A reimplementation of the MSCRED model in Keras, without the Attention layers (since they were not implemented in Keras during the time). A full reimplementation and application experiment could be found here:
Since its a while ago, I cannot provide any testing, Keras version, documentation, maintenance, input/output data for testing here at the moment. I might in the future if the need has increased. Hopefully it serves as a reference to help your own implementations. Tell me if theres anything missing, but I make no promises.
NOTE! that the MIT part of the license refers to my works reimplementing the original MSCRED in Keras. I do not claim any license rights over any rights that were preserved by the original authers, see https://github.com/7fantasysz/MSCRED