prompt example being truncated

[khimaros]

with stable-diffusion-v1-5, the default tokenizer only allows 77 tokens. are you having desired results with the prompt included in your wanostyle example? i'm seeing this warning and poor image results:

The following part of your input was truncated because CLIP can only handle sequences up to 77 tokens: ['_ offset : 1 >, closed shirt, anime screencap, scar under eye, ready to fight, black eyes']

[haofanwang]

This is the limitation of fixed CLIP text encoder. Try a more compact prompt.