stephenrawls commented on issue #14208: Add support for fast variable-length LSTM URL: https://github.com/apache/incubator-mxnet/pull/14208#issuecomment-487345467 @szha finally got around to rebasing my PR and pushing the update. The centos-gpu build is failing. The new var-length unit test is failing with this error: ``` ERROR: test_gluon_gpu.test_layer_bidirectional_varseqlength ---------------------------------------------------------------------- Traceback (most recent call last): File "/usr/lib/python3.6/site-packages/nose/case.py", line 198, in runTest self.test(*self.arg) File "/work/mxnet/tests/python/gpu/../unittest/common.py", line 177, in test_new orig_test(*args, **kwargs) File "/work/mxnet/tests/python/gpu/../unittest/common.py", line 112, in test_new assert_raises((MXNetError, RuntimeError), orig_test, *args, **kwargs) File "/usr/lib64/python3.6/unittest/case.py", line 750, in assertRaises return context.handle('assertRaises', args, kwargs) File "/usr/lib64/python3.6/unittest/case.py", line 195, in handle callable_obj(*args, **kwargs) File "/work/mxnet/tests/python/gpu/test_gluon_gpu.py", line 290, in test_layer_bidirectional_varseqlength check_layer_bidirectional_varseqlen(7, 5) File "/work/mxnet/tests/python/gpu/test_gluon_gpu.py", line 267, in check_layer_bidirectional_varseqlen net_output = net(data, sequence_length).asnumpy() File "/work/mxnet/python/mxnet/gluon/block.py", line 540, in __call__ out = self.forward(*args) File "/work/mxnet/python/mxnet/gluon/block.py", line 912, in forward self._deferred_infer_shape(x, *args) File "/work/mxnet/python/mxnet/gluon/block.py", line 793, in _deferred_infer_shape raise ValueError(error_msg) ValueError: Deferred initialization failed because shape cannot be inferred. Error in operator lstm_rnn0: Shape inconsistent, Provided = [10], inferred shape=(2,10,7) -------------------- >> begin captured stdout << --------------------- infer_shape error. Arguments: data0: (11, 10, 5) data1: (10,) ``` I'm not sure why yet, because when I run the unit test code on my own machine with my own build it works. Also the unix-gpu build is failing for the unit test `test_operator.test_custom_op_exc` which seems unrelated to my PR. I'll see if I can figure out why the centos-gpu build is failing (even though it works for me on my own machine) tomorrow. But if you have any ideas / advice please let me know.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
