Textual Inversion: Added custom training image size and number of repeats per input image in a single epoch

Pytorch 1.13 and later will rename _TypedStorage to TypedStorage, so check for TypedStorage and use _TypedStorage if it is not available. Currently this is needed so that nightly builds of PyTorch work correctly.

Updated txt2img screenshot (UI as of Oct 9th) for github webui / README.md

* add updateOnBackgroundChange
* typo fixes.
* reindent to 4 spaces

Fixes #1049

When using the 'Sampler' or 'Checkpoint' options, if one of the entered
names has a typo, an error will only be thrown once the `draw_xy_grid`
loop reaches that name. This can waste a lot of time for large grids
with a typo near the end of a list, since the script needs to start over
and re-generate any earlier images to finish making the grid.

Also fixing typo in variable name in `draw_xy_grid`.

This reverts commit e2930f98.