I have created this LoRa for personal purposes but I have decided to share it with the community in case it might be of help to any artists struggling to come up with new ideas for drawings.

This model is meant to recreate the feeling of 'thumbnails' or rough sketches with very loose details focusing more on the shapes rather than the details. In other words, it's meant to recreate the very early stages of a drawing. If your image comes out without fingers or such 'imperfections' then it did it's job right. Must of the time the generated image will be in greyscale. The best way to get this result in my opinion is by not using negative prompts nor any embedding meant to help the image look 'better'.

I would say to avoid altogether the use of 'color' tags, such as 'blonde hair', 'red socks', 'blue eyes', etc. although sometimes it would be fine even if they are in your prompt as you can see in the showcase gallery.

It should work just fine with "Character" LoRas although I didn't test it with any.

I couldn't decide between 2 epochs. One makes more loose rough sketches, which was my initial goal, and the other makes a bit more 'refined' sketch but they are sketches none the less.

Activation Word (both): R0u6hSk3tch, sketch

Weight V1a: 1.1

Weight V1b: 1.0

Now please test this LoRa and show your results in the gallery down below. Merry Christmas! Hoe! Hoe! Hoe!