Hello, Sorry, I thought this issue with GPU in new backend was solved, but I realized I was mistakenly running a test in CPU.. I use the latest git version (8b9f73365e4932f1c005a0a37b907d28985fbc5f) in local environment. In my case it seems to be a problem with Python 3, since Python 2 does not have the issue. Details below.
Thanks, Pablo ---------------------------- Name: Theano Version: 0.9.0b1 Summary: Optimizing compiler for evaluating mathematical expressions on CPUs and GPUs. Home-page: http://deeplearning.net/software/theano/ Author: LISA laboratory, University of Montreal Author-email: [email protected] License: BSD Location: /opt/libs/python3_ env/lib/python3.5/site-packagesRequires: numpy, scipy, six ---------------------------- GIT version: Theano 8b9f73365e4932f1c005a0a37b907d28985fbc5f libgpuarray 61e7ff8e4a76d78afc97d337337898560c03f247 ---------------------------- In Python 3.5.3 or 3.6.0: import os os.environ['THEANO_FLAGS'] = """ device=cuda0, floatX=float32, """ import theano Using cuDNN version 5105 on context None ERROR (theano.gpuarray): Could not initialize pygpu, support disabled Traceback (most recent call last): File "/opt/libs/python3_env/lib/python3.5/site-packages/theano/gpuarray/__init__.py", line 164, in <module> use(config.device) File "/opt/libs/python3_env/lib/python3.5/site-packages/theano/gpuarray/__init__.py", line 151, in use init_dev(device) File "/opt/libs/python3_env/lib/python3.5/site-packages/theano/gpuarray/__init__.py", line 100, in init_dev pygpu.blas.gemm(0, tmp, tmp, 0, tmp, overwrite_c=True) File "pygpu/blas.pyx", line 129, in pygpu.blas.gemm (pygpu/blas.c:3354) File "pygpu/blas.pyx", line 44, in pygpu.blas.pygpu_blas_rgemm (pygpu/blas.c:2011) pygpu.gpuarray.GpuArrayException: (b'Unsupported operation', 5) ---------------------------- In Python 2.7.13: import os os.environ['THEANO_FLAGS'] = """ device=cuda0, floatX=float32, """ import theano Using cuDNN version 5105 on context None Mapped name None to device cuda0: GeForce GTX 960M (0000:01:00.0) On Sunday, January 29, 2017 at 3:59:36 PM UTC+8, PNM wrote: > > Hello, > As for me, the issue in Python3 is already solved in git version > 3343d912717ee85c5c2e0572cfae94581b35e32b. > Thanks, > Pablo > > On Saturday, January 28, 2017 at 9:16:50 PM UTC+8, PNM wrote: >> >> Hello, >> I found the same issue with a git version in local environment. In my >> case it seems to be a problem with Python 3, since Python 2 does not have >> the issue. Details below. >> Thanks, >> Pablo >> >> ---------------------------- >> Name: Theano >> Version: 0.9.0b1 >> Summary: Optimizing compiler for evaluating mathematical expressions on >> CPUs and GPUs. >> Home-page: http://deeplearning.net/software/theano/ >> Author: LISA laboratory, University of Montreal >> Author-email: [email protected] >> License: BSD >> Location: /opt/libs_2017_01_26/python3_env/lib/python3.5/site-packages >> Requires: numpy, scipy, six >> ---------------------------- >> GIT version: >> Theano 70129ffb66320140275be7f75152e792c4647510 >> libgpuarray d838f6a43bcc56fb280a8b40fdbc03276b3eb7fd >> ---------------------------- >> In Python 3.5.3 or 3.6.0: >> >> import os >> os.environ['THEANO_FLAGS'] = """ >> device=cuda0, >> floatX=float32, >> """ >> import theano >> >> Using cuDNN version 5105 on context None >> ERROR (theano.gpuarray): Could not initialize pygpu, support disabled >> Traceback (most recent call last): >> File >> "/opt/libs_2017_01_26/python3_env/lib/python3.5/site-packages/theano/gpuarray/__init__.py", >> >> line 164, in <module> >> use(config.device) >> File >> "/opt/libs_2017_01_26/python3_env/lib/python3.5/site-packages/theano/gpuarray/__init__.py", >> >> line 151, in use >> init_dev(device) >> File >> "/opt/libs_2017_01_26/python3_env/lib/python3.5/site-packages/theano/gpuarray/__init__.py", >> >> line 100, in init_dev >> pygpu.blas.gemm(0, tmp, tmp, 0, tmp, overwrite_c=True) >> File "pygpu/blas.pyx", line 129, in pygpu.blas.gemm (pygpu/blas.c:3354) >> File "pygpu/blas.pyx", line 44, in pygpu.blas.pygpu_blas_rgemm >> (pygpu/blas.c:2011) >> pygpu.gpuarray.GpuArrayException: (b'Unsupported operation', 5) >> >> ---------------------------- >> In Python 2.7.13: >> >> import os >> os.environ['THEANO_FLAGS'] = """ >> device=cuda0, >> floatX=float32, >> """ >> import theano >> >> Using cuDNN version 5105 on context None >> Mapped name None to device cuda0: GeForce GTX 960M (0000:01:00.0) >> >> -- --- You received this message because you are subscribed to the Google Groups "theano-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
