Mathias Lechner

Hi @ZhangMengEnAPI I also get a 404. I think @raminmh knows what happened with the repo.

This does not seem to a bug in keras-ncp but may originate somewhere else, because I cannot reproduce the error in colab: https://colab.research.google.com/drive/10101d3RuRTJqStSlJpTcPtFKQFCkgBjQ?usp=sharing

I have no knowledge about TF with Metal. Colab (w and w/o GPU) is a good starting point to get an estimate of what training times to expect. The training...

hi @zengxiang01 Thanks for reporting the bug. It seems that the bug is caused by the use of an LTC model inside a ```tf.keras.layers.Bidirectional``` module. I will be working on...

Hi @zengxiang01 I have pushed a potential fix (keras-ncp version 2.0.7). Can you please update to this version and test your code?

Hi @sprakashdash Glad that you liked our work! There are a few issues in the ```QNetwork_w_LTC```: First, ```in_features``` are the size of the input tensor (=number of input features), which...

The LTC usually requires higher learning rates than feedforward networks. So, in our examples, 0.01 or 0.005 (autonomous driving) worked okay for supervised learning. As you mentioned, RL typically works...

Hi, ```[batch size, 1, 256]``` is the correct shape, i.e., a batch of length 1 sequences, each having 256 features. The output tensor of the LTC network should be then...