I am using this notebook which is combination of 3 notesbooks (linked in the second cell) for finetuning the Inception V3 for training the last layers on a cats and dogs dataset (from Kaggle).
import torchvision.models as models
model_ft = models.inception_v3(pretrained=True)
for param in model_ft.parameters():
param.requires_grad = False
# Parameters of newly constructed modules have requires_grad=True by default
# Handle the auxilary net
num_ftrs = model_ft.AuxLogits.fc.in_features
model_ft.AuxLogits.fc = nn.Linear(num_ftrs, 2)
# Handle the primary net
num_ftrs = model_ft.fc.in_features
model_ft.fc = nn.Linear(num_ftrs, 2)
I am trying to understand if I have done this correctly, so I decided to use the visualization module from PyTorch trainings. However, I get some problem in visualization in cell [57].
Error is:
ValueError Traceback (most recent call last)
Input In [57], in <module>
----> 1 visualize_model(model_ft)
Input In [56], in visualize_model(model, num_images)
15 for j in range(inputs.size()[0]):
16 images_so_far += 1
---> 17 ax = plt.subplot(num_images//2, 2, images_so_far)
18 ax.axis('off')
19 ax.set_title('predicted: {}'.format(class_names[preds[j]]))
File ~/research/venv/dpcc/lib/python3.8/site-packages/matplotlib/pyplot.py:1268, in subplot(*args, **kwargs)
1265 fig = gcf()
1267 # First, search for an existing subplot with a matching spec.
-> 1268 key = SubplotSpec._from_subplot_args(fig, args)
1270 for ax in fig.axes:
1271 # if we found an axes at the position sort out if we can re-use it
1272 if hasattr(ax, 'get_subplotspec') and ax.get_subplotspec() == key:
1273 # if the user passed no kwargs, re-use
File ~/research/venv/dpcc/lib/python3.8/site-packages/matplotlib/gridspec.py:608, in SubplotSpec._from_subplot_args(figure, args)
606 else:
607 if not isinstance(num, Integral) or num < 1 or num > rows*cols:
--> 608 raise ValueError(
609 f"num must be 1 <= num <= {rows*cols}, not {num!r}")
610 i = j = num
611 return gs[i-1:j]
ValueError: num must be 1 <= num <= 8, not 9
Also, based on the output I get in cell 38, is it fair to say that since the first element in output tensor is larger and I only have two classes, so it is a cat? I actually am testing on a random cat image from Web for this cell.
