Closed DmitryKey closed 9 years ago
You can ignore them. It is complicated detail.
The UserWarning is now removed if you update Theano to the latest development version.
The RuntimeWarning is related by numpy, but you can ignore it.
On Tue, Feb 24, 2015 at 12:28 AM, Dmitry Kan notifications@github.com wrote:
Hi!
Upon starting training from a saved state, I see this warning in the logs. Should I be worried about it and can it be fixed?
2015-02-24 07:24:28,242: experiments.nmt.encdec: DEBUG: Build log-likelihood computation graph 2015-02-24 07:24:28,313: groundhog.layers.cost_layers: DEBUG: Get grads 2015-02-24 07:24:29,292: groundhog.layers.cost_layers: DEBUG: Got grads 2015-02-24 07:24:29,292: experiments.nmt.encdec: DEBUG: Build sampling computation graph /Users/dmitry/Theano/theano/sandbox/rng_mrg.py:1195: UserWarning: MRG_RandomStreams Can't determine #streams from size (Elemwise{Cast{int32}}.0), guessing 60*256 nstreams = self.n_streams(size) 2015-02-24 07:24:29,666: experiments.nmt.encdec: DEBUG: Create auxiliary variables 2015-02-24 07:24:29,666: experiments.nmt.encdec: DEBUG: Compile sampler /Users/dmitry/Theano/theano/scan_module/scan_perform_ext.py:133: RuntimeWarning: numpy.ndarray size changed, may indicate binary incompatibility
— Reply to this email directly or view it on GitHub https://github.com/lisa-groundhog/GroundHog/issues/25.
thanks @nouiz, that is a relief :)
Hi!
Upon starting training from a saved state, I see this warning in the logs. Should I be worried about it and can it be fixed?