def __iter__(self):
# NB: we use 'imap' and not 'map' here, so that in Python 2 we get a
# generator and don't eagerly perform all the indexes. This could
# save us work, and also helps keep trace ordering deterministic
# (e.g., if you zip(*hiddens), the eager map will force all the
# indexes of hiddens[0] before hiddens[1], while the generator
# map will interleave them.)
return iter(imap(lambda i: self[i], range(self.size(0))))
def __iter__(self):
# NB: we use 'imap' and not 'map' here, so that in Python 2 we get a
# generator and don't eagerly perform all the indexes. This could
# save us work, and also helps keep trace ordering deterministic
# (e.g., if you zip(*hiddens), the eager map will force all the
# indexes of hiddens[0] before hiddens[1], while the generator
# map will interleave them.)
return iter(imap(lambda i: self[i], range(self.size(0))))
def __iter__(self):
# NB: we use 'imap' and not 'map' here, so that in Python 2 we get a
# generator and don't eagerly perform all the indexes. This could
# save us work, and also helps keep trace ordering deterministic
# (e.g., if you zip(*hiddens), the eager map will force all the
# indexes of hiddens[0] before hiddens[1], while the generator
# map will interleave them.)
if self.dim() == 0:
raise TypeError('iteration over a 0-d tensor')
if torch._C._get_tracing_state():
warnings.warn('Iterating over a tensor might cause the trace to be incorrect. '
'Passing a tensor of different shape won\'t change the number of '
'iterations executed (and might lead to errors or silently give '
'incorrect results).', category=RuntimeWarning)
return iter(imap(lambda i: self[i], range(self.size(0))))
