AttributeError: Can't get attribute 'my_func' on <module '__main__' from ''>

I want to create Python scripts, based on a notebook, to get a runtime using the same .pkl file.

Gist: model.pkl

On this line:

learn = load_learner('model.pkl', cpu=True)

I get this error:

(project) daniel@ubuntu-pcs:~/PycharmProjects/project$ python 
Traceback (most recent call last):
  File "", line 6, in <module>
    from src.train.train_model import train
  File "/home/daniel/PycharmProjects/project/src/train/", line 17, in <module>
    learn = load_learner('yasmine-sftp/export_2.pkl', cpu=True)  # to run on GPU
  File "/home/daniel/miniconda3/envs/project/lib/python3.6/site-packages/fastai/", line 384, in load_learner
    res = torch.load(fname, map_location='cpu' if cpu else None, pickle_module=pickle_module)
  File "/home/daniel/miniconda3/envs/project/lib/python3.6/site-packages/torch/", line 607, in load
    return _load(opened_zipfile, map_location, pickle_module, **pickle_load_args)
  File "/home/daniel/miniconda3/envs/project/lib/python3.6/site-packages/torch/", line 882, in _load
    result = unpickler.load()
  File "/home/daniel/miniconda3/envs/project/lib/python3.6/site-packages/torch/", line 875, in find_class
    return super().find_class(mod_name, name)
AttributeError: Can't get attribute 'Tf' on <module '__main__' from ''>

This is because in order to open the .pkl file, I need the original function that was used to train it.

Thankfully, looking back at the notebook, Tf(o) is there:

def Tf(o):
    return '/mnt/scratch2/DLinTHDP/PathLAKE/Version_4_fastai/Dataset/CD8/Train/masks/'+f'{o.stem}_P{o.suffix}'

However, anywhere I place Tf(o) in my Python scripts I still get the same error.

Where should I put Tf(o)?

In error message: <module '__main__' from ''> seems to suggest to put it in main() or under if __name__ ....

I’ve tried everywhere. Importing Tf(o) also doesn’t work.

Python Scripts

import glob
from pathlib import Path

from train_model import train

ROOT = Path("folder/path")  # Detection Folder

def main(root: Path):

if __name__ == '__main__':

from pathlib import Path

from import *

folder_path = Path('.')

learn = load_learner('model.pkl', cpu=True)  # AttributeError
learn.load('model_3C_34_CELW_V_1.1')  # weights

def train(root: Path):
    # ...

I cannot inspect the file:

(project) daniel@ubuntu-pcs:~/PycharmProjects/project$ python -m pickletools -a model.pkl
Traceback (most recent call last):
  File "/home/daniel/miniconda3/envs/project/lib/python3.6/", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/home/daniel/miniconda3/envs/project/lib/python3.6/", line 85, in _run_code
    exec(code, run_globals)
  File "/home/daniel/miniconda3/envs/project/lib/python3.6/", line 2830, in <module>
    args.indentlevel, annotate)
  File "/home/daniel/miniconda3/envs/project/lib/python3.6/", line 2394, in dis
    for opcode, arg, pos in genops(pickle):
  File "/home/daniel/miniconda3/envs/project/lib/python3.6/", line 2242, in _genops
    arg = opcode.arg.reader(data)
  File "/home/daniel/miniconda3/envs/project/lib/python3.6/", line 373, in read_stringnl_noescape
    return read_stringnl(f, stripquotes=False)
  File "/home/daniel/miniconda3/envs/project/lib/python3.6/", line 359, in read_stringnl
    data = codecs.escape_decode(data)[0].decode("ascii")
UnicodeDecodeError: 'ascii' codec can't decode byte 0x80 in position 63: ordinal not in range(128)

Still a ML novice, but I can suggest a general troubleshooting approach to individually deal with separate concerns…

  1. Start with a single file that does both training and inferencing, without saving/loading a model
  2. Between the training and inferencing code insert two lines (the middle of which we’ll call Point A), that exports the learner then immediately reloads it.
  3. Duplicate the file, then working upwards from Point A slowly delete training-code until you learn what breaks the inferencing-code.


Why I get that error is bc the Tf() function was used to train the model.pkl file, in the same namespace (because it was done in a notebook file).

This article states:

pickle is lazy and does not serialize class definitions or function
definitions. Instead it saves a reference of how to find the class
(the module it lives in and its name)


There’s an extension to pickle called dill, that does serialise Python objects and functions etc. (not references) PyPI

1 Like