Fastai-v4_"05_pet_breeds" FileNotFoundError

Part 1 (2019)
1

I am getting error while i try to run this piece of code from the fastbook,here is the code

    dblock1 = DataBlock(blocks=(ImageBlock(), CategoryBlock()),
                       get_y=parent_label,
                       item_tfms=Resize(460))
   dls1 = dblock1.dataloaders([(Path.cwd()/'images'/'grizzly.jpg')]*100, bs=8)
   dls1.train.get_idxs = lambda: Inf.ones
   x,y = dls1.valid.one_batch()
   _,axs = subplots(1, 2)

and the error being -

FileNotFoundError: [Errno 2] No such file or directory: '/content/images/grizzly.jpg'

error being specificly for the line-

dls1 = dblock1.dataloaders([(Path.cwd()/'images'/'grizzly.jpg')]*100, bs=8)

Could anyone clarify what is going wrong?i am running it on google colab

2

sessions on colab get recycled pretty quickly if you leave them idle which means all the files vanish. when you come back it automatically reconnects but you’ll need to re-run previous cells which your current cell depends on. in your case probably just re-run everything from the top.

3

i tried re-running the notebook but the error persisted,could you try it once yourself just in case you face the same issue

4

Are you working on the cloned directory? It shows the path /content/images/ as the directory where the grizzly.jpg was not found. Maybe you have cloned the github repository but aren’t working where the grizzly.jpg is present.
If you have cloned the course-v4 on the google drive i.e. within content/My Drive/course-v4, then you need to go there to access the file or you could copy the course-v4 directory to content/ to access the files where you are right now.

2 Likes
5

Hi, I have the impression that the folder we just uploaded/untared is with pets not with bears. Perhaps this line is from a previous notebook. I have tried the following succesfully:

Try:
dls1 = dblock1.dataloaders([(path/'images'/'samoyed_167.jpg')]*100, bs=8)
instead of
dls1 = dblock1.dataloaders([(Path.cwd()/'images'/'grizzly.jpg')]*100, bs=8)

Cheers,
M

1 Like