Fixed PyTorch 0.4.0 changes (changes torch.cuda.float tensors to float)

fixes facebookresearch#51

src/dico_builder.py

@@ -162,6 +162,8 @@ def build_dictionary(src_emb, tgt_emb, params, s2t_candidates=None, t2s_candidat
     elif params.dico_build == 'T2S':
         dico = t2s_candidates
     else:
+        s2t_candidates = s2t_candidates.numpy() if isinstance(s2t_candidates, torch.Tensor) else s2t_candidates
+        t2s_candidates = t2s_candidates.numpy() if isinstance(t2s_candidates, torch.Tensor) else t2s_candidates
         s2t_candidates = set([(a, b) for a, b in s2t_candidates])
         t2s_candidates = set([(a, b) for a, b in t2s_candidates])
         if params.dico_build == 'S2T|T2S':

src/evaluation/evaluator.py

@@ -9,6 +9,7 @@
 from copy import deepcopy
 import numpy as np
 from torch.autograd import Variable
+from torch import Tensor as torch_tensor
 
 from . import get_wordsim_scores, get_crosslingual_wordsim_scores, get_wordanalogy_scores
 from . import get_word_translation_accuracy
@@ -202,6 +203,7 @@ def dist_mean_cosine(self, to_log):
                 mean_cosine = -1e9
             else:
                 mean_cosine = (src_emb[dico[:dico_max_size, 0]] * tgt_emb[dico[:dico_max_size, 1]]).sum(1).mean()
+            mean_cosine = mean_cosine.item() if isinstance(mean_cosine, torch_tensor) else mean_cosine
             logger.info("Mean cosine (%s method, %s build, %i max size): %.5f"
                         % (dico_method, _params.dico_build, dico_max_size, mean_cosine))
             to_log['mean_cosine-%s-%s-%i' % (dico_method, _params.dico_build, dico_max_size)] = mean_cosine