Willi Gierke

`KerasLayer` returns one layer that cannot be split into more layers. Instead, I'd recommend using the accompanying feature vector model (https://tfhub.dev/google/imagenet/resnet_v1_50/feature_vector/5) which does not contain the top layer. Thus, you...

From a quick glance, it looks like your input layers are 1D (`tf.keras.Input(shape=(128,)`) while the model expects 2D inputs: ```Option 3: Positional arguments (3 total): * {'input_type_ids': TensorSpec(shape=(None, None), ......

It seems like the PR would currently try to add two new Colabs. Could you remove them and only modify the tf2_object_detection Colab by fixing the typos?

Cool! :) Maybe [this](https://github.com/WGierke/git_better/blob/99c7d85e6b781cd0deb976c9642499516de42c4e/server/django_server/utils.py) can help you.

[Top-tier conferences](http://www.guide2research.com/topconf/) such as CVPR, NIPS, ECCV, ICML, ICCV, ICLR ... would be the most interesting to everyone I guess :) I'd also be interested in papers with code of...

Once #237 is merged, I'd like to work on "segmentation test(s) should pass for luna scans".

I took a quick glance at the identification problems. When running the tests, I get a `RuntimeError: $ Torch: not enough memory: you tried to allocate 21GB. Buy new RAM!...

@reubano could you do me a favor and check out my branch [229_fix_segmentation](https://github.com/concept-to-clinic/concept-to-clinic/compare/master...WGierke:229_fix_segmentation?expand=1)? All I did was to apply the patch for the segmentation tests and merge the current master....

@isms I can understand that RAM usage is not a hard constraint. However, this makes it impossible for me (and others with less than 32GB RAM) to debug (and improve...

@isms Alright, thanks a lot for the insights and the feedback. I'll have a look at how I can solve this problem for me. Thanks!