I am writing following code:
######################################################
from theano.gpuarray.dnn import dnn_conv
lX = T.tensor4()
def convLayer(X, w1, w2, w3, w4):
h1 = dnn_conv(img=X, kerns=w1, subsample=(2, 2), border_mode=(1, 1))
h2 = dnn_conv(img=h1, kerns=w2, subsample=(2, 2), border_mode=(1, 1))
h3 = dnn_conv(img=h2, kerns=w3, subsample=(2, 2), border_mode=(1, 1))
h4 = dnn_conv(img=h3, kerns=w4, subsample=(2, 2), border_mode=(1, 1))
h4 = T.flatten(h4, 2)
return h4
lX_conv = convLayer(lX, w1, w2, w3, w4)
######################################################
I get an error like:
Traceback (most recent call last):
File "<input>", line 1, in <module>
File "<input>", line 2, in convLayer
File
"/home/xi/Projects/Python/env/local/lib/python2.7/site-packages/theano/gpuarray/dnn.py",
line 970, in dnn_conv
ctx_name = infer_context_name(img, kerns)
File
"/home/xi/Projects/Python/env/local/lib/python2.7/site-packages/theano/gpuarray/basic_ops.py",
line 122, in infer_context_name
raise ValueError("Could not infer context from inputs")
ValueError: Could not infer context from inputs
I have tested this code on two configuration:
1. theano 0.9.dev3 + cuda 8 + cudnn 5.1
2. theano.9.0.dev1 + cuda 7.5 + cudnn 5.0
In the past I am import cudnn from *theano.sandbox**.cuda.dnn*, everything runs
good, however, not sure what wrong is when I try to import cudnn from
*theano.gpuarray.dnn*
Can anyone helps or give some suggestion?
--
---
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.
