Hi Fred, 

Yes, it was fixed, but when upgrading to the new Theano-0.9.0.dev4 the 
problem arises again. Turning back to Theano-0.9.0.dev4 solves the problem:



  File 
"/home/samtzai/anaconda2/lib/python2.7/site-packages/pymc3/theanof.py", 
line 60, in gradient
    return tt.concatenate([gradient1(f, v) for v in vars], axis=0)
  File 
"/home/samtzai/anaconda2/lib/python2.7/site-packages/pymc3/theanof.py", 
line 48, in gradient1
    return tt.flatten(tt.grad(f, v, disconnected_inputs='warn'))
  File 
"/home/samtzai/anaconda2/lib/python2.7/site-packages/theano/gradient.py", 
line 553, in grad
    grad_dict, wrt, cost_name)
  File 
"/home/samtzai/anaconda2/lib/python2.7/site-packages/theano/gradient.py", 
line 1317, in _populate_grad_dict
    rval = [access_grad_cache(elem) for elem in wrt]
  File 
"/home/samtzai/anaconda2/lib/python2.7/site-packages/theano/gradient.py", 
line 1272, in access_grad_cache
    term = access_term_cache(node)[idx]
  File 
"/home/samtzai/anaconda2/lib/python2.7/site-packages/theano/gradient.py", 
line 965, in access_term_cache
    output_grads = [access_grad_cache(var) for var in node.outputs]
  File 
"/home/samtzai/anaconda2/lib/python2.7/site-packages/theano/gradient.py", 
line 1272, in access_grad_cache
    term = access_term_cache(node)[idx]
  File 
"/home/samtzai/anaconda2/lib/python2.7/site-packages/theano/gradient.py", 
line 965, in access_term_cache
    output_grads = [access_grad_cache(var) for var in node.outputs]
  File 
"/home/samtzai/anaconda2/lib/python2.7/site-packages/theano/gradient.py", 
line 1272, in access_grad_cache
    term = access_term_cache(node)[idx]
  File 
"/home/samtzai/anaconda2/lib/python2.7/site-packages/theano/gradient.py", 
line 1189, in access_term_cache
    i, term.type.dtype))
TypeError: Elemwise{mul,no_inplace}.grad illegally  returned an 
integer-valued variable. (Input index 1, dtype bool)

Regards








El lunes, 17 de noviembre de 2014, 21:58:39 (UTC+1), nouiz escribió:
>
> Hi,
>
> My guess is that this is fixed in the development version of Theano. Just 
> update Theano as done here:
>
>
> http://www.deeplearning.net/software/theano/install.html#bleeding-edge-install-instructions
>
> If that do not fix it, tell me. But from memory, this was fixed.
>
> Fred
>
> On Mon, Nov 17, 2014 at 2:32 PM, Ashutosh Modi <[email protected] 
> <javascript:>> wrote:
>
>> Hi,
>>
>> I have a complicated architecture and I am calculating gradients wrt to 
>> diff parameters, but I am getting following strange error, what is this 
>> indicative of? I am not getting any clue why is this error popping up?
>>
>> *TypeError: Join.grad illegally  returned an integer-valued variable. 
>> (Input index 1, dtype int32)*
>> The sequence steps which lead to this error are: 
>>
>> File "/localDisk/work/workspaces/python/sc/src/learner.py", line 304, in 
>> createUpdates
>> currentGrad = T.grad(self.cost,p)
>> File "/Library/Python/2.7/site-packages/Theano/theano/gradient.py", line 
>> 438, in grad
>> outputs, wrt, consider_constant)
>> File "/Library/Python/2.7/site-packages/Theano/theano/gradient.py", line 
>> 809, in _populate_var_to_app_to_idx
>> account_for(output)
>> File "/Library/Python/2.7/site-packages/Theano/theano/gradient.py", line 
>> 805, in account_for
>> account_for(ipt)
>> File "/Library/Python/2.7/site-packages/Theano/theano/gradient.py", line 
>> 805, in account_for
>> account_for(ipt)
>> File "/Library/Python/2.7/site-packages/Theano/theano/gradient.py", line 
>> 805, in account_for
>> account_for(ipt)
>> File "/Library/Python/2.7/site-packages/Theano/theano/gradient.py", line 
>> 805, in account_for
>> account_for(ipt)
>> File "/Library/Python/2.7/site-packages/Theano/theano/gradient.py", line 
>> 805, in account_for
>> account_for(ipt)
>> File "/Library/Python/2.7/site-packages/Theano/theano/gradient.py", line 
>> 805, in account_for
>> account_for(ipt)
>> File "/Library/Python/2.7/site-packages/Theano/theano/gradient.py", line 
>> 805, in account_for
>> account_for(ipt)
>> File "/Library/Python/2.7/site-packages/Theano/theano/gradient.py", line 
>> 805, in account_for
>> account_for(ipt)
>> File "/Library/Python/2.7/site-packages/Theano/theano/gradient.py", line 
>> 805, in account_for
>> account_for(ipt)
>> File "/Library/Python/2.7/site-packages/Theano/theano/gradient.py", line 
>> 780, in account_for
>> connection_pattern = _node_to_pattern(app)
>> File "/Library/Python/2.7/site-packages/Theano/theano/gradient.py", line 
>> 665, in _node_to_pattern
>> connection_pattern = node.op.connection_pattern(node)
>> File 
>> "/Library/Python/2.7/site-packages/Theano/theano/scan_module/scan_op.py", 
>> line 1373, in connection_pattern
>> ils)
>> File 
>> "/Library/Python/2.7/site-packages/Theano/theano/scan_module/scan_op.py", 
>> line 1301, in compute_gradient
>> known_grads={y: g_y}, wrt=x)
>> File "/Library/Python/2.7/site-packages/Theano/theano/gradient.py", line 
>> 529, in grad
>> grad_dict, wrt, cost_name)
>> File "/Library/Python/2.7/site-packages/Theano/theano/gradient.py", line 
>> 1213, in _populate_grad_dict
>> rval = [access_grad_cache(elem) for elem in wrt]
>> File "/Library/Python/2.7/site-packages/Theano/theano/gradient.py", line 
>> 1173, in access_grad_cache
>> term = access_term_cache(node)[idx]
>> File "/Library/Python/2.7/site-packages/Theano/theano/gradient.py", line 
>> 895, in access_term_cache
>> output_grads = [access_grad_cache(var) for var in node.outputs]
>> File "/Library/Python/2.7/site-packages/Theano/theano/gradient.py", line 
>> 1173, in access_grad_cache
>> term = access_term_cache(node)[idx]
>> File "/Library/Python/2.7/site-packages/Theano/theano/gradient.py", line 
>> 895, in access_term_cache
>> output_grads = [access_grad_cache(var) for var in node.outputs]
>> File "/Library/Python/2.7/site-packages/Theano/theano/gradient.py", line 
>> 1173, in access_grad_cache
>> term = access_term_cache(node)[idx]
>> File "/Library/Python/2.7/site-packages/Theano/theano/gradient.py", line 
>> 1091, in access_term_cache
>> term.type.dtype))
>>
>> TypeError: Join.grad illegally returned an integer-valued variable. 
>> (Input index 1, dtype int32)
>>
>> -- 
>>
>> --- 
>> 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] <javascript:>.
>> For more options, visit https://groups.google.com/d/optout.
>>
>
>

-- 

--- 
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.

Reply via email to