search

deel-ai / oodeel

Simple, compact, and hackable post-hoc deep OOD detection for already trained tensorflow or pytorch image classifiers.
https://deel-ai.github.io/oodeel/
MIT License
52 stars 2 forks source link

Homogenize preprocess_fn between torch and tf #56

Closed y-prudent closed 1 year ago

y-prudent commented 1 year ago

Update

Now preprocess_fn provided as an argument of OODDataset.prepare method need to be defined for both frameworks with:

def preprocess_fn(*inputs):
    # (...)
    return tuple(...)

(as it was the case for tf before)

y-prudent commented 1 year ago

@paulnovello @cofri Should we remove the augment_fn which seem redundant with preprocess_fn? If a user wants some DA for training only, he can do it through a preprocess_fn specific to the training dataset.

github-actions[bot] commented 1 year ago

☂️ Python Coverage

current status: ✅

Overall Coverage

Lines Covered Coverage Threshold Status
1615 1453 90% 70% 🟢

New Files

No new covered files...

Modified Files

File Coverage Status
oodeel/datasets/torch_data_handler.py 89% 🟢
TOTAL 89% 🟢

updated for commit: 29f6787 by action🐍