Using callback on adapt_fit yields `TypeError: expected str, bytes or os.PathLike object, not NoneType`

[icarosadero]

Is there an existing issue for this?

• I have searched the existing issues

Bug description

If you define a function on the callback of the method adapt_fit along with a frequency, an error will show up mentioning that there is no logdir specified to save checkpoints on the save method of the solver.base. Example:

progress_bar = lambda num_steps, solver: print(f"Embedding... {num_steps}")
U = embedding.fit_transform(X, adapt=True, callback_frequency=100, callback=progress_bar)

Results in:

Traceback (most recent call last):
  File "/project/.venv/lib/python3.12/site-packages/streamlit/runtime/scriptrunner/exec_code.py", line 88, in exec_func_with_error_handling
    result = func()
             ^^^^^^
  File "/project/.venv/lib/python3.12/site-packages/streamlit/runtime/scriptrunner/script_runner.py", line 579, in code_to_exec
    exec(code, module.__dict__)
  File "/project/app.py", line 119, in <module>
    main()
  File "/project/app.py", line 78, in main
    U = embedding.fit_transform(X, adapt=True, callback_frequency=100, callback=progress_bar)
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/project/.venv/lib/python3.12/site-packages/sklearn/utils/_set_output.py", line 157, in wrapped
    data_to_wrap = f(self, X, *args, **kwargs)
                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/project/.venv/lib/python3.12/site-packages/cebra/integrations/sklearn/cebra.py", line 1289, in fit_transform
    self.fit(X,
  File "/project/.venv/lib/python3.12/site-packages/cebra/integrations/sklearn/cebra.py", line 1188, in fit
    self._adapt_fit(X,
  File "/project/.venv/lib/python3.12/site-packages/cebra/integrations/sklearn/cebra.py", line 1127, in _adapt_fit
    self._partial_fit(*self.state_,
  File "/project/.venv/lib/python3.12/site-packages/cebra/integrations/sklearn/cebra.py", line 1036, in _partial_fit
    solver.fit(
  File "/project/.venv/lib/python3.12/site-packages/cebra/solver/base.py", line 214, in fit
    self.save(logdir, f"checkpoint_{num_steps:#07d}.pth")
  File "/project/.venv/lib/python3.12/site-packages/cebra/solver/base.py", line 346, in save
    if not os.path.exists(os.path.dirname(logdir)):
                          ^^^^^^^^^^^^^^^^^^^^^^^
  File "<frozen posixpath>", line 181, in dirname
TypeError: expected str, bytes or os.PathLike object, not NoneType

Following the error message, the call to solver.fit is setting logdir = None, which is then passed to base.save, which then tried to save in None.

Operating System

Kubuntu 24.04

CEBRA version

0.4.0

Device type

GPU

Code of Conduct

• I agree to follow this project's Code of Conduct

[MMathisLab]

Can you pip install 0.5.0rc1 and see if error persists?

[icarosadero]

Can you pip install 0.5.0rc1 and see if error persists?

Just did, seems like it's the same error. Is cebra.__version__ supposed to be empty in this version?

[MMathisLab]

hey @icarosadero

• okay thanks for checking, we will address, thanks for PR Update base.py #224 !

• version working for me; what the error?

  CEBRA/cebra/__init__.py

  Line 69 in c32ed67

  __version__ = "0.5.0rc1"

[icarosadero]

hey @icarosadero

• okay thanks for checking, we will address, thanks for PR Update base.py #224 !

• version working for me; what the error?

    [CEBRA/cebra/__init__.py](https://github.com/AdaptiveMotorControlLab/CEBRA/blob/c32ed67ab8bc1ff1afa4e47e1d09bb26ce0a3270/cebra/__init__.py#L69)
  
  
       Line 69
    in
    [c32ed67](/AdaptiveMotorControlLab/CEBRA/commit/c32ed67ab8bc1ff1afa4e47e1d09bb26ce0a3270)
  
  
  
  
  
      
        
         __version__ = "0.5.0rc1"
  

Nevermind, I bet it is a problem with my REPL