Implement W2V model

[tdrose]

As an additional model for the package. Use previous work from Dominik for the dataloaders.

Try to incorporate intensities instead of just thresholds.

[tdrose]

Information for the implementation:

• http://mccormickml.com/2017/01/11/word2vec-tutorial-part-2-negative-sampling/
• https://towardsdatascience.com/word2vec-with-pytorch-implementing-original-paper-2cd7040120b0
• https://pytorch.org/tutorials/beginner/nlp/word_embeddings_tutorial.html

[tdrose]

• Implement Patch Dataset dataloader (using RandomCrop, or sample only from tissue slice region (only pixels that have at least one non-zero ionimage))
• Patched image parameter optimization
• Make new evaluation metrics to better explore/understand model performance
• Evaluate on Different Dataloader scenarios