Model initialized with `skip_model_load` does not have a compiled path
- Dominant language
- Python
- Stars
- 5.4k
- Forks
- 850
- Avg merge
- 4d 5h
- Merged PRs (30d)
- 10
Description
Code:
```python
import coremltools as ct
m = ct.models.MLModel("", skip_model_load=True)
m.get_compiled_model_path()
```
Outputs:
```
----> 1 m.get_compiled_model_path()
File ~/miniforge3/envs/prod/lib/python3.8/site-packages/coremltools/models/model.py:528, in MLModel.get_compiled_model_path(self)
520 def get_compiled_model_path(self):
521 """
522 Returns the path for the underlying compiled ML Model.
523
(...)
526
527 """
--> 528 return self.__proxy__.get_compiled_model_path()
AttributeError: 'NoneType' object has no attribute 'get_compiled_model_path'
```
Correct behavior:
Return `None` without error.
Contributor guide
Research direction
Start in coremltools/models/model.py at MLModel.get_compiled_model_path(), then trace how skip_model_load=True leaves __proxy__ unset. Reproduce the shown call and confirm that it returns None without raising an AttributeError.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- machine-learning
- Issue type
- Bug
- Difficulty
- 2/5
- Estimated time
- 1-3 hours
- Activity status
- Stale
- Clarity
- Clearly specified
- Newbie friendliness
- 48/100