testing error

[jrdyhwh]

Hello, I am a deep learning newbie currently testing your model on a dataset as a proposal for an MS project but when I run it using the provided code, I get this: /usr/local/lib/python3.7/dist-packages/torchvision/transforms/transforms.py:333: UserWarning: Argument interpolation should be of type InterpolationMode instead of int. Please, use InterpolationMode enum. "Argument interpolation should be of type InterpolationMode instead of int. " model [IRWGANModel] was created

Information

task: dibcoirwgan

phase: test

gan_type: lsgan

netD: gl

trainA_size: 93

trainB_size: 84

testA_size: 16

testB_size: 16

Weight

lambda_A: 10

lambda_B: 10

lambda_identity: 1

Model Specific

beta_mode: AB

threshold: 0.1

batch_size: 20

lambda_nos_A: 1

lambda_nos_B: 1

-------------- Networks loaded ---------------- [Network gen_a2b] Total number of parameters : 11.379 M [Network gen_b2a] Total number of parameters : 11.379 M [Network dis_a] Total number of parameters : 13.969 M [Network dis_b] Total number of parameters : 13.969 M [Network beta_net_a] Total number of parameters : 2.757 M [Network beta_net_b] Total number of parameters : 2.757 M

[*] testing start! Traceback (most recent call last): File "main.py", line 117, in test(model, opt, test_loader_a, test_loader_b) File "main.py", line 38, in test dict_a, dict_b = misc.test_fid(test_loader_a, model.gen_a2b, test_loader_b, model.gen_b2a, model.run_dir, opt) File "/content/drive/MyDrive/Colab Notebooks/IrwGAN/models/misc.py", line 125, in test_fid metric_dict_AB = torch_fidelity.calculate_metrics(input1=real_b_path, input2=fake_b_path, **eval_args) File "/usr/local/lib/python3.7/dist-packages/torch_fidelity/metrics.py", line 239, in calculate_metrics featuresdict_1 = extract_featuresdict_from_input_id_cached(1, feat_extractor, **kwargs) File "/usr/local/lib/python3.7/dist-packages/torch_fidelity/utils.py", line 372, in extract_featuresdict_from_input_id_cached featuresdict = fn_recompute() File "/usr/local/lib/python3.7/dist-packages/torch_fidelity/utils.py", line 360, in fn_recompute return extract_featuresdict_from_input_id(input_id, feat_extractor, **kwargs) File "/usr/local/lib/python3.7/dist-packages/torch_fidelity/utils.py", line 342, in extract_featuresdict_from_input_id input = prepare_input_from_id(input_id, **kwargs) File "/usr/local/lib/python3.7/dist-packages/torch_fidelity/utils.py", line 275, in prepare_input_from_id return prepare_input_from_descriptor(input_desc, **kwargs) File "/usr/local/lib/python3.7/dist-packages/torch_fidelity/utils.py", line 230, in prepare_input_from_descriptor vassert(len(input) > 0, f'No samples found in {input} with samples_find_deep={samples_find_deep}') File "/usr/local/lib/python3.7/dist-packages/torch_fidelity/helpers.py", line 9, in vassert raise ValueError(message) ValueError: No samples found in [] with samples_find_deep=False

has anyone ever come across this?

[Mid-Push]

Hi, it looks like that some directories are empty. Could you please go to models/misc.py 125 and add print(fake_b_path, real_b_path) and print(fake_a_path, real_a_path) ?

If you are sure that you have images in these four directories, I will have a look this weekend.