On my machine (AWS instance) that doesn’t have GPU access, I always see these two error messages:
ERROR (theano.gpuarray): Could not initialize pygpu, support disabled
Traceback (most recent call last):
....blah blah blah...
and
WARNING (theano.tensor.blas): Using NumPy C-API based implementation for BLAS functions.
These fill up my experiment logs and make important stuff hard to find. Is there some way I could suppress them?