In that train.py script i got an issue regarding eval() missing 1 required positional argument: 'fake_dir' how to solve it i tried out so many things but its not working about audiodeepfakedetection HOT 9 CLOSED

How did you execute the script? What have you tried? What are the outputs? You need to provide more information.

from audiodeepfakedetection.

from audiodeepfakedetection.

I go through the terminal and run python train.py file and i got the following issue after giving the real dataset and fake dataset path such as
real_dir=data/real/wavs",
fake_dir=data/fake/generated_audio/ljspeech_melgan
and i got the following error :

2022-12-28 11:33:55,219 - INFO - Batch size: 256, seed: None, epochs: 20

Experiment Failed: ShallowCNN_lfcc_I

2022-12-28 11:33:55,219 - ERROR - eval() missing 1 required positional argument: 'fake_dir'
Traceback (most recent call last):
File "/home/soumya/deepfake_detection/AudioDeepFakeDetection/train.py", line 599, in main
experiment(
File "/home/soumya/deepfake_detection/AudioDeepFakeDetection/train.py", line 447, in experiment
train(
File "/home/soumya/deepfake_detection/AudioDeepFakeDetection/train.py", line 143, in train
Model = eval(model_classname)
TypeError: eval() missing 1 required positional argument: 'fake_dir'

from audiodeepfakedetection.

Hi @soumyajee, thanks for the info. The exception is due to a bug in my train.py indeed, there was a conflict between my custom function eval() and Python's built-in function eval(). At line 143, where your error occurred, I intended to use the built-in eval().

I have committed a fix to this bug in the master branch, please use the latest code to run again.

from audiodeepfakedetection.

Again i got the following issue while running trin.py file:
2022-12-29 16:54:21,966 - ERROR -
Traceback (most recent call last):
File "/home/soumya/deepfake_detection/AudioDeepFakeDetection/train.py", line 599, in main
experiment(
File "/home/soumya/deepfake_detection/AudioDeepFakeDetection/train.py", line 447, in experiment
train(
File "/home/soumya/deepfake_detection/AudioDeepFakeDetection/train.py", line 165, in train
assert melgan_dir.is_dir()
AssertionError

from audiodeepfakedetection.

@soumyajee Can you run python train.py --debug_all and see what the result is? It seems like a data path issue.

from audiodeepfakedetection.

I got the following issue while running train.py --debug_all its following:
Traceback (most recent call last):
File "/home/soumya/deepfake_detection/AudioDeepFakeDetection/train.py", line 621, in
main()
File "/home/soumya/deepfake_detection/AudioDeepFakeDetection/train.py", line 589, in main
debug(args.real, args.fake, args.device)
AttributeError: 'Namespace' object has no attribute 'real'

from audiodeepfakedetection.

@soumyajee you are not using my latest code, this issue has been fixed in commit 5e29c11

from audiodeepfakedetection.

Thanks for ur support we have already solve that issue now our scripts running fine

from audiodeepfakedetection.