Stable Diffusion WebUI 图生图 DeepBooru反推提示词 结果翻车了_blip反推 bert-base-uncased stable-diffusion-webui

图生图界面，上传图片，点击“DeepBooru反推提示词”

 图形界面没有反应，终端日志

Downloading: "https://github.com/AUTOMATIC1111/TorchDeepDanbooru/releases/download/v1/model-resnet_custom_v3.pt" to /home/yeqiang/Downloads/ai/stable-diffusion-webui/models/torch_deepdanbooru/model-resnet_custom_v3.pt██████████████████████████| 16/16 [00:03<00:00,  5.84it/s]
 
Traceback (most recent call last):
  File "/home/yeqiang/Downloads/ai/stable-diffusion-webui/venv/lib/python3.10/site-packages/gradio/routes.py", line 422, in run_predict
    output = await app.get_blocks().process_api(
  File "/home/yeqiang/Downloads/ai/stable-diffusion-webui/venv/lib/python3.10/site-packages/gradio/blocks.py", line 1323, in process_api
    result = await self.call_function(
  File "/home/yeqiang/Downloads/ai/stable-diffusion-webui/venv/lib/python3.10/site-packages/gradio/blocks.py", line 1051, in call_function
    prediction = await anyio.to_thread.run_sync(
  File "/home/yeqiang/Downloads/ai/stable-diffusion-webui/venv/lib/python3.10/site-packages/anyio/to_thread.py", line 33, in run_sync
    return await get_asynclib().run_sync_in_worker_thread(
  File "/home/yeqiang/Downloads/ai/stable-diffusion-webui/venv/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 877, in run_sync_in_worker_thread
    return await future
  File "/home/yeqiang/Downloads/ai/stable-diffusion-webui/venv/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 807, in run
    result = context.run(func, *args)
  File "/home/yeqiang/Downloads/ai/stable-diffusion-webui/modules/ui.py", line 994, in <lambda>
    fn=lambda *args: process_interrogate(interrogate_deepbooru, *args),
  File "/home/yeqiang/Downloads/ai/stable-diffusion-webui/modules/ui.py", line 142, in process_interrogate
    return [interrogation_function(ii_singles[mode]), None]
  File "/home/yeqiang/Downloads/ai/stable-diffusion-webui/modules/ui.py", line 169, in interrogate_deepbooru
    prompt = deepbooru.model.tag(image)
  File "/home/yeqiang/Downloads/ai/stable-diffusion-webui/modules/deepbooru.py", line 43, in tag
    self.start()
  File "/home/yeqiang/Downloads/ai/stable-diffusion-webui/modules/deepbooru.py", line 34, in start
    self.load()
  File "/home/yeqiang/Downloads/ai/stable-diffusion-webui/modules/deepbooru.py", line 28, in load
    self.model.load_state_dict(torch.load(files[0], map_location="cpu"))
IndexError: list index out of range

github网络不通导致

手动处理

cd /home/yeqiang/Downloads/ai/stable-diffusion-webui/models/torch_deepdanbooru/
wget https://ghproxy.com/https://github.com/AUTOMATIC1111/TorchDeepDanbooru/releases/download/v1/model-resnet_custom_v3.pt

日志

yeqiang@yeqiang-MS-7B23:~/Downloads/ai/stable-diffusion-webui/models/torch_deepdanbooru$ wget https://ghproxy.com/https://github.com/AUTOMATIC1111/TorchDeepDanbooru/releases/download/v1/model-resnet_custom_v3.pt
--2023-07-07 15:11:11--  https://ghproxy.com/https://github.com/AUTOMATIC1111/TorchDeepDanbooru/releases/download/v1/model-resnet_custom_v3.pt
正在解析主机 ghproxy.com (ghproxy.com)... 144.24.81.189
正在连接 ghproxy.com (ghproxy.com)|144.24.81.189|:443... 已连接。
已发出 HTTP 请求，正在等待回应... 200 OK
长度： 644144887 (614M) [application/octet-stream]
正在保存至: ‘model-resnet_custom_v3.pt’
 
model-resnet_custom_v3.pt     100%[=================================================>] 614.30M  2.86MB/s    用时 3m 27s
 
2023-07-07 15:14:40 (2.97 MB/s) - 已保存 ‘model-resnet_custom_v3.pt’ [644144887/644144887])

成功获取

1girl, 3d, architecture, blurry, blurry background, blurry foreground, breasts, brown hair, building, cherry blossoms, city, cityscape, cosplay photo, cowboy shot, day, depth of field, east asian architecture, flower, lips, long hair, looking at viewer, medium breasts, midriff, motion blur, navel, outdoors, photo \(medium\), photo background, pink flower, scrunchie, shirt, shorts, skyscraper, solo, street, tree, white shirt

Stable Diffusion WebUI 集成 LoRA模型，给自己做一张壁纸 Ubuntu22.04 rtx2060 6G_hkNaruto的博客-CSDN博客

参考上文，下载模型，添加上LoRA模型参数，名称参考如下。

效果不理想。。。翻车了

抢救一下

正向提示词

1girl, 3d, architecture, blurry, blurry background, blurry foreground, breasts, brown hair, building, cherry blossoms, city, cityscape, cosplay photo, cowboy shot, day, depth of field, east asian architecture, flower, lips, long hair, looking at viewer, medium breasts, midriff, motion blur, navel, outdoors, photo \(medium\), photo background, pink flower, scrunchie, shirt, shorts, skyscraper, solo, street, tree, white shirt,best quality, ultra high res, (photorealistic:1.4), 1girl,   <lora:koreanDollLikeness_v15:1> ,<lora:yaeMikoRealistic_yaemikoMixed:1>,

反向提示词

paintings, sketches, (worst quality:2),(low quality:2),(normal quality:2), lowres, normal quality, ((monochrome)), ((grayscale)), skin spots, acnes, skin blemishes, age spot, glans,{{{bad_hands}}},badhand, bad_feet,

注意：采样方法选择DPM++ SDE Karras，影响较大！

再次出图

 过得去了