When passing a model_path to the FeatureExtractor, weights are loaded twice. One the pretrained once and then the given ones. Avoid the unnecessary load. |
||
---|---|---|
.. | ||
data | ||
engine | ||
losses | ||
metrics | ||
models | ||
optim | ||
utils | ||
__init__.py |
When passing a model_path to the FeatureExtractor, weights are loaded twice. One the pretrained once and then the given ones. Avoid the unnecessary load. |
||
---|---|---|
.. | ||
data | ||
engine | ||
losses | ||
metrics | ||
models | ||
optim | ||
utils | ||
__init__.py |