issue about TensorType

[wangdaxiao]

when I run demos it appears problem like this:

eeedl@eeedl-OptiPlex-7040:~$ cortex-rbm-demo
Loading experiment from /home/eeedl/anaconda2/lib/python2.7/site-packages/cortex-0.12a0-py2.7.egg/rbm_mnist.yaml
Experiment hyperparams: {'dataset_args': {'dataset': 'mnist',
'source': '$data/basic/mnist_binarized_salakhutdinov.pkl.gz'},
'dim_h': 200,
'inference_args': {'n_chains': 10, 'n_steps': 1, 'persistent': True},
'learning_args': {'epochs': 1000, 'learning_rate': 0.01, 'optimizer': 'sgd'},
'name': 'rbm_mnist',
'test_every': 10}
Saving to /home/eeedl/cortex/program/rbm_mnist
Dataset args: {'dataset': 'mnist',
'source': '$data/basic/mnist_binarized_salakhutdinov.pkl.gz'}
Learning args: {'batch_size': 100,
'epochs': 1000,
'excludes': [],
'learning_rate': 0.01,
'learning_rate_schedule': None,
'optimizer': 'sgd',
'optimizer_args': {},
'valid_batch_size': 100,
'valid_key': 'nll',
'valid_sign': '+'}
Inference args: {'n_chains': 10, 'n_steps': 1, 'persistent': True}
---Setting up data--------------------------------------------------------------------------------------------------------------------------------------
Loading mnist (train) from /home/eeedl/cortex/program/basic/mnist_binarized_salakhutdinov.pkl.gz
Loading mnist (valid) from /home/eeedl/cortex/program/basic/mnist_binarized_salakhutdinov.pkl.gz
---Setting model and variables--------------------------------------------------------------------------------------------------------------------------
---Loading model and forming graph----------------------------------------------------------------------------------------------------------------------
Print profile for tparams (name, shape)
rbm_W (784, 200)
rbm_visible_z (784,)
rbm_hidden_z (200,)
---Getting cost-----------------------------------------------------------------------------------------------------------------------------------------
---Test functions---------------------------------------------------------------------------------------------------------------------------------------
---Setting final tparams and save function--------------------------------------------------------------------------------------------------------------
Learned model params: ['rbm_W', 'rbm_visible_z', 'rbm_hidden_z']
Saved params: ['rbm_W', 'rbm_visible_z', 'rbm_hidden_z']
---Getting gradients and building optimizer.------------------------------------------------------------------------------------------------------------
Traceback (most recent call last):
File "/home/eeedl/anaconda2/bin/cortex-rbm-demo", line 11, in
load_entry_point('cortex==0.12a0', 'console_scripts', 'cortex-rbm-demo')()
File "/home/eeedl/anaconda2/lib/python2.7/site-packages/cortex-0.12a0-py2.7.egg/cortex/demos/demos_basic/init.py", line 30, in run_rbm_demo
run_demo(yaml_file, train)
File "/home/eeedl/anaconda2/lib/python2.7/site-packages/cortex-0.12a0-py2.7.egg/cortex/demos/demos_basic/init.py", line 18, in run_demo
train(**exp_dict)
File "/home/eeedl/anaconda2/lib/python2.7/site-packages/cortex-0.12a0-py2.7.egg/cortex/demos/demos_basic/rbm_mnist.py", line 167, in train
[X], cost, tparams, constants, updates, extra_outs, **learning_args)
File "/home/eeedl/anaconda2/lib/python2.7/site-packages/cortex-0.12a0-py2.7.egg/cortex/utils/training.py", line 291, in set_optimizer
extra_outs=extra_outs, **optimizer_args)
File "/home/eeedl/anaconda2/lib/python2.7/site-packages/cortex-0.12a0-py2.7.egg/cortex/utils/op.py", line 301, in sgd
f_update = theano.function(lr, [], updates=pup, profile=profile)
File "/home/eeedl/anaconda2/lib/python2.7/site-packages/theano/compile/function.py", line 326, in function
output_keys=output_keys)
File "/home/eeedl/anaconda2/lib/python2.7/site-packages/theano/compile/pfunc.py", line 449, in pfunc
no_default_updates=no_default_updates)
File "/home/eeedl/anaconda2/lib/python2.7/site-packages/theano/compile/pfunc.py", line 208, in rebuild_collect_shared

**# **# __raise TypeError(err_msg, err_sug)

TypeError: ('An update must have the same type as the original shared variable (shared_var=W, shared_var.type=TensorType(float32, matrix), update_val=Elemwise{sub,no_inplace}.0, update_val.type=TensorType(float64, matrix)).', 'If the difference is related to the broadcast pattern, you can call the tensor.unbroadcast(var, axis_to_unbroadcast[, ...]) function to remove broadcastable dimensions.')****

and I use theano 0.9.0.

[carricky]

Same issue here. Also using theano 0.9.0