Skip to content
This repository has been archived by the owner on Nov 17, 2023. It is now read-only.

[numpy] Fix test failures after using new shape semantics for zero-dim/size tensors #14454

Closed
reminisce opened this issue Mar 18, 2019 · 7 comments

Comments

@reminisce
Copy link
Contributor

reminisce commented Mar 18, 2019

Branch: /~https://github.com/apache/incubator-mxnet/tree/numpy

CPU withou MKLDNN

Folder: tests/python/unittest

Errors/Failures (CPU) Assignee Status
test_contrib_control_flow.test_contrib_rnn @zheng-da Fixed by #14520
test_dynamic_shape.test_dynamic_shape @junrushao1994 Fixed by #14538
test_gluon_contrib.test_convrnn Resolved
test_gluon_contrib.test_convlstm Resolved
test_gluon_contrib.test_convgru Resolved
test_gluon_contrib.test_conv_fill_shape Resolved
test_gluon_contrib.test_lstmp Resolved
test_gluon_contrib.test_vardrop Resolved
test_gluon_contrib.test_contrib_unroll @junrushao1994 Resolved
test_gluon_data.test_recordimage_dataset @szha Resolved
test_gluon_model_zoo.test_models Resolved
test_gluon.test_deconv @reminisce Resolved
test_gluon.test_symbol_block Resolved
test_gluon.test_deferred_init Resolved
test_gluon.test_hybrid_stale_cache Resolved
test_gluon.test_fill_shape_deferred Resolved
test_gluon.test_dtype Resolved
test_gluon.test_fill_shape_load Resolved
test_gluon.test_inline Resolved
test_gluon.test_req Resolved
test_gluon.test_save_load Resolved
test_gluon.test_symbol_block_save_load Resolved
test_gluon.test_hybrid_multi_context Resolved
test_gluon.test_hook Resolved
test_gluon.test_summary Resolved
test_gluon.test_legacy_save_params Resolved
test_gluon.test_conv2d_16c Resolved
test_gluon.test_group_conv2d_16c Resolved
test_gluon.test_concat Resolved
test_gluon.test_reshape_conv Resolved
test_gluon.test_slice_conv Resolved
test_gluon.test_slice_conv_slice_conv Resolved
test_gluon.test_reshape_dense Resolved
test_gluon.test_slice_dense Resolved
test_gluon.test_slice_dense_slice_dense Resolved
test_gluon.test_reshape_dense_reshape_dense Resolved
test_gluon.test_slice_dense_reshape_dense Resolved
test_gluon.test_reshape_dense_slice_dense Resolved
test_gluon.test_slice_batchnorm Resolved
test_gluon.test_slice_batchnorm_reshape_batchnorm Resolved
test_gluon_rnn.test_rnn Resolved
test_gluon_rnn.test_lstm Resolved
test_gluon_rnn.test_lstm_forget_bias Resolved
test_gluon_rnn.test_lstm_cpu_inference Resolved
test_gluon_rnn.test_gru Resolved
test_gluon_rnn.test_residual Resolved
test_gluon_rnn.test_residual_bidirectional Resolved
test_gluon_rnn.test_stack Resolved
test_gluon_rnn.test_hybridstack Resolved
test_gluon_rnn.test_bidirectional Resolved
test_gluon_rnn.test_layer_bidirectional Resolved
test_gluon_rnn.test_zoneout Resolved
test_gluon_rnn.test_unroll_layout Resolved
test_gluon_rnn.test_rnn_cells_export_import Resolved
test_gluon_rnn.test_rnn_layers_fp32 Resolved
test_gluon_rnn.test_rnn_unroll_variant_length Resolved
test_gluon_rnn.test_cell_fill_shape Resolved
test_gluon_rnn.test_layer_fill_shape Resolved
test_gluon_rnn.test_bidirectional_unroll_valid_length Resolved
test_gluon_trainer.test_trainer_multi_layer_init Resolved
test_infer_shape.test_incomplete_infer_elewise @reminisce Resolved
test_infer_shape.test_incomplete_infer_mlp @reminisce Resolved
test_infer_shape.test_incomplete_infer_slicechannel @reminisce Resolved
test_infer_shape.test_incomplete_infer_convolution @reminisce Resolved
test_infer_shape.test_incomplete_infer_concat @reminisce Resolved
test_init.test_default_init @reminisce Resolved
test_init.test_variable_init @reminisce Resolved
test_io.test_NDArrayIter_csr @eric-haibin-lin Resolved
test_io.test_LibSVMIter @eric-haibin-lin Fixed locally
test_ndarray.test_ndarray_setitem @reminisce Resolved
test_module.test_module_states Resolved
test_module.test_module_switch_bucket Resolved
test_module.test_executor_group Resolved
test_module.test_factorization_machine_module @eric-haibin-lin Fixed by sparse slice
test_operator.test_lstm_sym @reminisce Resolved
test_operator.test_lstm_bidirectional @reminisce Resolved
test_operator.test_gru_sym @reminisce Resolved
test_operator.test_gru_bidirectional @reminisce Resolved
test_operator.test_rnntanh_sym @reminisce Resolved
test_operator.test_rnntanh_bidirectional @reminisce Resolved
test_operator.test_rnnrelu_sym @reminisce Resolved
test_operator.test_rnnrelu_bidirectional @reminisce Resolved
test_operator.test_boolean_mask @reminisce Resolved
test_operator.test_custom_op @reminisce Resolved
test_operator.test_multi_proposal_op @reminisce Resolved
test_operator.test_diag @reminisce Resolved
test_predictor.test_predictor Resolved
test_rnn.test_deprecated Resolved
test_rnn.test_rnn Resolved
test_rnn.test_lstm Resolved
test_rnn.test_lstm_forget_bias Resolved
test_rnn.test_gru Resolved
test_rnn.test_residual Resolved
test_rnn.test_residual_bidirectional Resolved
test_rnn.test_stack Resolved
test_rnn.test_bidirectional Resolved
test_rnn.test_zoneout Resolved
test_rnn.test_unfuse Resolved
test_rnn.test_convrnn Resolved
test_rnn.test_convlstm Resolved
test_rnn.test_convgru Resolved
test_sparse_ndarray.test_sparse_nd_slice @reminisce Resolved
test_sparse_operator.test_sparse_slice @eric-haibin-lin Resolved
test_subgraph_op.test_subgraph_exe @reminisce Resolved
test_subgraph.test_make_subgraph Resolved
test_symbol.test_symbol_infer_shape Resolved
test_thread_local.test_createblock Resolved
test_kvstore.test_single_kv_pair @eric-haibin-lin Resolved

Folder: tests/python/train

Errors/Failures (CPU) Assignee Status
test_autograd.test_autograd Resolved
test_bucketing.test_bucket_module Resolved
test_sparse_fm.test_factorization_machine_module @eric-haibin-lin Fixed by sparse slice

Folder: tests/python/gpu

GPU

Errors/Failures (GPU) Assignee Status
test_operator_gpu.test_pooling_versions @reminisce Fixed by #14577
test_operator_gpu.test_global_pooling @reminisce Fixed by #14577
test_operator_gpu.test_upsampling_bilinear_with_type @junrushao1994 Fixed by #14557
test_operator_gpu.test_embedding_with_type @eric-haibin-lin Cannot reproduce with 10K runs
test_operator_gpu.test_deconvolution @szha Cannot reproduce
test_operator_gpu.test_deformable_convolution_with_type @zheng-da Fixed by #14577
test_operator_gpu.test_deformable_convolution_options @zheng-da Fixed by #14577

CPU with MKLDNN

Folder: tests/python/unittest

Errors/Failures (CPU) Assignee Status
test_gluon_rnn.test_rnn_cells @lihaofd Resolved
test_io.test_LibSVMIter May have been fixed on master
Test factorization machine model with sparse operators Resolved
test_operator.test_lstm_bidirectional @lihaofd Resolved
test_operator.test_gru_bidirectional @lihaofd Resolved
test_operator.test_rnntanh_bidirectional @lihaofd Resolved
test_operator.test_rnnrelu_bidirectional @lihaofd Resolved
test_operator.test_concat @TaoLv Resolved
test_operator.test_slice @TaoLv Resolved
test operators which don't implement FComputeEx or FStatefulComputeEx @ZhennanQin Resolved

Folder: tests/python/train

Errors/Failures (CPU) Assignee Status
Test factorization machine model with sparse operators Resolved

Folder: tests/python/mkl

Errors/Failures (CPU) Assignee Status
test_subgraph.test_pos_single_concat Resolved

CPP TEST

[----------] 2 tests from ACTIVATION_PERF
[ RUN      ] ACTIVATION_PERF.ExecuteBidirectional
Segmentation fault (core dumped)
tests/cpp/unittest.mk:68: recipe for target 'runtest' failed
make: *** [runtest] Error 139

train.log
unittest.log
gpu.log
unittest_mkldnn.log

@mxnet-label-bot
Copy link
Contributor

Hey, this is the MXNet Label Bot.
Thank you for submitting the issue! I will try and suggest some labels so that the appropriate MXNet community members can help resolve it.
Here are my recommended labels: Test

@junrushao
Copy link
Member

Cannot reproduce test_gluon_contrib.test_contrib_unroll now, did I miss anything?

@reminisce
Copy link
Contributor Author

I took a look at unittest.log. test_gluon_contrib.test_contrib_unroll should have been fixed by
/~https://github.com/apache/incubator-mxnet/pull/14520/files#diff-18a3d5a32b024eb7f4d6560001898c4cL755
Thanks for working on it.

@pengzhao-intel
Copy link
Contributor

@TaoLv will help taking look the MKLDNN related issues in the list :)

@pengzhao-intel
Copy link
Contributor

cc @juliusshufan

@TaoLv
Copy link
Member

TaoLv commented Mar 28, 2019

@reminisce Do you have steps for reproducing those failures on local machine?

@reminisce
Copy link
Contributor Author

@TaoLv Let's talk offline today.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Projects
None yet
Development

No branches or pull requests

5 participants