Closed DenSckriva closed 2 months ago
Little update If i have in prompt : "score_9, score_8_up, score_7_up, score_6_up, masterpiece" and negative : "bad-artist, boring_e621, lowres, bad anatomy, text, error, cropped, (low quality, worst quality:1.4), normal quality, jpeg artifacts, signature, watermark, username, blurry, monochrome, human,"
The module work perfectly.
With the same prompt but in negative : "rt by bad-artist, lowres, text, error, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry, bad anatomy, bad hands, ugly, morbid, extra fingers, missing fingers, extra digits, poorly drawn hands, mutation, extra limbs, gross proportions, missing arms, mutated hands, long neck, mutilated, mutilated hands, poorly drawn face, deformed, bad anatomy, malformed limbs, missing legs, fewer digits, extra tails, extra limbs, disembodied tail, EasyNegativeV2"
Not work, ValueError: Incompatible shapes for attention inputs: blabla.
A rapid test show me a token limite in prompt (and neg prompt) at 75. <=75 it's OK >75 nope. I don't know why at this moment
I have same issue as you, and i follow this disable "Pad prompt/negative prompt to be same length" now its work, not sure if this work same for you
I've found a definitive solution, to this issue, open config.json
and set "batch_cond_uncond"
on true
Hello !
Maybe you can help me with my error. Since updating automatic1111, I have been unable to generate an image with this tool.
"ValueError: Incompatible shapes for attention inputs: query.shape: torch.Size([2, 12288, 8, 40]) key.shape : torch.Size([1, 77, 8, 40]) value.shape: torch.Size([1, 77, 8, 40]) HINT: We don't support broadcasting, please use
expand
yourself before callingmemory_efficient_attention
if you need to"I have try with and without xformer, same problem. :(
The complete console message :
Thanks for your help