More fixes for recent PyTorch (incl. topk issue) (#113)

README.md

@@ -24,7 +24,7 @@ If you use the code in your paper, then please cite it as:
 * Python version 3.6
 * A [PyTorch installation](http://pytorch.org/)
 
-Currently fairseq-py requires PyTorch version >= 0.3.0.
+Currently fairseq-py requires PyTorch version >= 0.4.0.
 Please follow the instructions here: https://github.com/pytorch/pytorch#installation.
 
 If you use Docker make sure to increase the shared memory size either with `--ipc=host` or `--shm-size` as command line

fairseq/sequence_generator.py

@@ -271,8 +271,8 @@ class SequenceGenerator(object):
             if step < maxlen:
                 if prefix_tokens is not None and step < prefix_tokens.size(1):
                     probs_slice = probs.view(bsz, -1, probs.size(-1))[:, 0, :]
-                    cand_scores = probs_slice.gather(
-                        dim=1,
+                    cand_scores = torch.gather(
+                        probs_slice, dim=1,
                         index=prefix_tokens[:, step].view(-1, 1).data
                     ).expand(-1, cand_size)
                     cand_indices = prefix_tokens[:, step].view(-1, 1).expand(bsz, cand_size).data
@@ -341,17 +341,17 @@ class SequenceGenerator(object):
             # get the top beam_size active hypotheses, which are just the hypos
             # with the smallest values in active_mask
             active_hypos, _ignore = buffer('active_hypos'), buffer('_ignore')
-            active_mask.topk(
-                k=beam_size, dim=1, largest=False,
-                out=(_ignore, active_hypos),
+            torch.topk(
+                active_mask, k=beam_size, dim=1, largest=False,
+                out=(_ignore, active_hypos)
             )
             active_bbsz_idx = buffer('active_bbsz_idx')
-            cand_bbsz_idx.gather(
-                dim=1, index=active_hypos,
+            torch.gather(
+                cand_bbsz_idx, dim=1, index=active_hypos,
                 out=active_bbsz_idx,
             )
-            active_scores = cand_scores.gather(
-                dim=1, index=active_hypos,
+            active_scores = torch.gather(
+                cand_scores, dim=1, index=active_hypos,
                 out=scores[:, step].view(bsz, beam_size),
             )
             active_bbsz_idx = active_bbsz_idx.view(-1)

generate.py

@@ -127,7 +127,10 @@ def main(args):
                             hypo['positional_scores'].tolist(),
                         ))
                     ))
-                    print('A-{}\t{}'.format(sample_id, ' '.join(map(str, alignment))))
+                    print('A-{}\t{}'.format(
+                        sample_id,
+                        ' '.join(map(lambda x: str(utils.item(x)), alignment))
+                    ))
 
                 # Score only the top hypothesis
                 if i == 0:

requirements.txt

 cffi
 numpy
-torch>=0.3.0
+torch>=0.4.0
 tqdm