Adding additional functionality for plotting 2D PCA on learned embeddings and representations in Int2Int https://github.com/f-charton/Int2Int

Parameters to add to parser

--pca_id 

as the name of the experiment to be added to the plot

--pca_plot

as either 0 (default) or 1. Choose 1 for plotting some 2D PCA with arguments to follow

--pca_initial

as either 0 (default) or 1. Choose 1 for plotting the initial embedding, otherwise 0 for plotting hidden state representations.

--pca_layer

If pca_initial is 0, then pick which layer. From 1 to 3 for intermediate layers, -1 to default to the last layer.

--pca_legend

as either 0 (default) or 1. Choose 1 to add a legend to your plot that specifies each separate color.

--pca_labels

Determines the coloring of labels in the PCA plot. By default, labels are colored using a rainbow gradient:

• Based on tokens if pca_initial == 1
• Based on input indices when plotting hidden states (pca_initial == 0)
• To use custom label coloring and manually group labels into clusters with the same color, provide a path to a JSON file containing a dictionary:
  • Keys: Integers in the range [0, n-1], where n is the number of colors.
  • Values: Lists of either:
    • Words (if pca_initial == 1) to assign the same color to related words.
    • Input indices (if pca_initial == 0) to assign the same color to specific input positions.