Heya, I have one question about transform. When we apply transform to the planet data - changing the size in this case, I am assuming (from reading the source code in https://github.com/fastai/fastai/blob/master/fastai/data_block.py#L498) that the transform is applied to both training and validation - please do correct me if this is not correct.
In the absence of test set and since we base our evaluation performance based on validation set, shouldn’t we restrict transform to training set only?
I’m having a CUDA out of memory issue. I have restarted the instance and killed the kernal an reopened the notebook with no success. Each time I have tried to recreate the learner and then load my resnet-stage-2 model to continue training the gpu runs out of memory.
If I don’t load my old model I can’t continue to train it, am I correct?
Is there any way that we can interpret the multi-label model?
Hello all,
If the learning rate curve doesn’t follow that initial down-slope followed by an up-slope does that indicate an issue? It also might be my plotting settings as I don’t seem to plot the curve to x=1e-01.
I have problem understanding how changing the input size of the image doesnt affect the inputs/outputs of the network. Changing the input size in a convolutional layer should definitely change the output right? How can i use transfer learning with a bigger image size then, i just dont get it.
Even I was having the same doubt. The architecture, I guess, does change. Please anyone correct me if I am wrong!
I searched over net for transfer learning with different sizes and it led me to this blog post from Adrian Rosebrock which confirms the change in architecture.
I also came across this thread from the forum where Jeremy says that fastai automatically handles it and later in the same thread mentions about having used adaptive pooling to do so.
I am still new to course (and the techniques), digging through the excellent fastai library to find where and how exactly this is handled. Any pointers will be useful though!
Could somebody tell me the meaning of ‘/f’ in the lambada expression:
get_y_fn = lambda x: path_lbl/f’{x.stem}_P{x.suffix}’
thanks
Hello,
I’m getting a syntax error when I try to run the following in windows:
Any suggestions please?
Cheers.
During Segmentation Learning topic, I didn’t understand why Jeremy used acc_camvid as accuracy metric instead of accuracy. Any suggestion?
Hi all,
I am having trouble with visualizing where the Kaggle data for the planet Amazon exercise exists on my instance using Salamander. From the previous example, I can navigate using Jupyter Notebooks to see the images of bears/teddy bears. But I cannot see the same for the planets example. What am I missing?
Thanks,
Luke
Hi, I’m trying to run the planets notebook, and I’m getting a memory error when trying to find the learning rate. I tried reducing the batch size to 32 by running data = (src.transform(tfms, size=256).databunch(bs=32).normalize(imagenet_stats)), but that didn’t work, so I also tried batch sizes of 16, 4, 2, and 1, always getting the same memory error. I’ve searched the forum and found others on this thread having the same problem, but changing the batch size seemed to work for them.
Am I changing the batch size incorrectly, or is there something else I should try?
Thanks for your insight!
If train loss > valid loss = underfit and train loss < valid loss = overfit is true,
does this mean a good model should have similar levels of training loss and validation loss?
Another thing that bothers me is the way he uses the learning rate. In some videos, he uses two different numbers for the same type of learning plot. Lesson(1) - slice(1e-3, 1e-5) and Lesson(2) - slice(3e-3, 3e-5).
Hope someone answer this.
Hi, @ghubs. I found this post on learning rates very helpful: Determining when you are overfitting, underfitting, or just right?.
Regarding the notation for the learning rate: are you familiar with scientific notation? If not, Khan Academy is a fantastic free resource, with a good series on the topic: https://www.khanacademy.org/math/pre-algebra/pre-algebra-exponents-radicals#pre-algebra-scientific-notation
If you’re already familiar with scientific notation, you might be used to seeing it notated as something times 10 to a power, and this notation using e is actually equivalent. The ‘e’ represents “10 to the power of,” not the number e! https://en.wikipedia.org/wiki/Scientific_notation#E-notation
So 1e-3 == 1 * 10^(-3) == 0.001. If the number in the decimal place needs to be something other than 1, you’ll use that number. For instance, if you needed to express 0.05 in this notation, you’d have 5e-2, i.e. “5 times 10 to the power of -2”.
I hope this helps!
That is very helpful. Thank you Laura 
Also the links are amazing. Now I understand that choosing the learning rate is not only dependent on the LR plot but also to do with the intuition and experience gained from running many different models and datasets.
can we use slice(1e-5, 1e-3) as explained in lesson1 instead of 3e-02?
Hi 
I had a quick question regarding accuracy (in a regular classifier not multi classifier). Jeremy said that basically what happens is we use argmax to find the index/label of the predicted class and compare it to the actual and then we take a mean.
What I can’t figure out is where does a mean come from? What are we taking a mean of? ie if we have a predict value with an index of 4 for example and an actual with index of 5 when we compare those (4==5) how can we take a mean on that?
I feel like it has to do something that these might be vector/matrices but I was hoping someone could clarify this up a bit.
Where are you running your Jupiter notebook at?
I think windows commands assume you are running your notebook locally on your windows machine or windows VM and in case that you are doing that my best guess would be to run last two commands and not first two ones.
In case that you are running your notebook on google cloud and similar platforms just run first two regardless of if you have a windows/Mac/linux machine since that’s running in the cloud. Hopefully that helps!
Hi, @novarac23! I’m new at this, too, so this might not be correct, but reading the docs for the accuracy function:
def accuracy(input:Tensor, targs:Tensor)->Rank0Tensor:
"Computes accuracy with `targs` when `input` is bs * n_classes."
n = targs.shape[0]
input = input.argmax(dim=-1).view(n,-1)
targs = targs.view(n,-1)
return (input==targs).float().mean()
It looks like the == operator is returning not True or False but instead a tensor of numeric values, which are then converted to floats, and then it takes the mean of those floats? I’m not sure what it’s doing to generate the numeric values instead of the boolean values we’d expect from ==, so hopefully someone else will chime in, too.
It’s a Boolean. So we take if they’re the same, it’s 1. If not. It’s 0. Then we sum up all those correct and incorrect predictions, and take the average. This gets us accuracy. Total amount of 1’s / total
Ah, that makes sense! Thanks. 