[GitHub] julian90 opened a new issue #10782: Multiple calls to mx.sym.Variable with same name: Dangerous / unexpected behavior?

[GitBox <gi...@apache.org>]

julian90 opened a new issue #10782: Multiple calls to mx.sym.Variable with same name: Dangerous / unexpected behavior?
URL: https://github.com/apache/incubator-mxnet/issues/10782
 
 
   ## Description
   Multiple calls to `mx.sym.Variable` with same name e.g. calling `mx.sym.Variable('a')` twice seems to result in unexpected and therefore possibly dangerous behavior in the `forward` and `backward` call.
   
   ## Environment info (Required)
   ----------Python Info----------
   Version      : 3.5.2
   Compiler     : GCC 5.4.0 20160609
   Build        : ('default', 'Nov 23 2017 16:37:01')
   Arch         : ('64bit', 'ELF')
   ------------Pip Info-----------
   Version      : 10.0.1
   Directory    : /u/schamper/software/pyenvs/py3env_seye-1.18.1/lib/python3.5/site-packages/pip
   ----------MXNet Info-----------
   Version      : 1.1.0
   Directory    : /u/schamper/software/pyenvs/py3env_seye-1.18.1/lib/python3.5/site-packages/mxnet
   Commit Hash   : 07a83a0325a3d782513a04f47d711710972cb144
   ----------System Info----------
   Platform     : Linux-4.13.0-36-generic-x86_64-with-Ubuntu-16.04-xenial
   system       : Linux
   node         : valin
   release      : 4.13.0-36-generic
   version      : #40~16.04.1-Ubuntu SMP Fri Feb 16 23:25:58 UTC 2018
   ----------Hardware Info----------
   machine      : x86_64
   processor    : x86_64
   Architecture:          x86_64
   CPU op-mode(s):        32-bit, 64-bit
   Byte Order:            Little Endian
   CPU(s):                4
   On-line CPU(s) list:   0-3
   Thread(s) per core:    1
   Core(s) per socket:    4
   Socket(s):             1
   NUMA node(s):          1
   Vendor ID:             GenuineIntel
   CPU family:            6
   Model:                 58
   Model name:            Intel(R) Core(TM) i5-3470 CPU @ 3.20GHz
   Stepping:              9
   CPU MHz:               3192.812
   CPU max MHz:           3600.0000
   CPU min MHz:           1600.0000
   BogoMIPS:              6385.62
   Virtualization:        VT-x
   L1d cache:             32K
   L1i cache:             32K
   L2 cache:              256K
   L3 cache:              6144K
   NUMA node0 CPU(s):     0-3
   Flags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm cpuid_fault epb pti retpoline tpr_shadow vnmi flexpriority ept vpid fsgsbase smep erms xsaveopt dtherm ida arat pln pts
   ----------Network Test----------
   Setting timeout: 10
   Timing for PYPI: https://pypi.python.org/pypi/pip, DNS: 0.0093 sec, LOAD: 0.3110 sec.
   Timing for MXNet: https://github.com/apache/incubator-mxnet, DNS: 0.0011 sec, LOAD: 0.8138 sec.
   Timing for Gluon Tutorial(cn): https://zh.gluon.ai, DNS: 0.9853 sec, LOAD: 0.6936 sec.
   Timing for FashionMNIST: https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/dataset/fashion-mnist/train-labels-idx1-ubyte.gz, DNS: 0.0264 sec, LOAD: 0.2201 sec.
   Timing for Gluon Tutorial(en): http://gluon.mxnet.io, DNS: 0.1136 sec, LOAD: 0.0287 sec.
   
   
   ## Minimum reproducible example
   [multiple_calls_to_sym_var.py.zip](https://github.com/apache/incubator-mxnet/files/1966650/multiple_calls_to_sym_var.py.zip)
   
   Copied content of `multiple_calls_to_sym_var.py`
   
   ```
   import mxnet as mx
   
   var_f = mx.sym.Variable('a') * mx.sym.Variable('a')
   gA = mx.nd.zeros(1)
   ex_f = var_f.bind(ctx=mx.cpu(), args={'a':mx.nd.array([3])}, args_grad={"a": gA})
   print(ex_f.forward())
   # results in 9 as expected
   
   ex_f.backward(out_grads=mx.nd.array([1]))
   print(gA)
   # results in 3 but expected gradient would be 6 (the two calls to sym.Variable
   # seem to have created two versions of 'a').
   # Up to now the program did not crash, so the wrong(?) usage of
   # mx.sym.Variable(...) is not obvious to the programmer
   
   print(ex_f.grad_dict)
   # Only after e.g. reading the grad_dict the program crashes and the wrong usage
   # becomes obvious: ValueError: Duplicate names detected, ['a', 'a']
   ```
   
   ## Steps to reproduce
   (Paste the commands you ran that produced the error.)
   
   `python3 multiple_calls_to_sym_var.py`
   
   ## Remarks
   Is this behavior intentional and useful in any scenario? Do I miss something obvious? Maybe there is a guideline, I am not aware of, that `mx.sym.Variable` should not be called twice with the same name and then be used in the same computational graph, for most applications. But if so, mxnet should probably warn the user a bit more.

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services