Saving trained models
xloffree opened this issue · 1 comments
Hi,
I have asked about this in a previous, closed issue (#14), but the solution given is not working.
Has this solution worked for you with PiML? When I try this solution, this line:
with open('name_model.pkl', 'rb') as file: clf_load = dill.load(file)
results in a recursion error every time. I tried changing the recursion limit but even with a recursion limit pf 10000000 I still run into this error. Increasing the recursion limit indefinitely just causes the kernel to crash.
The error is as follows:
RecursionError Traceback (most recent call last)
Cell In [39], line 2
1 with open('name_model.pkl', 'rb') as file:
----> 2 clf_load = dill.load(file)
File /n/holylfs05/LABS/liang_lab_l3/Lab/piml_py39_shared/lib/python3.9/site-packages/dill/_dill.py:272, in load(file, ignore, **kwds)
266 def load(file, ignore=None, **kwds):
267 """
268 Unpickle an object from a file.
269
270 See :func:loads for keyword arguments.
271 """
--> 272 return Unpickler(file, ignore=ignore, **kwds).load()
File /n/holylfs05/LABS/liang_lab_l3/Lab/piml_py39_shared/lib/python3.9/site-packages/dill/_dill.py:419, in Unpickler.load(self)
418 def load(self): #NOTE: if settings change, need to update attributes
--> 419 obj = StockUnpickler.load(self)
420 if type(obj).module == getattr(_main_module, 'name', 'main'):
421 if not self._ignore:
422 # point obj class to main
File piml/models/glm.py:32, in piml.models.glm.GLMRegressor.getattr()
File piml/models/glm.py:32, in piml.models.glm.GLMRegressor.getattr()
[... skipping similar frames: piml.models.glm.GLMRegressor.getattr at line 32 (9999967 times)]
File piml/models/glm.py:32, in piml.models.glm.GLMRegressor.getattr()
RecursionError: maximum recursion depth exceeded while calling a Python object
Any help with this would be very appreciated. When using PiML for research purposes, being able to save a trained model is essential for reproducibility. Thank you!