生成的都是噪点图

[TimeWings]

我clone repo之后，把models文件夹复制进去，不管是txt2img还是img2img，跑出来的都是噪点图 中途还报了缺少sklearn模块的错误，不过我自己解决了，不知道有没有影响

[sotokisehiro]

I'm having a similar issue in my environment. (There is no problem with the AUTOMATIC1111 version)

[JC-Array]

I was able to replicate the results for this image: https://github.com/7eu7d7/APT-stable-diffusion-auto-prompt/blob/master/imgs/p1.png?raw=true

By copying the image into PNG Info here is the following parameters: masterpiece, best quality, ptun-d Negative prompt: ptun-d-uc, lowres, bad anatomy, bad hands, text, error, missing fingers, extra fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry Steps: 25, Sampler: Euler a, CFG scale: 7, Seed: 1316503418, Size: 768x512, Model hash: e6e8e1fc

Running this on AUTOMATIC1111's version I get the following image:

However I was about to recreate it properly on this repo and product similar images.

If you are getting a static image check that your using the correct model. Hash: e6e8e1fc. Aka NAI's animefull-latest.

[MEIZU16]

同样的问题

[IrisRainbowNeko]

我的模型是在hq done上练的，如果用不一样的可能出bug

[zozyo]

# sha256sum novelaileakpt2/random_stableckpt/animefullhq-done/model.ckpt 
a7529df02340e5b4c3870c894c1ae84f22ea7b37fd0633e5bacfad9618228032  novelaileakpt2/random_stableckpt/animefullhq-done/model.ckpt
# sha256sum novelaileakpt2/random_stableckpt/animefull-latest/model.ckpt 
a7529df02340e5b4c3870c894c1ae84f22ea7b37fd0633e5bacfad9618228032  novelaileakpt2/random_stableckpt/animefull-latest/model.ckpt

no difference

[fenglindubu]

ModuleNotFoundError: No module named 'sklearn' pip3 install scikit-learn我运行这个还是报错啊，请问怎么解决的

[TimeWings]

我的模型是在hq done上练的，如果用不一样的可能出bug

我换成part2的animefullhq-done成功出图了，出一次图后再换成其他模型也正常出图了，神奇

[TimeWings]

ModuleNotFoundError: No module named 'sklearn' pip3 install scikit-learn我运行这个还是报错啊，请问怎么解决的

安装sklearn的依赖，在venv/Scripts/里打开命令行运行python -m pip install scikit-learn

[sotokisehiro]

我的模型是在hq done上练的，如果用不一样的可能出bug

Thank you for the advice. After changing the model, I was able to confirm that the image can be generated. I'm trying "Prompt Thuing" and it works great!