finished docs apart from examples. Changed name of DeepPreprocessor to DensePreprocessor for consistency. Added raise error if model components do not have an 'output_dim' attribute

Deal with when either deeptext or deepimage does not exist.

fixed bug related to the activation function in the case of multiclass classification. F.cross_entropy already applies logSoftmax

refined the documentation for the WarmUp class. Fix an issue regarding to the model type. Added a test for the warm up functionalities

refined documentation in the WideDeep class. Change the reduction method for the FocalLoss from sum to mean

improved deep_dense syntax. Replaced Adam with AdamW as default. Added n_cpus to n_workers in Dataloader

Added support for sparse matrices. Added the possibility of using a head for the deep branch. Started with the examples

changed order of input params in a number of methods to sensical order. Also fixed a number of bugs in the predict methods