add sd-webui api doc and script #559
Conversation
onediff_sd_webui_extensions/api.py
Outdated
| # Enable OneDiff speed up | ||
| "script_name": "onediff_diffusion_model", | ||
| # If you are using OneDiff Enterprise, add the field below to enable quant feature | ||
| # "script_args" : [{"0": True}], |
There was a problem hiding this comment.
增加这里的字段就可以使用 OneDiff 的加速功能，下面的 script_args : [{"0": True}] 是开启量化功能
onediff_sd_webui_extensions/api.py
Outdated
| } | ||
| call_txt2img_api(**payload) | ||
|
|
||
| # OneDiff sd-webui doesn't support img2img now, the feature is under urgent development |
There was a problem hiding this comment.
OneDiff sd-webui 似乎不支持 img2img，所以这里先注释掉了
There was a problem hiding this comment.
直接删除吧。 img2img 之后另外准备一个新的脚本
onediff_sd_webui_extensions/api.py
Outdated
| @@ -0,0 +1,170 @@ | |||
| from datetime import datetime | |||
There was a problem hiding this comment.
新建个 api 目录吧
新建了一个 api_examples 目录
onediff_sd_webui_extensions/api.py
Outdated
| @@ -0,0 +1,170 @@ | |||
| from datetime import datetime | |||
There was a problem hiding this comment.
这个作为 example，单独放到一个 examples 目录吧。现在是 txt2img 的例子，未来还有 img2img, controlnet 的例子
onediff_sd_webui_extensions/api.py
Outdated
| @@ -0,0 +1,170 @@ | |||
| from datetime import datetime | |||
| import urllib.request | |||
There was a problem hiding this comment.
这里注释加一下原始来源吧，方便以后维护这个脚本的人定位更新。
而且可以注释里，把因为适配 onediff 改动的地方强调下。让后面维护的人，以及用户，都知道改动不用特别多、怎么改。
| @@ -98,6 +98,28 @@ If you possess a OneDiff Enterprise license key, you can access instructions on | |||
|
|
|||
| Note: The quantified model doesn't currently support LoRA loading, and related features are under emergency development. | |||
|
|
|||
| ## API | |||
There was a problem hiding this comment.
| ## API | |
| ## Invoke OneDiff by API |
”用 api 方式调用onediff"，具体英文是否正确你再确认下吧。
| To use the OneDiff-based sd-webui API, you only need to add a `"script": "onediff_diffusion_model"` field to the request to speed up model inference. | ||
|
|
||
| For OneDiff Enterprise, add `"script_args" : [{"0": True}]` to use the quantization feature. |
There was a problem hiding this comment.
我觉得这些东西放 py 脚本里，作为注释就可以了
There was a problem hiding this comment.
我是这么想的，对于经常用 api 调用 sd-webui 的用户，看到文档里这么写就能明白怎么用了，省去了一步扒代码的过程
There was a problem hiding this comment.
那也可以保留。不过我建议换一下顺序：先介绍如何跑起来。然后再解释在干什么。
因为大部分看这个文档的，是想要 “howto”，就是照着就能复现，复现成功了，再看能干嘛
|
|
||
| For OneDiff Enterprise, add `"script_args" : [{"0": True}]` to use the quantization feature. | ||
|
|
||
| Check file `extensions/onediff_sd_webui_extensions/api.py` for more details. |
There was a problem hiding this comment.
| Check file `extensions/onediff_sd_webui_extensions/api.py` for more details. | |
| Check file [api.py](./api.py) for more details. |
改成超链接（这个路径只是当前的例子。需要合并前根据实际情况更新）
|
|
||
| Then you can get the images returned by sd-webui client at `api_out/txt2img/`. | ||
|
|
||
| Note: OneDiff sd-webui only supports txt2img now, img2img and ControlNet is under urgent development. |
onediff_sd_webui_extensions/api.py
Outdated
| # If you are using OneDiff Enterprise, add the field below to enable quant feature | ||
| # "script_args" : [{"0": True}], | ||
|
|
||
| # example args for x/y/z plot |
There was a problem hiding this comment.
这些是原来脚本就有的注释吧。
不知道有特殊意义没有，为了干净可以删掉？
onediff_sd_webui_extensions/api.py
Outdated
| } | ||
| call_txt2img_api(**payload) | ||
|
|
||
| # OneDiff sd-webui doesn't support img2img now, the feature is under urgent development |
There was a problem hiding this comment.
直接删除吧。 img2img 之后另外准备一个新的脚本
api 代码参考了 sd-webui 的 wiki：https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/API ，改动是在 request 的 data 中增加了对应的字段。
由于现在 OneDiff 还不支持 img2img，所以注释掉了脚本中对应的部分。