Lesson 1: What is 360/360 and 32/32 while training the model

Hey all,

Just wondering why/where the 360/360 and 32/32 comes from while training the model.

It seems to be picked by default. Why does it go from 1 - 360?

Just trying to look thru the code, and not able to figure this out.

Any suggestions welcome.



1 Like

Another doubt – In the code @jeremy sets the size (sz=224) before passing it the function tfms_from_model.
Why was 224 picked? I saw that reducing the size from 224 to say 24, reduced the accuracy also to 75%. But curious why 224 was picked by default.

That’s because resnet, the model we use for this experiment, takes in images of size 224 x 224 x 3 as input. 224 is the image width and height.

1 Like

The images don’t start at 224 x 224, as ResNet34 was trained on, so a few calls down this function uses openCV to scale the images appropriately when preparing the data.

I’m surprised you did as well as 75%, since the images you created setting sz = 24 would have been roughly 1% the size of what the model “expected”, ignoring color-depth.

As for your original question, I’m not too sure; something to do with loading the pretrained model?


360 is the number of batches it takes to precompute the activations (~23000 images / 64 batch size) in the train set.

As @A_TF57 says, resnet34 was originally trained on 224x224 images IIRC, so it’s a reasonable starting point. We’ll get better results still with larger images. We’ll see that on Monday. Frankly, there’s nothing that special about 224x224, and I guess it’s kinda habit (it used to be that we had to use the exact same dimensions as the original model used, but that’s not the case any more).


ohhhhhh Excellent question, thank you.

What is the 32/32?

Validation set precomputed activations.

1 Like

While I run this part of the code,

data = ImageClassifierData.from_paths(PATH, tfms=tfms_from_model(resnet34, sz)) learn = ConvLearner.pretrained(resnet34, data, precompute=True)

I see that there are multiple GPU processes created and I wonder why?! The first process that takes up higher amount of GPU memory is the the actual training process, I presume. And what about the other smaller processes?

The other processes are for the preprocessing. They don’t actually need to use the GPU, so it would be nice to find a way to stop them taking up GPU memory…

Exactly what I thought. The DataLoaders use multiple workers to read data and apply transformations on the CPU is what I assumed. Correct me if I’m wrong.

the sum of val data is 1000+1000=2000,so 2000/64(batch_size)=31.25,This is 32/32