ASCII code error

[morpheuslord]

To enable the following instructions: AVX2 FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags. 2023-02-24 12:27:40.098752: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT Traceback (most recent call last): File "/home/morpheuslord/Desktop/PassGAN-master/sample.py", line 74, in <module> charmap = pickle.load(f) ^^^^^^^^^^^^^^ UnicodeDecodeError: 'ascii' codec can't decode byte 0x83 in position 0: ordinal not in range(128)

So this is what error i am getting in addition with many bracket missing errors which were easy to locate and correct but this error i am not able to wrap my head around

[pronoym99]

I am also getting the same error. I have now converted the entire repository to Python 3.8 using the 2to3.8 command available in Unix systems. However, I continue to get the same error.

[morpheuslord]

same i tried using 2.7 and also 3.11 and 3.10 but same error

[Gurkinator1]

this is relatively easy to fix. after some googling I found out that you can load python2 pickles in python3 by specifying the encoding:

line 73-76 in sample.py

 with open(os.path.join(args.input_dir, 'charmap.pickle'), 'rb') as f:
     charmap = pickle.load(f, encoding="latin1")
 
 with open(os.path.join(args.input_dir, 'inv_charmap.pickle'), 'rb') as f:
     inv_charmap = pickle.load(f, encoding="latin1")

I just changed those lines + added brackets to all print statements and that was it. I get a different error, TypeError: Variable is unhashable. but thats a seperate issue

[RarMane]

this is relatively easy to fix. after some googling I found out that you can load python2 pickles in python3 by specifying the encoding:

line 73-76 in sample.py

 with open(os.path.join(args.input_dir, 'charmap.pickle'), 'rb') as f:
     charmap = pickle.load(f, encoding="latin1")
 
 with open(os.path.join(args.input_dir, 'inv_charmap.pickle'), 'rb') as f:
     inv_charmap = pickle.load(f, encoding="latin1")

I just changed those lines + added brackets to all print statements and that was it. I get a different error, TypeError: Variable is unhashable. but thats a seperate issue

Did you solved problem with "TypeError: Variable is unhashable."?

[Gurkinator1]

Did you solved problem with "TypeError: Variable is unhashable."?

As far as I know it's caused by newer tensorflow versions. I "fixed" it by changing lines 28-30 of file PassGAN/tflib/__init__.py

- result = _params[name]
+ result = _params[name].ref()
i = 0
- while result in _param_aliases:
+ while result in _param_aliases.keys(): #not sure if this is required

However now tensorflow has problems transforming the value into a tensor. in File "PassGAN/tflib/ops/linear.py", line 133, in Linear: ValueError: Attempt to convert a value (<Reference wrapping <tf.Variable 'Generator.Input/Generator.Input.W:0' shape=(128, 1280) dtype=float32, numpy= array([...], dtype=float32)>>) with an unsupported type (<class 'tensorflow.python.util.object_identity.Reference'>) to a Tensor.

I dont know enough about tensorflow to fix this. I tried using tf.identity to manually dereference the value, but that didnt work either. This project uses both python2 and outdated tensorflow versions. I dont know if its even worth the trouble
¯\(ツ)/¯