search

AlexKuhnle / ShapeWorld

MIT License
58 stars 18 forks source link

Reading pre-loaded datasets tries to open `world_features` unsuccessfully #9

Closed tomsherborne closed 6 years ago

tomsherborne commented 6 years ago

[Not urgent]

Hi Alex,

I've noticed that if you open a dataset that has saved the ResNet features using generate() not tf_util.batch_records() then there is an exception thrown as deserialize_value() tries to open world_features.txt (the offending line is 436 in dataset.py). This doesn't seem to be a useful file as the full features aren't saved in non-TFRecord form so a condition to ignore them could be included.

I know you require the generation of TFRecords for the ResNet features but in cases such as designing a manual object detector (such as what I am doing now) and I need the world_model I have to use the generate() call.

Hope this helps!

AlexKuhnle commented 6 years ago

This should be fixed now, however, it's not (yet) optimal, since the data is serialized as a txt file, which for these feature files doesn't make so much sense. Anyway, I'll close this for now, and might add some improvements later on.