Hans Gaiser
Hans Gaiser
Hopefully https://github.com/dbrgn/tealdeer/pull/212 will fix this issue :).
Are there any plans for this? Would be nice to implement it in [keras-retinanent](https://github.com/fizyr/keras-retinanet) :)
`load_model` is implemented in `keras.models`. Try: ``` model = keras.models.load_model(model_path) ```
Why not wrap the model in a new model so that it appears as one big layer? You can name that model whatever you want then. On Sun, 2 Dec...
@0x00b1 , any feedback on this?
What is the status of this PR? I am currently looking at workarounds but this solution would be highly preferred.
I am doing something similar. It isn't pretty, but it doesn't require an extra tool: ```shell printf '[' > compile_commands.json find ./build_isolated -type f -name 'compile_commands.json' -exec sh -c "cat...
Hm yeah I agree, it makes sense for `SubImage` to be `Clone` and `Copy` :). I submitted a PR to fix this issue: https://github.com/image-rs/image/pull/1582.
This is actually addressed in https://github.com/broadinstitute/keras-rcnn/pull/71
So there'd be one layer computing the RPN loss and one layer to compute the RCNN loss? I like that idea!