Local Ubuntu/GPU module incompatibility issue

Hi all,
I am working through the course with the folks over at https://twimlai.com/twiml-x-fast-ai/
I am using Kai Lichtenberg’s setup and it seems to work well most of the time however I have received a couple of odd errors when replicating Jeremy’s code as follows:
From the Lecture 2 section on building your own image classifier:

accuracy(log_preds, y)

TypeError Traceback (most recent call last)
in ()
----> 1 accuracy(log_preds, y)
2 # p = np.max(preds)
3 # (preds==y).mean()

/fastai/courses/dl1/fastai/metrics.py in accuracy(preds, targs)
8 def accuracy(preds, targs):
----> 9 preds = torch.max(preds, dim=1)[1]
10 return (preds==targs).float().mean()

TypeError: torch.max received an invalid combination of arguments - got (numpy.ndarray, dim=int), but expected one of:

  • (torch.FloatTensor source)
  • (torch.FloatTensor source, torch.FloatTensor other)
    didn’t match because some of the keywords were incorrect: dim
  • (torch.FloatTensor source, int dim)
  • (torch.FloatTensor source, int dim, bool keepdim)

and in Lecture 2’s section on dog breeds:


TypeError Traceback (most recent call last)
in ()
----> 1 row_sz, col_sz = list(zip(*size_d.values()))

TypeError: zip argument #1 must support iteration

Since it worked for Jeremy something must be different or has changed. Any suggestions?

did you had problems with lesson1.ipynb?

Yes. Everything worked well. My results were not quite as good as Jeremy’s but were well within a percentage point. No errors.

Resolved. Typo on my part. Thanks