Hello,
I am running the interactive translation (via PyTorch Hub) example.
When I run the command
en2de = torch.hub.load('pytorch/fairseq', 'transformer.wmt16.en-de', tokenizer='moses', bpe='subword_nmt')
I am getting the following error:
File "torch/hub/pytorch_fairseq_master/fairseq/checkpoint_utils.py", line 140, in load_checkpoint_to_cpu
from fairseq.fb_pathmgr import fb_pathmgr
ImportError: No module named 'fairseq.fb_pathmgr'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/lib/python3.5/site-packages/torch/hub.py", line 363, in load
model = entry(*args, **kwargs)
File "torch/hub/pytorch_fairseq_master/fairseq/models/fairseq_model.py", line 172, in from_pretrained
**kwargs,
File "torch/hub/pytorch_fairseq_master/fairseq/hub_utils.py", line 54, in from_pretrained
arg_overrides=kwargs,
File "torch/hub/pytorch_fairseq_master/fairseq/checkpoint_utils.py", line 178, in load_model_ensemble_and_task
state = load_checkpoint_to_cpu(filename, arg_overrides)
File "torch/hub/pytorch_fairseq_master/fairseq/checkpoint_utils.py", line 145, in load_checkpoint_to_cpu
except (ModuleNotFoundError, ImportError):
NameError: name 'ModuleNotFoundError' is not defined
Can you please help me resolve this?
Thanks!
Same here, same error
same error
This should be fixed.
I'm still having this error with fairseq 0.9.0
same here
And same here.
And same here.
Still have the issue
@myleott I am having same issue. Anybody able to fix it or find a solution?
Same issue here @myleott
Most helpful comment
I'm still having this error with fairseq 0.9.0