Lesson 1 In-Class Discussion ✅

[ymittal23 (Yash Mittal)]

1. If you like a post, it’s better to “like” it with the rather than commenting. It saves traffic in the Forums and makes it easier for everyone to find posts.
   A- Not applicable
2. Please be mindful of looking to see if a topic exists before starting a new thread.
   A- I didn’t find this topic after searching
3. Do a quick search of the forums to see if your question is already under discussion.
   A- Tried but didn’t find
4. Do not @ people if you are not referencing them for a specific reason that requires the attention of that forum member. Be specially mindful in mentioning Jeremy and Rachel.
   A- Didn’t find the answer for problem by anybody else so I tagged you
5. Regional Study Group discussions should be posted in the relevant category related to timezones.
   A- Not applicable

Sorry, if I have done something wrong by mistake.

[jeremy (Jeremy Howard)]

Do not @ people if you are not referencing them for a specific reason that requires the attention of that forum member. Be specially mindful in mentioning Jeremy and Rachel

This seems quite clear?..

[ymittal23 (Yash Mittal)]

Ok sorry could you direct me where else can I ask my question if it is not answered on forum

[Neural-instinct (Vignesh J)]

Hi,
I would like to know how get_transforms() method works in detail.
Providing the link to the repository [https://github.com/fastai/fastai/blob/master/fastai/vision/transform.py]
1.What are affines and why it is used?
2.How changing brightness,contrast,cropping or zooming would help in satellite images,for example finding water resources.It can be any use case.
Just need a basic explanation of how this is used to manipulate images.Does this manipulation correlate to higher accuracy?
Also, How does this transformation techniques correspond to this paper[http://cs231n.stanford.edu/reports/2017/pdfs/300.pdf]
Thanks!

[lesscomfortable (Francisco Ingham)]

Hey, @ymittal23! This is the correct place to ask but you should not tag Jeremy for code-specific questions.

Tensor b is double the specified batch size because the validation set batch size is double the training set batch size in fastai (and the problem here concerns the validation set).

When does this bug arise? When you call fit or when it is finishing an epoch?

[lesscomfortable (Francisco Ingham)]

How does your csv file look like? Can you send a screenshot?

[balnazzar (Andrea de Luca)]

to summarize I don’t think there is any direct relationship at all between the two.

Visualizing Learning rate vs Batch size

[rreilly (Rebecca Reilly)]

Can anyone help with this error in the lesson 1 notebook:

OSError: cannot identify image file ‘/home/jupyter/.fastai/data/oxford-iiit-pet/images/Birman_141.jpg’

I’ve re-pulled the repo and continue to get the same error

[ymittal23 (Yash Mittal)]

The bug is arising when epoch is completed. The error is coming with fbeta matric only with accuracy it is working fine.

[basu (bhaskar rao)]

how access the file names when we use learn.predict(is_test=True) ?

data.test_ds.fnames shows ‘ImageClassificationDataset’ object has no attribute ‘test_ds’

[lesscomfortable (Francisco Ingham)]

Can you run len(data.valid_ds) and tell me what is the output?

[basu (bhaskar rao)]

it prints the len ‘110924’

[ymittal23 (Yash Mittal)]

111110
If we restart kernel and run it again it will be different but is almost in same range.
For eg it changed to 110812 when I run the kernel again by restarting.

[tank13]

Getting a 404 from https://docs-dev.fast.ai/troubleshoot.html. Did the documentation move?

[Pranav7296 (Pranav Bajaj)]

Thank you

[rreilly (Rebecca Reilly)]

Experiencing a corrupted file in lesson 1 pets data, anyone know how to get around this? The file exists, and I get “OSError: cannot identify image file ‘/home/jupyter/.fastai/data/oxford-iiit-pet/images/Birman_141.jpg’” when I attempt the Jupiter notebook

[Kaspar (Kaspar Lund)]

have you tried to delete the dataset and let the notebook download it again

[rreilly (Rebecca Reilly)]

I haven’t, I wasn’t sure if our notebook was downloading from a URL or if the zipped data existed from the git pull?

[Kaspar (Kaspar Lund)]

have a look in the notebook. i belive that the line untar also downloads.
anyway you could also just delete the offensive file

[rreilly (Rebecca Reilly)]

I deleted all the pet data and re-ran the notebook, but even when it downloads from the untar, I get the same error

OSError: cannot identify image file ‘/home/jupyter/.fastai/data/oxford-iiit-pet/images/Birman_141.jpg’