operator_repeat.*' # Tile is not compliant with ONNX yet
'|test_.*pool_.*same.*' # Does not support pool same.
'|test_maxpool_with_argmax.*' # MaxPool outputs indices in different format.
'|test_convtranspose.*' # ConvTranspose needs some more complicated translation
'|test_mvn.*' # MeanVarianceNormalization is experimental and not supported.
'|test_dynamic_slice.*' # MeanVarianceNormalization is experimental and not supported.
'|test_constantlike.*' # Needs implementation
'|test_eyelike.*' # Needs implementation
')')
# Quick patch to unbreak master CI, is working on the debugging.
backend_test.exclude('(test_cast_.*'
'|test_Conv1d_.*cuda'
'|test_Conv3d_groups_cuda'
'|test_rnn_seq_length'
'|test_operator_add.*_cuda'
'|test_operator_lstm_cuda'
'|test_operator_rnn.*_cuda'
'|test_lrn_default_cuda)')
# Temporarily skip some ONNX backend tests with broadcasting.
backend_test.exclude('(test_pow_bcast'
')')
# Skip vgg to speed up CI
if 'JENKINS_URL' in os.environ:
backend_test.exclude(r'(test_vgg19|test_vgg)')
# import all test cases at global scope to make them visible to python.unittest
globals().update(backend_test
.enable_report()
.test_cases)
if __name__ == '__main__':
unittest.main()