extern/easydiffusion
1
0
Fork 1
mirror of https://github.com/easydiffusion/easydiffusion.git synced 2025-02-03 03:59:20 +01:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #229 from cmdr2/beta

Browse Source 
Live Preview, In-Painting Editor, New Sampler Options, VRAM (Memory) Usage Improvements, Progress Bar, Re-organized UI with more space for images, Update to the latest fork version
This commit is contained in:
cmdr2 2022-09-23 20:31:21 +05:30 committed by GitHub
commit a35454b1b3
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
12 changed files with 956 additions and 149 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
scripts/on_sd_start.bat
View File

@ -15,7 +15,7 @@
@call git reset --hard @call git reset --hard
@call git pull @call git pull
@call git checkout d154155d4c0b43e13ec1f00eb72b7ff9d522fcf9 @call git checkout f6cfebffa752ee11a7b07497b8529d5971de916c
@call git apply ..\ui\sd_internal\ddim_callback.patch @call git apply ..\ui\sd_internal\ddim_callback.patch
@call git apply ..\ui\sd_internal\env_yaml.patch @call git apply ..\ui\sd_internal\env_yaml.patch
@ -33,7 +33,7 @@
) )
@cd stable-diffusion @cd stable-diffusion
@call git checkout d154155d4c0b43e13ec1f00eb72b7ff9d522fcf9 @call git checkout f6cfebffa752ee11a7b07497b8529d5971de916c
@call git apply ..\ui\sd_internal\ddim_callback.patch @call git apply ..\ui\sd_internal\ddim_callback.patch
@call git apply ..\ui\sd_internal\env_yaml.patch @call git apply ..\ui\sd_internal\env_yaml.patch

4
scripts/on_sd_start.sh
View File

@ -16,7 +16,7 @@ if [ -e "scripts/install_status.txt" ] && [ `grep -c sd_git_cloned scripts/insta
git reset --hard git reset --hard
git pull git pull
git checkout d154155d4c0b43e13ec1f00eb72b7ff9d522fcf9 git checkout f6cfebffa752ee11a7b07497b8529d5971de916c
git apply ../ui/sd_internal/ddim_callback.patch git apply ../ui/sd_internal/ddim_callback.patch
git apply ../ui/sd_internal/env_yaml.patch git apply ../ui/sd_internal/env_yaml.patch
@ -34,7 +34,7 @@ else
fi fi
cd stable-diffusion cd stable-diffusion
git checkout d154155d4c0b43e13ec1f00eb72b7ff9d522fcf9 git checkout f6cfebffa752ee11a7b07497b8529d5971de916c
git apply ../ui/sd_internal/ddim_callback.patch git apply ../ui/sd_internal/ddim_callback.patch
git apply ../ui/sd_internal/env_yaml.patch git apply ../ui/sd_internal/env_yaml.patch

522
ui/index.html
View File

@ -1,6 +1,8 @@
<!DOCTYPE html> <!DOCTYPE html>
<html> <html>
<meta name="viewport" content="width=device-width, initial-scale=1.0"> <meta name="viewport" content="width=device-width, initial-scale=1.0">
<link rel="icon" type="image/png" href="/media/favicon-16x16.png" sizes="16x16">
<link rel="icon" type="image/png" href="/media/favicon-32x32.png" sizes="32x32">
<style> <style>
body { body {
font-family: Arial, Helvetica, sans-serif; font-family: Arial, Helvetica, sans-serif;
@ -19,7 +21,8 @@
} }
#prompt { #prompt {
width: 100%; width: 100%;
height: 50pt; height: 65pt;
box-sizing: border-box;
} }
@media screen and (max-width: 600px) { @media screen and (max-width: 600px) {
#prompt { #prompt {
@ -27,7 +30,7 @@
} }
} }
.image_preview_container { .image_preview_container {
display: none; /* display: none; */
margin-top: 10pt; margin-top: 10pt;
} }
.image_clear_btn { .image_clear_btn {
@ -45,14 +48,14 @@
font-family: Verdana; font-family: Verdana;
font-size: 8pt; font-size: 8pt;
} }
#editor-settings-entries { .settings-box ul {
font-size: 9pt; font-size: 9pt;
margin-bottom: 5px; margin-bottom: 5px;
padding-left: 10px; padding-left: 10px;
list-style-type: none; list-style-type: none;
} }
#editor-settings-entries li { .settings-box li {
padding-bottom: 3pt; padding-bottom: 4pt;
} }
.editor-slider { .editor-slider {
transform: translateY(30%); transform: translateY(30%);
@ -60,6 +63,9 @@
#outputMsg { #outputMsg {
font-size: small; font-size: small;
} }
#progressBar {
font-size: small;
}
#footer { #footer {
font-size: small; font-size: small;
padding-left: 10pt; padding-left: 10pt;
@ -102,23 +108,26 @@
} }
#container { #container {
width: 75%; width: 90%;
margin-left: auto; margin-left: auto;
margin-right: auto; margin-right: auto;
} }
@media screen and (max-width: 1400px) { @media screen and (max-width: 1800px) {
#container { #container {
width: 100%; width: 100%;
} }
} }
#meta small { #logo small {
font-size: 11pt; font-size: 11pt;
} }
#editor { #editor {
padding: 5px; padding: 5px;
} }
#editor label { #editor label {
font-weight: bold; font-weight: normal;
}
.settings-box label small {
color: rgb(153, 153, 153);
} }
#preview { #preview {
padding: 5px; padding: 5px;
@ -169,6 +178,9 @@
.col-50 { .col-50 {
flex: 50%; flex: 50%;
} }
.col-fixed-10 {
flex: 0 0 400pt;
}
.col-free { .col-free {
flex: 1; flex: 1;
} }
@ -247,16 +259,19 @@
display: none; display: none;
} }
#server-status { #server-status {
display: inline;
float: right; float: right;
transform: translateY(-5pt);
} }
#server-status-color { #server-status-color {
width: 8pt; /* width: 8pt;
height: 8pt; height: 8pt;
border-radius: 4pt; border-radius: 4pt; */
background-color: rgb(128, 87, 0); font-size: 14pt;
color: rgb(128, 87, 0);
/* background-color: rgb(197, 1, 1); */ /* background-color: rgb(197, 1, 1); */
float: left; /* transform: translateY(15%); */
transform: translateY(15%); display: inline;
} }
#server-status-msg { #server-status-msg {
color: rgb(128, 87, 0); color: rgb(128, 87, 0);
@ -271,18 +286,155 @@
height: 23px; height: 23px;
transform: translateY(25%); transform: translateY(25%);
} }
#inpaintingEditor {
width: 300pt;
height: 300pt;
margin-top: 5pt;
}
.drawing-board-canvas-wrapper {
background-size: 100% 100%;
}
#inpaintingEditor canvas {
opacity: 0.6;
}
#enable_mask {
margin-top: 8pt;
}
#top-nav {
padding-top: 3pt;
padding-bottom: 15pt;
}
#top-nav .icon {
padding-right: 4pt;
font-size: 14pt;
transform: translateY(1pt);
}
#logo {
display: inline;
}
#logo h1 {
display: inline;
}
#top-nav-items {
list-style-type: none;
display: inline;
float: right;
}
#top-nav-items > li {
float: left;
display: inline;
padding-left: 20pt;
cursor: default;
}
#initial-text {
padding-top: 15pt;
padding-left: 4pt;
}
.settings-subheader {
font-size: 10pt;
font-weight: bold;
}
.pl-5 {
padding-left: 5pt;
}
#system-settings {
width: 360pt;
transform: translateX(-100%) translateX(70pt);
padding-top: 10pt;
padding-bottom: 10pt;
}
#system-settings ul {
margin: 0;
padding: 0;
}
#system-settings li {
padding-left: 5pt;
}
#community-links {
list-style-type: none;
margin: 0;
padding: 12pt;
padding-bottom: 0pt;
transform: translateX(-15%);
}
#community-links li {
padding-bottom: 12pt;
display: block;
font-size: 10pt;
}
#community-links li .fa-fw {
padding-right: 2pt;
}
#community-links li a {
color: white;
text-decoration: none;
}
.dropdown {
overflow: hidden;
}
.dropdown-content {
display: none;
position: absolute;
z-index: 2;
background: rgb(18, 18, 19);
border: 2px solid rgb(37, 38, 41);
border-radius: 7px;
padding: 5px;
margin-bottom: 15px;
box-shadow: 0 20px 28px 0 rgba(0, 0, 0, 0.15), 0 6px 20px 0 rgba(0, 0, 0, 0.15);
}
.dropdown:hover .dropdown-content {
display: block;
}
</style> </style>
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.2.0/css/all.min.css">
<link rel="stylesheet" href="/media/drawingboard.min.css">
<script src="/media/jquery-3.6.1.min.js"></script>
<script src="/media/drawingboard.min.js"></script>
</html> </html>
<body> <body>
<div id="container"> <div id="container">
<div class="flex-container"> <div id="top-nav">
<div id="editor" class="col-50"> <div id="logo">
<div id="meta"> <h1>Stable Diffusion UI <small>v2.16 <span id="updateBranchLabel"></span></small></h1>
<div id="server-status"> </div>
<div id="server-status-color">&nbsp;</div> <ul id="top-nav-items">
<span id="server-status-msg">Stable Diffusion is starting..</span> <li class="dropdown">
<span><i class="fa fa-comments icon"></i> Help & Community</span>
<ul id="community-links" class="dropdown-content">
<li><a href="https://github.com/cmdr2/stable-diffusion-ui/blob/main/Troubleshooting.md" target="_blank"><i class="fa-solid fa-circle-question fa-fw"></i> Usual problems and solutions</a></li>
<li><a href="https://discord.com/invite/u9yhsFmEkB" target="_blank"><i class="fa-brands fa-discord fa-fw"></i> Discord user community</a></li>
<li><a href="https://github.com/cmdr2/stable-diffusion-ui" target="_blank"><i class="fa-brands fa-github fa-fw"></i> Source code on GitHub</a></li>
</ul>
</li>
<li class="dropdown">
<span><i class="fa fa-gear icon"></i> Settings</span>
<div id="system-settings" class="panel-box settings-box dropdown-content">
<ul id="system-settings-entries">
<li><b class="settings-subheader">System Settings</b></li>
<br/>
<li><input id="save_to_disk" name="save_to_disk" type="checkbox"> <label for="save_to_disk">Automatically save to <input id="diskPath" name="diskPath" size="40" disabled></label></li>
<li><input id="sound_toggle" name="sound_toggle" type="checkbox" checked> <label for="sound_toggle">Play sound on task completion</label></li>
<li><input id="turbo" name="turbo" type="checkbox" checked> <label for="turbo">Turbo mode <small>(generates images faster, but uses an additional 1 GB of GPU memory)</small></label></li>
<li><input id="use_cpu" name="use_cpu" type="checkbox"> <label for="use_cpu">Use CPU instead of GPU <small>(warning: this will be *very* slow)</small></label></li>
<li><input id="use_full_precision" name="use_full_precision" type="checkbox"> <label for="use_full_precision">Use full precision <small>(for GPU-only. warning: this will consume more VRAM)</small></label></li>
<!-- <li><input id="allow_nsfw" name="allow_nsfw" type="checkbox"> <label for="allow_nsfw">Allow NSFW Content (You confirm you are above 18 years of age)</label></li> -->
<br/>
<li><input id="use_beta_channel" name="use_beta_channel" type="checkbox"> <label for="use_beta_channel">🔥Beta channel. Get the latest features immediately (but could be less stable). Please restart the program after changing this.</label></li>
</ul>
</div> </div>
<h1>Stable Diffusion UI <small>v2.1 <span id="updateBranchLabel"></span></small></h1> </li>
</ul>
</div>
<div class="flex-container">
<div id="editor" class="col-fixed-10">
<div id="server-status">
<div id="server-status-color">●</div>
<span id="server-status-msg">Stable Diffusion is starting..</span>
</div> </div>
<div id="editor-inputs"> <div id="editor-inputs">
<div id="editor-inputs-prompt" class="row"> <div id="editor-inputs-prompt" class="row">
@ -291,10 +443,15 @@
</div> </div>
<div id="editor-inputs-init-image" class="row"> <div id="editor-inputs-init-image" class="row">
<label for="init_image"><b>Initial Image:</b> (optional) </label> <input id="init_image" name="init_image" type="file" /> </button><br/> <label for="init_image"><b>Initial Image:</b> (optional) </label> <input id="init_image" name="init_image" type="file" /><br/>
<div id="init_image_preview_container" class="image_preview_container"> <div id="init_image_preview_container" class="image_preview_container">
<img id="init_image_preview" src="" width="100" height="100" /> <img id="init_image_preview" src="" width="100" height="100" />
<button id="init_image_clear" class="image_clear_btn">X</button> <button class="init_image_clear image_clear_btn">X</button>
<br/>
<input id="enable_mask" name="enable_mask" type="checkbox"> <label for="enable_mask">In-Painting (select the area which the AI will paint into)</label>
<div id="inpaintingEditor"></div>
</div> </div>
</div> </div>
@ -310,22 +467,25 @@
<div class="line-separator">&nbsp;</div> <div class="line-separator">&nbsp;</div>
<div id="editor-settings" class="panel-box"> <div id="editor-settings" class="panel-box settings-box">
<h4 class="collapsible">Advanced Settings</h4> <h4 class="collapsible">Image Settings</h4>
<ul id="editor-settings-entries" class="collapsible-content"> <ul id="editor-settings-entries" class="collapsible-content">
<li><input id="use_face_correction" name="use_face_correction" type="checkbox" checked> <label for="use_face_correction">Fix incorrect faces and eyes (uses GFPGAN)</label></li> <li><b class="settings-subheader">Image Settings</b></li>
<li> <li class="pl-5"><label for="seed">Seed:</label> <input id="seed" name="seed" size="10" value="30000"> <input id="random_seed" name="random_seed" type="checkbox" checked> <label for="random_seed">Random Image</label></li>
<input id="use_upscale" name="use_upscale" type="checkbox"> <label for="use_upscale">Upscale the image to 4x resolution using </label> <li class="pl-5"><label for="num_outputs_total">Number of images to make:</label> <input id="num_outputs_total" name="num_outputs_total" value="1" size="1"> <label for="num_outputs_parallel">Generate in parallel:</label> <input id="num_outputs_parallel" name="num_outputs_parallel" value="1" size="1"> (images at once)</li>
<select id="upscale_model" name="upscale_model"> <li id="samplerSelection" class="pl-5"><label for="sampler">Sampler:</label>
<option value="RealESRGAN_x4plus" selected>RealESRGAN_x4plus</option> <select id="sampler" name="sampler">
<option value="RealESRGAN_x4plus_anime_6B">RealESRGAN_x4plus_anime_6B</option> <option value="plms" selected>plms</option>
<option value="ddim">ddim</option>
<option value="heun">heun</option>
<option value="euler">euler</option>
<option value="euler_a">euler_a</option>
<option value="dpm2">dpm2</option>
<option value="dpm2_a">dpm2_a</option>
<option value="lms">lms</option>
</select> </select>
</li> </li>
<li><input id="show_only_filtered_image" name="show_only_filtered_image" type="checkbox" checked> <label for="show_only_filtered_image">Show only the corrected/upscaled image</label></li> <li class="pl-5"><label>Image Size: </label>
<br/>
<li><label for="seed">Seed:</label> <input id="seed" name="seed" size="10" value="30000"> <input id="random_seed" name="random_seed" type="checkbox" checked> <label for="random_seed">Random Image</label></li>
<li><label for="num_outputs_total">Number of images to make:</label> <input id="num_outputs_total" name="num_outputs_total" value="1" size="4"> <label for="num_outputs_parallel">Generate in parallel:</label> <input id="num_outputs_parallel" name="num_outputs_parallel" value="1" size="4"> (images at once)</li>
<li><label for="width">Width:</label>
<select id="width" name="width" value="512"> <select id="width" name="width" value="512">
<option value="128">128 (*)</option> <option value="128">128 (*)</option>
<option value="192">192</option> <option value="192">192</option>
@ -346,9 +506,7 @@
<option value="1536">1536</option> <option value="1536">1536</option>
<option value="1792">1792</option> <option value="1792">1792</option>
<option value="2048">2048</option> <option value="2048">2048</option>
</select> </select> <label for="width"><small>(width)</small></label>
</li>
<li><label for="height">Height:</label>
<select id="height" name="height" value="512"> <select id="height" name="height" value="512">
<option value="128">128 (*)</option> <option value="128">128 (*)</option>
<option value="192">192</option> <option value="192">192</option>
@ -370,19 +528,27 @@
<option value="1792">1792</option> <option value="1792">1792</option>
<option value="2048">2048</option> <option value="2048">2048</option>
</select> </select>
<label for="height"><small>(height)</small></label>
</li> </li>
<li><label for="num_inference_steps">Number of inference steps:</label> <input id="num_inference_steps" name="num_inference_steps" size="4" value="50"></li> <li class="pl-5"><label for="num_inference_steps">Number of inference steps:</label> <input id="num_inference_steps" name="num_inference_steps" size="4" value="50"></li>
<li><label for="guidance_scale_slider">Guidance Scale:</label> <input id="guidance_scale_slider" name="guidance_scale_slider" class="editor-slider" value="75" type="range" min="10" max="200"> <input id="guidance_scale" name="guidance_scale" size="4"></li> <li class="pl-5"><label for="guidance_scale_slider">Guidance Scale:</label> <input id="guidance_scale_slider" name="guidance_scale_slider" class="editor-slider" value="75" type="range" min="10" max="500"> <input id="guidance_scale" name="guidance_scale" size="4"></li>
<li><span id="prompt_strength_container"><label for="prompt_strength_slider">Prompt Strength:</label> <input id="prompt_strength_slider" name="prompt_strength_slider" class="editor-slider" value="80" type="range" min="0" max="99"> <input id="prompt_strength" name="prompt_strength" size="4"><br/></span></li> <li class="pl-5"><span id="prompt_strength_container"><label for="prompt_strength_slider">Prompt Strength:</label> <input id="prompt_strength_slider" name="prompt_strength_slider" class="editor-slider" value="80" type="range" min="0" max="99"> <input id="prompt_strength" name="prompt_strength" size="4"><br/></span></li>
<li>&nbsp;</li>
<li><input id="save_to_disk" name="save_to_disk" type="checkbox"> <label for="save_to_disk">Automatically save to <input id="diskPath" name="diskPath" size="40" disabled></label></li>
<li><input id="sound_toggle" name="sound_toggle" type="checkbox" checked> <label for="sound_toggle">Play sound on task completion</label></li>
<li><input id="turbo" name="turbo" type="checkbox" checked> <label for="turbo">Turbo mode (generates images faster, but uses an additional 1 GB of GPU memory)</label></li>
<li><input id="use_cpu" name="use_cpu" type="checkbox"> <label for="use_cpu">Use CPU instead of GPU (warning: this will be *very* slow)</label></li>
<li><input id="use_full_precision" name="use_full_precision" type="checkbox"> <label for="use_full_precision">Use full precision (for GPU-only. warning: this will consume more VRAM)</label></li>
<!-- <li><input id="allow_nsfw" name="allow_nsfw" type="checkbox"> <label for="allow_nsfw">Allow NSFW Content (You confirm you are above 18 years of age)</label></li> -->
<br/> <br/>
<li><input id="use_beta_channel" name="use_beta_channel" type="checkbox"> <label for="use_beta_channel">🔥Beta channel. Get the latest features immediately (but could be less stable). Please restart the program after changing this.</label></li>
<li><b class="settings-subheader">Render Settings</b></li>
<li class="pl-5"><input id="stream_image_progress" name="stream_image_progress" type="checkbox"> <label for="stream_image_progress">Show a live preview of the image <small>(consumes more VRAM, slightly slower image generation)</small></label></li>
<li class="pl-5"><input id="use_face_correction" name="use_face_correction" type="checkbox" checked> <label for="use_face_correction">Fix incorrect faces and eyes <small>(uses GFPGAN)</small></label></li>
<li class="pl-5">
<input id="use_upscale" name="use_upscale" type="checkbox"> <label for="use_upscale">Upscale the image to 4x resolution using </label>
<select id="upscale_model" name="upscale_model">
<option value="RealESRGAN_x4plus" selected>RealESRGAN_x4plus</option>
<option value="RealESRGAN_x4plus_anime_6B">RealESRGAN_x4plus_anime_6B</option>
</select>
</li>
<li class="pl-5"><input id="show_only_filtered_image" name="show_only_filtered_image" type="checkbox" checked> <label for="show_only_filtered_image">Show only the corrected/upscaled image</label></li>
<br/>
<li><small>The system-related settings have been moved to the top-right corner.</small></li>
</ul> </ul>
</div> </div>
@ -393,10 +559,15 @@
</div> </div>
</div> </div>
<div id="preview" class="col-50"> <div id="preview" class="col-free">
<div id="preview-prompt">Type a prompt and press the "Make Image" button.<br/><br/>You can set an "Initial Image" if you want to guide the AI.<br/><br/>You can also add modifiers like "Realistic", "Pencil Sketch", "ArtStation" etc by browsing through the "Image Modifiers" section and selecting the desired modifiers.<br/><br/>Click "Advanced Settings" for additional settings like seed, image size, number of images to generate etc.<br/><br/>Enjoy! :)</div> <div id="preview-prompt">
<div id="initial-text">
Type a prompt and press the "Make Image" button.<br/><br/>You can set an "Initial Image" if you want to guide the AI.<br/><br/>You can also add modifiers like "Realistic", "Pencil Sketch", "ArtStation" etc by browsing through the "Image Modifiers" section and selecting the desired modifiers.<br/><br/>Click "Advanced Settings" for additional settings like seed, image size, number of images to generate etc.<br/><br/>Enjoy! :)
</div>
</div>
<div id="outputMsg"></div> <div id="outputMsg"></div>
<div id="progressBar"></div>
<div id="current-images" class="img-preview"> <div id="current-images" class="img-preview">
</div> </div>
</div> </div>
@ -428,11 +599,14 @@ const MODIFIERS_PANEL_OPEN_KEY = "modifiersPanelOpen"
const USE_FACE_CORRECTION_KEY = "useFaceCorrection" const USE_FACE_CORRECTION_KEY = "useFaceCorrection"
const USE_UPSCALING_KEY = "useUpscaling" const USE_UPSCALING_KEY = "useUpscaling"
const SHOW_ONLY_FILTERED_IMAGE_KEY = "showOnlyFilteredImage" const SHOW_ONLY_FILTERED_IMAGE_KEY = "showOnlyFilteredImage"
const STREAM_IMAGE_PROGRESS_KEY = "streamImageProgress"
const HEALTH_PING_INTERVAL = 5 // seconds const HEALTH_PING_INTERVAL = 5 // seconds
const MAX_INIT_IMAGE_DIMENSION = 768 const MAX_INIT_IMAGE_DIMENSION = 768
const IMAGE_REGEX = new RegExp('data:image/[A-Za-z]+;base64') const IMAGE_REGEX = new RegExp('data:image/[A-Za-z]+;base64')
let sessionId = new Date().getTime()
let promptField = document.querySelector('#prompt') let promptField = document.querySelector('#prompt')
let numOutputsTotalField = document.querySelector('#num_outputs_total') let numOutputsTotalField = document.querySelector('#num_outputs_total')
let numOutputsParallelField = document.querySelector('#num_outputs_parallel') let numOutputsParallelField = document.querySelector('#num_outputs_parallel')
@ -445,8 +619,8 @@ let widthField = document.querySelector('#width')
let heightField = document.querySelector('#height') let heightField = document.querySelector('#height')
let initImageSelector = document.querySelector("#init_image") let initImageSelector = document.querySelector("#init_image")
let initImagePreview = document.querySelector("#init_image_preview") let initImagePreview = document.querySelector("#init_image_preview")
// let maskImageSelector = document.querySelector("#mask") let maskImageSelector = document.querySelector("#mask")
// let maskImagePreview = document.querySelector("#mask_preview") let maskImagePreview = document.querySelector("#mask_preview")
let turboField = document.querySelector('#turbo') let turboField = document.querySelector('#turbo')
let useCPUField = document.querySelector('#use_cpu') let useCPUField = document.querySelector('#use_cpu')
let useFullPrecisionField = document.querySelector('#use_full_precision') let useFullPrecisionField = document.querySelector('#use_full_precision')
@ -456,23 +630,27 @@ let diskPathField = document.querySelector('#diskPath')
let useBetaChannelField = document.querySelector("#use_beta_channel") let useBetaChannelField = document.querySelector("#use_beta_channel")
let promptStrengthSlider = document.querySelector('#prompt_strength_slider') let promptStrengthSlider = document.querySelector('#prompt_strength_slider')
let promptStrengthField = document.querySelector('#prompt_strength') let promptStrengthField = document.querySelector('#prompt_strength')
let samplerField = document.querySelector('#sampler')
let samplerSelectionContainer = document.querySelector("#samplerSelection")
let useFaceCorrectionField = document.querySelector("#use_face_correction") let useFaceCorrectionField = document.querySelector("#use_face_correction")
let useUpscalingField = document.querySelector("#use_upscale") let useUpscalingField = document.querySelector("#use_upscale")
let upscaleModelField = document.querySelector("#upscale_model") let upscaleModelField = document.querySelector("#upscale_model")
let showOnlyFilteredImageField = document.querySelector("#show_only_filtered_image") let showOnlyFilteredImageField = document.querySelector("#show_only_filtered_image")
let updateBranchLabel = document.querySelector("#updateBranchLabel") let updateBranchLabel = document.querySelector("#updateBranchLabel")
let streamImageProgressField = document.querySelector("#stream_image_progress")
let makeImageBtn = document.querySelector('#makeImage') let makeImageBtn = document.querySelector('#makeImage')
let stopImageBtn = document.querySelector('#stopImage') let stopImageBtn = document.querySelector('#stopImage')
let imagesContainer = document.querySelector('#current-images') let imagesContainer = document.querySelector('#current-images')
let initImagePreviewContainer = document.querySelector('#init_image_preview_container') let initImagePreviewContainer = document.querySelector('#init_image_preview_container')
let initImageClearBtn = document.querySelector('#init_image_clear') let initImageClearBtn = document.querySelector('.init_image_clear')
let promptStrengthContainer = document.querySelector('#prompt_strength_container') let promptStrengthContainer = document.querySelector('#prompt_strength_container')
// let maskSetting = document.querySelector('#mask_setting') // let maskSetting = document.querySelector('#editor-inputs-mask_setting')
// let maskImagePreviewContainer = document.querySelector('#mask_preview_container') // let maskImagePreviewContainer = document.querySelector('#mask_preview_container')
// let maskImageClearBtn = document.querySelector('#mask_clear') // let maskImageClearBtn = document.querySelector('#mask_clear')
let maskSetting = document.querySelector('#enable_mask')
let editorModifierEntries = document.querySelector('#editor-modifiers-entries') let editorModifierEntries = document.querySelector('#editor-modifiers-entries')
let editorModifierTagsList = document.querySelector('#editor-inputs-tags-list') let editorModifierTagsList = document.querySelector('#editor-inputs-tags-list')
@ -483,6 +661,7 @@ let previewPrompt = document.querySelector('#preview-prompt')
let showConfigToggle = document.querySelector('#configToggleBtn') let showConfigToggle = document.querySelector('#configToggleBtn')
// let configBox = document.querySelector('#config') // let configBox = document.querySelector('#config')
let outputMsg = document.querySelector('#outputMsg') let outputMsg = document.querySelector('#outputMsg')
let progressBar = document.querySelector("#progressBar")
let soundToggle = document.querySelector('#sound_toggle') let soundToggle = document.querySelector('#sound_toggle')
@ -491,11 +670,35 @@ let serverStatusMsg = document.querySelector('#server-status-msg')
let advancedPanelHandle = document.querySelector("#editor-settings .collapsible") let advancedPanelHandle = document.querySelector("#editor-settings .collapsible")
let modifiersPanelHandle = document.querySelector("#editor-modifiers .collapsible") let modifiersPanelHandle = document.querySelector("#editor-modifiers .collapsible")
let inpaintingEditorContainer = document.querySelector('#inpaintingEditor')
let inpaintingEditor = new DrawingBoard.Board('inpaintingEditor', {
color: "#ffffff",
background: false,
size: 30,
webStorage: false,
controls: [{'DrawingMode': {'filler': false}}, 'Size', 'Navigation']
})
let inpaintingEditorCanvasBackground = document.querySelector('.drawing-board-canvas-wrapper')
// let inpaintingEditorControls = document.querySelector('.drawing-board-controls')
// let inpaintingEditorMetaControl = document.createElement('div')
// inpaintingEditorMetaControl.className = 'drawing-board-control'
// let initImageClearBtnToolbar = document.createElement('button')
// initImageClearBtnToolbar.className = 'init_image_clear'
// initImageClearBtnToolbar.innerHTML = 'Remove Image'
// inpaintingEditorMetaControl.appendChild(initImageClearBtnToolbar)
// inpaintingEditorControls.appendChild(inpaintingEditorMetaControl)
let maskResetButton = document.querySelector('.drawing-board-control-navigation-reset')
maskResetButton.innerHTML = 'Clear'
maskResetButton.style.fontWeight = 'normal'
maskResetButton.style.fontSize = '10pt'
let serverStatus = 'offline' let serverStatus = 'offline'
let activeTags = [] let activeTags = []
let lastPromptUsed = '' let lastPromptUsed = ''
let taskStopped = true let taskStopped = true
let batchesDone = 0
function getLocalStorageItem(key, fallback) { function getLocalStorageItem(key, fallback) {
let item = localStorage.getItem(key) let item = localStorage.getItem(key)
@ -571,6 +774,10 @@ function isModifiersPanelOpenEnabled() {
return getLocalStorageBoolItem(MODIFIERS_PANEL_OPEN_KEY, false) return getLocalStorageBoolItem(MODIFIERS_PANEL_OPEN_KEY, false)
} }
function isStreamImageProgressEnabled() {
return getLocalStorageBoolItem(STREAM_IMAGE_PROGRESS_KEY, false)
}
function setStatus(statusType, msg, msgType) { function setStatus(statusType, msg, msgType) {
if (statusType !== 'server') { if (statusType !== 'server') {
return; return;
@ -578,12 +785,12 @@ function setStatus(statusType, msg, msgType) {
if (msgType == 'error') { if (msgType == 'error') {
// msg = '<span style="color: red">' + msg + '<span>' // msg = '<span style="color: red">' + msg + '<span>'
serverStatusColor.style.backgroundColor = 'red' serverStatusColor.style.color = 'red'
serverStatusMsg.style.color = 'red' serverStatusMsg.style.color = 'red'
serverStatusMsg.innerHTML = 'Stable Diffusion has stopped' serverStatusMsg.innerHTML = 'Stable Diffusion has stopped'
} else if (msgType == 'success') { } else if (msgType == 'success') {
// msg = '<span style="color: green">' + msg + '<span>' // msg = '<span style="color: green">' + msg + '<span>'
serverStatusColor.style.backgroundColor = 'green' serverStatusColor.style.color = 'green'
serverStatusMsg.style.color = 'green' serverStatusMsg.style.color = 'green'
serverStatusMsg.innerHTML = 'Stable Diffusion is ready' serverStatusMsg.innerHTML = 'Stable Diffusion is ready'
serverStatus = 'online' serverStatus = 'online'
@ -630,14 +837,37 @@ async function healthCheck() {
} }
} }
function makeImageElement(width, height) {
let imgItem = document.createElement('div')
imgItem.className = 'imgItem'
let img = document.createElement('img')
img.width = parseInt(width)
img.height = parseInt(height)
imgItem.appendChild(img)
imagesContainer.appendChild(imgItem)
return imgItem
}
// makes a single image. don't call this directly, use makeImage() instead // makes a single image. don't call this directly, use makeImage() instead
async function doMakeImage(reqBody) { async function doMakeImage(reqBody, batchCount) {
if (taskStopped) { if (taskStopped) {
return return
} }
let res = '' let res = ''
let seed = reqBody['seed'] let seed = reqBody['seed']
let numOutputs = parseInt(reqBody['num_outputs'])
let images = []
function makeImageContainers(numImages) {
for (let i = images.length; i < numImages; i++) {
images.push(makeImageElement(reqBody.width, reqBody.height))
}
}
try { try {
res = await fetch('/image', { res = await fetch('/image', {
@ -648,15 +878,82 @@ async function doMakeImage(reqBody) {
body: JSON.stringify(reqBody) body: JSON.stringify(reqBody)
}) })
let reader = res.body.getReader()
let textDecoder = new TextDecoder()
let finalJSON = ''
let prevTime = -1
while (true) {
try {
let t = new Date().getTime()
const {value, done} = await reader.read()
if (done) {
break
}
let timeTaken = (prevTime === -1 ? -1 : t - prevTime)
let jsonStr = textDecoder.decode(value)
try {
let stepUpdate = JSON.parse(jsonStr)
if (stepUpdate.step === undefined) {
finalJSON += jsonStr
} else {
let batchSize = stepUpdate.total_steps
let overallStepCount = stepUpdate.step + batchesDone * batchSize
let totalSteps = batchCount * batchSize
let percent = 100 * (overallStepCount / totalSteps)
percent = (percent > 100 ? 100 : percent)
percent = percent.toFixed(0)
stepsRemaining = totalSteps - overallStepCount
stepsRemaining = (stepsRemaining < 0 ? 0 : stepsRemaining)
timeRemaining = (timeTaken === -1 ? '' : stepsRemaining * timeTaken) // ms
outputMsg.innerHTML = `Batch ${batchesDone+1} of ${batchCount}`
progressBar.innerHTML = `Generating image(s): ${percent}%`
if (timeTaken !== -1) {
progressBar.innerHTML += `<br>Time remaining (approx): ${millisecondsToStr(timeRemaining)}`
}
progressBar.style.display = 'block'
if (stepUpdate.output !== undefined) {
makeImageContainers(numOutputs)
for (idx in stepUpdate.output) {
let imgItem = images[idx]
let img = imgItem.firstChild
let tmpImageData = stepUpdate.output[idx]
img.src = tmpImageData['path'] + '?t=' + new Date().getTime()
}
}
}
} catch (e) {
finalJSON += jsonStr
}
prevTime = t
} catch (e) {
logError('Stable Diffusion had an error. Please check the logs in the command-line window.', res)
res = undefined
throw e
}
}
if (res.status != 200) { if (res.status != 200) {
if (serverStatus === 'online') { if (serverStatus === 'online') {
logError('Stable Diffusion had an error: ' + await res.text() + '. This happens sometimes. Maybe modify the prompt or seed a little bit?', res) logError('Stable Diffusion had an error: ' + await res.text(), res)
} else { } else {
logError("Stable Diffusion is still starting up, please wait. If this goes on beyond a few minutes, Stable Diffusion has probably crashed.", res) logError("Stable Diffusion is still starting up, please wait. If this goes on beyond a few minutes, Stable Diffusion has probably crashed. Please check the error message in the command-line window.", res)
} }
res = undefined res = undefined
progressBar.style.display = 'none'
} else { } else {
res = await res.json() res = JSON.parse(finalJSON)
progressBar.style.display = 'none'
if (res.status !== 'succeeded') { if (res.status !== 'succeeded') {
let msg = '' let msg = ''
@ -680,7 +977,10 @@ async function doMakeImage(reqBody) {
} }
} catch (e) { } catch (e) {
console.log('request error', e) console.log('request error', e)
logError('Stable Diffusion had an error. Please check the logs in the command-line window. <br/><br/>' + e + '<br/><pre>' + e.stack + '</pre>', res)
setStatus('request', 'error', 'error') setStatus('request', 'error', 'error')
progressBar.style.display = 'none'
res = undefined
} }
if (!res) { if (!res) {
@ -689,6 +989,8 @@ async function doMakeImage(reqBody) {
lastPromptUsed = reqBody['prompt'] lastPromptUsed = reqBody['prompt']
makeImageContainers(res.output.length)
for (let idx in res.output) { for (let idx in res.output) {
let imgBody = '' let imgBody = ''
let seed = 0 let seed = 0
@ -703,12 +1005,9 @@ async function doMakeImage(reqBody) {
continue continue
} }
let imgItem = document.createElement('div') let imgItem = images[idx]
imgItem.className = 'imgItem' let img = imgItem.firstChild
let img = document.createElement('img')
img.width = parseInt(reqBody.width)
img.height = parseInt(reqBody.height)
img.src = imgBody img.src = imgBody
let imgItemInfo = document.createElement('span') let imgItemInfo = document.createElement('span')
@ -726,19 +1025,19 @@ async function doMakeImage(reqBody) {
imgSaveBtn.className = 'imgSaveBtn' imgSaveBtn.className = 'imgSaveBtn'
imgSaveBtn.innerHTML = 'Download' imgSaveBtn.innerHTML = 'Download'
imgItem.appendChild(img)
imgItem.appendChild(imgItemInfo) imgItem.appendChild(imgItemInfo)
imgItemInfo.appendChild(imgSeedLabel) imgItemInfo.appendChild(imgSeedLabel)
imgItemInfo.appendChild(imgUseBtn) imgItemInfo.appendChild(imgUseBtn)
imgItemInfo.appendChild(imgSaveBtn) imgItemInfo.appendChild(imgSaveBtn)
imagesContainer.appendChild(imgItem)
imgUseBtn.addEventListener('click', function() { imgUseBtn.addEventListener('click', function() {
initImageSelector.value = null initImageSelector.value = null
initImagePreview.src = imgBody initImagePreview.src = imgBody
initImagePreviewContainer.style.display = 'block' initImagePreviewContainer.style.display = 'block'
inpaintingEditorContainer.style.display = 'none'
promptStrengthContainer.style.display = 'block' promptStrengthContainer.style.display = 'block'
maskSetting.checked = false
// maskSetting.style.display = 'block' // maskSetting.style.display = 'block'
@ -789,7 +1088,7 @@ async function makeImage() {
let validation = validateInput() let validation = validateInput()
if (validation['isValid']) { if (validation['isValid']) {
outputMsg.innerHTML = 'Fetching..' outputMsg.innerHTML = 'Starting..'
} else { } else {
if (validation['error']) { if (validation['error']) {
logError(validation['error']) logError(validation['error'])
@ -802,10 +1101,12 @@ async function makeImage() {
setStatus('request', 'fetching..') setStatus('request', 'fetching..')
makeImageBtn.innerHTML = 'Processing..' makeImageBtn.innerHTML = 'Processing..'
makeImageBtn.disabled = true
makeImageBtn.style.display = 'none' makeImageBtn.style.display = 'none'
stopImageBtn.style.display = 'block' stopImageBtn.style.display = 'block'
taskStopped = false taskStopped = false
batchesDone = 0
let seed = (randomSeedField.checked ? Math.floor(Math.random() * 10000000) : parseInt(seedField.value)) let seed = (randomSeedField.checked ? Math.floor(Math.random() * 10000000) : parseInt(seedField.value))
let numOutputsTotal = parseInt(numOutputsTotalField.value) let numOutputsTotal = parseInt(numOutputsTotalField.value)
@ -813,6 +1114,8 @@ async function makeImage() {
let batchCount = Math.ceil(numOutputsTotal / numOutputsParallel) let batchCount = Math.ceil(numOutputsTotal / numOutputsParallel)
let batchSize = numOutputsParallel let batchSize = numOutputsParallel
let streamImageProgress = (numOutputsTotal > 50 ? false : streamImageProgressField.checked)
let prompt = promptField.value let prompt = promptField.value
if (activeTags.length > 0) { if (activeTags.length > 0) {
let promptTags = activeTags.join(", ") let promptTags = activeTags.join(", ")
@ -822,6 +1125,7 @@ async function makeImage() {
previewPrompt.innerHTML = prompt previewPrompt.innerHTML = prompt
let reqBody = { let reqBody = {
session_id: sessionId,
prompt: prompt, prompt: prompt,
num_outputs: batchSize, num_outputs: batchSize,
num_inference_steps: numInferenceStepsField.value, num_inference_steps: numInferenceStepsField.value,
@ -831,7 +1135,10 @@ async function makeImage() {
// allow_nsfw: allowNSFWField.checked, // allow_nsfw: allowNSFWField.checked,
turbo: turboField.checked, turbo: turboField.checked,
use_cpu: useCPUField.checked, use_cpu: useCPUField.checked,
use_full_precision: useFullPrecisionField.checked use_full_precision: useFullPrecisionField.checked,
stream_progress_updates: true,
stream_image_progress: streamImageProgress,
show_only_filtered_image: showOnlyFilteredImageField.checked
} }
if (IMAGE_REGEX.test(initImagePreview.src)) { if (IMAGE_REGEX.test(initImagePreview.src)) {
@ -841,6 +1148,13 @@ async function makeImage() {
// if (IMAGE_REGEX.test(maskImagePreview.src)) { // if (IMAGE_REGEX.test(maskImagePreview.src)) {
// reqBody['mask'] = maskImagePreview.src // reqBody['mask'] = maskImagePreview.src
// } // }
if (maskSetting.checked) {
reqBody['mask'] = inpaintingEditor.getImg()
}
reqBody['sampler'] = 'ddim'
} else {
reqBody['sampler'] = samplerField.value
} }
if (saveToDiskField.checked && diskPathField.value.trim() !== '') { if (saveToDiskField.checked && diskPathField.value.trim() !== '') {
@ -855,10 +1169,6 @@ async function makeImage() {
reqBody['use_upscale'] = upscaleModelField.value reqBody['use_upscale'] = upscaleModelField.value
} }
if (showOnlyFilteredImageField.checked && (useUpscalingField.checked || useFaceCorrectionField.checked)) {
reqBody['show_only_filtered_image'] = showOnlyFilteredImageField.checked
}
let time = new Date().getTime() let time = new Date().getTime()
imagesContainer.innerHTML = '' imagesContainer.innerHTML = ''
@ -867,7 +1177,8 @@ async function makeImage() {
for (let i = 0; i < batchCount; i++) { for (let i = 0; i < batchCount; i++) {
reqBody['seed'] = seed + (i * batchSize) reqBody['seed'] = seed + (i * batchSize)
let success = await doMakeImage(reqBody) let success = await doMakeImage(reqBody, batchCount)
batchesDone++
if (success) { if (success) {
outputMsg.innerHTML = 'Processed batch ' + (i+1) + '/' + batchCount outputMsg.innerHTML = 'Processed batch ' + (i+1) + '/' + batchCount
@ -971,6 +1282,9 @@ useFullPrecisionField.checked = isUseFullPrecisionEnabled()
turboField.addEventListener('click', handleBoolSettingChange(USE_TURBO_MODE_KEY)) turboField.addEventListener('click', handleBoolSettingChange(USE_TURBO_MODE_KEY))
turboField.checked = isUseTurboModeEnabled() turboField.checked = isUseTurboModeEnabled()
streamImageProgressField.addEventListener('click', handleBoolSettingChange(STREAM_IMAGE_PROGRESS_KEY))
streamImageProgressField.checked = isStreamImageProgressEnabled()
diskPathField.addEventListener('change', handleStringSettingChange(DISK_PATH_KEY)) diskPathField.addEventListener('change', handleStringSettingChange(DISK_PATH_KEY))
saveToDiskField.addEventListener('click', function(e) { saveToDiskField.addEventListener('click', function(e) {
@ -1007,8 +1321,8 @@ function updateGuidanceScale() {
function updateGuidanceScaleSlider() { function updateGuidanceScaleSlider() {
if (guidanceScaleField.value < 0) { if (guidanceScaleField.value < 0) {
guidanceScaleField.value = 0 guidanceScaleField.value = 0
} else if (guidanceScaleField.value > 20) { } else if (guidanceScaleField.value > 50) {
guidanceScaleField.value = 20 guidanceScaleField.value = 50
} }
guidanceScaleSlider.value = guidanceScaleField.value * 10 guidanceScaleSlider.value = guidanceScaleField.value * 10
@ -1094,6 +1408,7 @@ checkRandomSeed()
function showInitImagePreview() { function showInitImagePreview() {
if (initImageSelector.files.length === 0) { if (initImageSelector.files.length === 0) {
initImagePreviewContainer.style.display = 'none' initImagePreviewContainer.style.display = 'none'
// inpaintingEditorContainer.style.display = 'none'
promptStrengthContainer.style.display = 'none' promptStrengthContainer.style.display = 'none'
// maskSetting.style.display = 'none' // maskSetting.style.display = 'none'
return return
@ -1106,9 +1421,10 @@ function showInitImagePreview() {
// console.log(file.name, reader.result) // console.log(file.name, reader.result)
initImagePreview.src = reader.result initImagePreview.src = reader.result
initImagePreviewContainer.style.display = 'block' initImagePreviewContainer.style.display = 'block'
inpaintingEditorContainer.style.display = 'none'
promptStrengthContainer.style.display = 'block' promptStrengthContainer.style.display = 'block'
samplerSelectionContainer.style.display = 'none'
// maskSetting.style.display = 'block' // maskSetting.checked = false
}) })
if (file) { if (file) {
@ -1118,24 +1434,37 @@ function showInitImagePreview() {
initImageSelector.addEventListener('change', showInitImagePreview) initImageSelector.addEventListener('change', showInitImagePreview)
showInitImagePreview() showInitImagePreview()
initImagePreview.addEventListener('load', function() {
inpaintingEditorCanvasBackground.style.backgroundImage = "url('" + this.src + "')"
// maskSetting.style.display = 'block'
// inpaintingEditorContainer.style.display = 'block'
})
initImageClearBtn.addEventListener('click', function() { initImageClearBtn.addEventListener('click', function() {
initImageSelector.value = null initImageSelector.value = null
// maskImageSelector.value = null // maskImageSelector.value = null
initImagePreview.src = '' initImagePreview.src = ''
// maskImagePreview.src = '' // maskImagePreview.src = ''
maskSetting.checked = false
initImagePreviewContainer.style.display = 'none' initImagePreviewContainer.style.display = 'none'
// inpaintingEditorContainer.style.display = 'none'
// maskImagePreviewContainer.style.display = 'none' // maskImagePreviewContainer.style.display = 'none'
// maskSetting.style.display = 'none' // maskSetting.style.display = 'none'
promptStrengthContainer.style.display = 'none' promptStrengthContainer.style.display = 'none'
samplerSelectionContainer.style.display = 'block'
})
maskSetting.addEventListener('click', function() {
inpaintingEditorContainer.style.display = (this.checked ? 'block' : 'none')
}) })
// function showMaskImagePreview() { // function showMaskImagePreview() {
// if (maskImageSelector.files.length === 0) { // if (maskImageSelector.files.length === 0) {
// maskImagePreviewContainer.style.display = 'none' // // maskImagePreviewContainer.style.display = 'none'
// return // return
// } // }
@ -1143,8 +1472,8 @@ initImageClearBtn.addEventListener('click', function() {
// let file = maskImageSelector.files[0] // let file = maskImageSelector.files[0]
// reader.addEventListener('load', function() { // reader.addEventListener('load', function() {
// maskImagePreview.src = reader.result // // maskImagePreview.src = reader.result
// maskImagePreviewContainer.style.display = 'block' // // maskImagePreviewContainer.style.display = 'block'
// }) // })
// if (file) { // if (file) {
@ -1157,8 +1486,32 @@ initImageClearBtn.addEventListener('click', function() {
// maskImageClearBtn.addEventListener('click', function() { // maskImageClearBtn.addEventListener('click', function() {
// maskImageSelector.value = null // maskImageSelector.value = null
// maskImagePreview.src = '' // maskImagePreview.src = ''
// maskImagePreviewContainer.style.display = 'none' // // maskImagePreviewContainer.style.display = 'none'
// }) // })
// https://stackoverflow.com/a/8212878
function millisecondsToStr(milliseconds) {
function numberEnding (number) {
return (number > 1) ? 's' : '';
}
var temp = Math.floor(milliseconds / 1000);
var hours = Math.floor((temp %= 86400) / 3600);
var s = ''
if (hours) {
s += hours + ' hour' + numberEnding(hours) + ' ';
}
var minutes = Math.floor((temp %= 3600) / 60);
if (minutes) {
s += minutes + ' minute' + numberEnding(minutes) + ' ';
}
var seconds = temp % 60;
if (!hours && minutes < 4 && seconds) {
s += seconds + ' second' + numberEnding(seconds);
}
return s;
}
</script> </script>
<script> <script>
function createCollapsibles(node) { function createCollapsibles(node) {
@ -1317,5 +1670,4 @@ async function init() {
init() init()
</script> </script>
</html> </html>

5
ui/media/drawingboard.min.css vendored Normal file
View File

File diff suppressed because one or more lines are too long

4
ui/media/drawingboard.min.js vendored Normal file
View File

File diff suppressed because one or more lines are too long

BIN
ui/media/favicon-16x16.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 466 B

BIN
ui/media/favicon-32x32.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 973 B

2
ui/media/jquery-3.6.1.min.js vendored Normal file
View File

File diff suppressed because one or more lines are too long

16
ui/sd_internal/__init__.py
View File

@ -1,6 +1,7 @@
import json import json
class Request: class Request:
session_id: str = "session"
prompt: str = "" prompt: str = ""
init_image: str = None # base64 init_image: str = None # base64
mask: str = None # base64 mask: str = None # base64
@ -11,6 +12,7 @@ class Request:
height: int = 512 height: int = 512
seed: int = 42 seed: int = 42
prompt_strength: float = 0.8 prompt_strength: float = 0.8
sampler: str = None # "ddim", "plms", "heun", "euler", "euler_a", "dpm2", "dpm2_a", "lms"
# allow_nsfw: bool = False # allow_nsfw: bool = False
precision: str = "autocast" # or "full" precision: str = "autocast" # or "full"
save_to_disk_path: str = None save_to_disk_path: str = None
@ -21,8 +23,12 @@ class Request:
use_upscale: str = None # or "RealESRGAN_x4plus" or "RealESRGAN_x4plus_anime_6B" use_upscale: str = None # or "RealESRGAN_x4plus" or "RealESRGAN_x4plus_anime_6B"
show_only_filtered_image: bool = False show_only_filtered_image: bool = False
stream_progress_updates: bool = False
stream_image_progress: bool = False
def json(self): def json(self):
return { return {
"session_id": self.session_id,
"prompt": self.prompt, "prompt": self.prompt,
"num_outputs": self.num_outputs, "num_outputs": self.num_outputs,
"num_inference_steps": self.num_inference_steps, "num_inference_steps": self.num_inference_steps,
@ -31,15 +37,18 @@ class Request:
"height": self.height, "height": self.height,
"seed": self.seed, "seed": self.seed,
"prompt_strength": self.prompt_strength, "prompt_strength": self.prompt_strength,
"sampler": self.sampler,
"use_face_correction": self.use_face_correction, "use_face_correction": self.use_face_correction,
"use_upscale": self.use_upscale, "use_upscale": self.use_upscale,
} }
def to_string(self): def to_string(self):
return f''' return f'''
session_id: {self.session_id}
prompt: {self.prompt} prompt: {self.prompt}
seed: {self.seed} seed: {self.seed}
num_inference_steps: {self.num_inference_steps} num_inference_steps: {self.num_inference_steps}
sampler: {self.sampler}
guidance_scale: {self.guidance_scale} guidance_scale: {self.guidance_scale}
w: {self.width} w: {self.width}
h: {self.height} h: {self.height}
@ -50,7 +59,10 @@ class Request:
use_full_precision: {self.use_full_precision} use_full_precision: {self.use_full_precision}
use_face_correction: {self.use_face_correction} use_face_correction: {self.use_face_correction}
use_upscale: {self.use_upscale} use_upscale: {self.use_upscale}
show_only_filtered_image: {self.show_only_filtered_image}''' show_only_filtered_image: {self.show_only_filtered_image}
stream_progress_updates: {self.stream_progress_updates}
stream_image_progress: {self.stream_image_progress}'''
class Image: class Image:
data: str # base64 data: str # base64
@ -71,13 +83,11 @@ class Image:
class Response: class Response:
request: Request request: Request
session_id: str
images: list images: list
def json(self): def json(self):
res = { res = {
"status": 'succeeded', "status": 'succeeded',
"session_id": self.session_id,
"request": self.request.json(), "request": self.request.json(),
"output": [], "output": [],
} }

320
ui/sd_internal/ddim_callback.patch
View File

@ -1,8 +1,26 @@
diff --git a/optimizedSD/ddpm.py b/optimizedSD/ddpm.py diff --git a/optimizedSD/ddpm.py b/optimizedSD/ddpm.py
index dcf7901..1f99adc 100644 index b967b55..35ef520 100644
--- a/optimizedSD/ddpm.py --- a/optimizedSD/ddpm.py
+++ b/optimizedSD/ddpm.py +++ b/optimizedSD/ddpm.py
@@ -528,7 +528,8 @@ class UNet(DDPM): @@ -22,7 +22,7 @@ from ldm.util import exists, default, instantiate_from_config
from ldm.modules.diffusionmodules.util import make_beta_schedule
from ldm.modules.diffusionmodules.util import make_ddim_sampling_parameters, make_ddim_timesteps, noise_like
from ldm.modules.diffusionmodules.util import make_beta_schedule, extract_into_tensor, noise_like
-from samplers import CompVisDenoiser, get_ancestral_step, to_d, append_dims,linear_multistep_coeff
+from .samplers import CompVisDenoiser, get_ancestral_step, to_d, append_dims,linear_multistep_coeff
def disabled_train(self):
"""Overwrite model.train with this function to make sure train/eval mode
@@ -506,6 +506,8 @@ class UNet(DDPM):
x_latent = noise if x0 is None else x0
# sampling
+ if sampler in ('ddim', 'dpm2', 'heun', 'dpm2_a', 'lms') and not hasattr(self, 'ddim_timesteps'):
+ self.make_schedule(ddim_num_steps=S, ddim_eta=eta, verbose=False)
if sampler == "plms":
self.make_schedule(ddim_num_steps=S, ddim_eta=eta, verbose=False)
@@ -528,39 +530,46 @@ class UNet(DDPM):
elif sampler == "ddim": elif sampler == "ddim":
samples = self.ddim_sampling(x_latent, conditioning, S, unconditional_guidance_scale=unconditional_guidance_scale, samples = self.ddim_sampling(x_latent, conditioning, S, unconditional_guidance_scale=unconditional_guidance_scale,
unconditional_conditioning=unconditional_conditioning, unconditional_conditioning=unconditional_conditioning,
@ -10,9 +28,69 @@ index dcf7901..1f99adc 100644
+ mask = mask,init_latent=x_T,use_original_steps=False, + mask = mask,init_latent=x_T,use_original_steps=False,
+ callback=callback, img_callback=img_callback) + callback=callback, img_callback=img_callback)
# elif sampler == "euler": elif sampler == "euler":
# cvd = CompVisDenoiser(self.alphas_cumprod) self.make_schedule(ddim_num_steps=S, ddim_eta=eta, verbose=False)
@@ -687,7 +688,8 @@ class UNet(DDPM): samples = self.euler_sampling(self.alphas_cumprod,x_latent, S, conditioning, unconditional_conditioning=unconditional_conditioning,
- unconditional_guidance_scale=unconditional_guidance_scale)
+ unconditional_guidance_scale=unconditional_guidance_scale,
+ img_callback=img_callback)
elif sampler == "euler_a":
self.make_schedule(ddim_num_steps=S, ddim_eta=eta, verbose=False)
samples = self.euler_ancestral_sampling(self.alphas_cumprod,x_latent, S, conditioning, unconditional_conditioning=unconditional_conditioning,
- unconditional_guidance_scale=unconditional_guidance_scale)
+ unconditional_guidance_scale=unconditional_guidance_scale,
+ img_callback=img_callback)
elif sampler == "dpm2":
samples = self.dpm_2_sampling(self.alphas_cumprod,x_latent, S, conditioning, unconditional_conditioning=unconditional_conditioning,
- unconditional_guidance_scale=unconditional_guidance_scale)
+ unconditional_guidance_scale=unconditional_guidance_scale,
+ img_callback=img_callback)
elif sampler == "heun":
samples = self.heun_sampling(self.alphas_cumprod,x_latent, S, conditioning, unconditional_conditioning=unconditional_conditioning,
- unconditional_guidance_scale=unconditional_guidance_scale)
+ unconditional_guidance_scale=unconditional_guidance_scale,
+ img_callback=img_callback)
elif sampler == "dpm2_a":
samples = self.dpm_2_ancestral_sampling(self.alphas_cumprod,x_latent, S, conditioning, unconditional_conditioning=unconditional_conditioning,
- unconditional_guidance_scale=unconditional_guidance_scale)
+ unconditional_guidance_scale=unconditional_guidance_scale,
+ img_callback=img_callback)
elif sampler == "lms":
samples = self.lms_sampling(self.alphas_cumprod,x_latent, S, conditioning, unconditional_conditioning=unconditional_conditioning,
- unconditional_guidance_scale=unconditional_guidance_scale)
+ unconditional_guidance_scale=unconditional_guidance_scale,
+ img_callback=img_callback)
+
+ yield from samples
if(self.turbo):
self.model1.to("cpu")
self.model2.to("cpu")
- return samples
-
@torch.no_grad()
def plms_sampling(self, cond,b, img,
ddim_use_original_steps=False,
@@ -599,10 +608,10 @@ class UNet(DDPM):
old_eps.append(e_t)
if len(old_eps) >= 4:
old_eps.pop(0)
- if callback: callback(i)
- if img_callback: img_callback(pred_x0, i)
+ if callback: yield from callback(i)
+ if img_callback: yield from img_callback(pred_x0, i)
- return img
+ yield from img_callback(img, len(iterator)-1)
@torch.no_grad()
def p_sample_plms(self, x, c, t, index, repeat_noise=False, use_original_steps=False, quantize_denoised=False,
@@ -706,7 +715,8 @@ class UNet(DDPM):
@torch.no_grad() @torch.no_grad()
def ddim_sampling(self, x_latent, cond, t_start, unconditional_guidance_scale=1.0, unconditional_conditioning=None, def ddim_sampling(self, x_latent, cond, t_start, unconditional_guidance_scale=1.0, unconditional_conditioning=None,
@ -22,13 +100,233 @@ index dcf7901..1f99adc 100644
timesteps = self.ddim_timesteps timesteps = self.ddim_timesteps
timesteps = timesteps[:t_start] timesteps = timesteps[:t_start]
@@ -710,6 +712,9 @@ class UNet(DDPM): @@ -730,10 +740,13 @@ class UNet(DDPM):
x_dec = self.p_sample_ddim(x_dec, cond, ts, index=index, use_original_steps=use_original_steps,
unconditional_guidance_scale=unconditional_guidance_scale, unconditional_guidance_scale=unconditional_guidance_scale,
unconditional_conditioning=unconditional_conditioning) unconditional_conditioning=unconditional_conditioning)
+
+ if callback: callback(i)
+ if img_callback: img_callback(x_dec, i)
+ if callback: yield from callback(i)
+ if img_callback: yield from img_callback(x_dec, i)
+
if mask is not None: if mask is not None:
return x0 * mask + (1. - mask) * x_dec - return x0 * mask + (1. - mask) * x_dec
+ x_dec = x0 * mask + (1. - mask) * x_dec
- return x_dec
+ yield from img_callback(x_dec, len(iterator)-1)
@torch.no_grad()
@@ -779,13 +792,16 @@ class UNet(DDPM):
@torch.no_grad()
- def euler_sampling(self, ac, x, S, cond, unconditional_conditioning = None, unconditional_guidance_scale = 1,extra_args=None,callback=None, disable=None, s_churn=0., s_tmin=0., s_tmax=float('inf'), s_noise=1.):
+ def euler_sampling(self, ac, x, S, cond, unconditional_conditioning = None, unconditional_guidance_scale = 1,extra_args=None,callback=None, disable=None, s_churn=0., s_tmin=0., s_tmax=float('inf'), s_noise=1.,
+ img_callback=None):
"""Implements Algorithm 2 (Euler steps) from Karras et al. (2022)."""
extra_args = {} if extra_args is None else extra_args
cvd = CompVisDenoiser(ac)
sigmas = cvd.get_sigmas(S)
x = x*sigmas[0]
+ print(f"Running Euler Sampling with {len(sigmas) - 1} timesteps")
+
s_in = x.new_ones([x.shape[0]]).half()
for i in trange(len(sigmas) - 1, disable=disable):
gamma = min(s_churn / (len(sigmas) - 1), 2 ** 0.5 - 1) if s_tmin <= sigmas[i] <= s_tmax else 0.
@@ -807,13 +823,18 @@ class UNet(DDPM):
d = to_d(x, sigma_hat, denoised)
if callback is not None:
callback({'x': x, 'i': i, 'sigma': sigmas[i], 'sigma_hat': sigma_hat, 'denoised': denoised})
+
+ if img_callback: yield from img_callback(x, i)
+
dt = sigmas[i + 1] - sigma_hat
# Euler method
x = x + d * dt
- return x
+
+ yield from img_callback(x, len(sigmas)-1)
@torch.no_grad()
- def euler_ancestral_sampling(self,ac,x, S, cond, unconditional_conditioning = None, unconditional_guidance_scale = 1,extra_args=None, callback=None, disable=None):
+ def euler_ancestral_sampling(self,ac,x, S, cond, unconditional_conditioning = None, unconditional_guidance_scale = 1,extra_args=None, callback=None, disable=None,
+ img_callback=None):
"""Ancestral sampling with Euler method steps."""
extra_args = {} if extra_args is None else extra_args
@@ -822,6 +843,8 @@ class UNet(DDPM):
sigmas = cvd.get_sigmas(S)
x = x*sigmas[0]
+ print(f"Running Euler Ancestral Sampling with {len(sigmas) - 1} timesteps")
+
s_in = x.new_ones([x.shape[0]]).half()
for i in trange(len(sigmas) - 1, disable=disable):
@@ -837,17 +860,22 @@ class UNet(DDPM):
sigma_down, sigma_up = get_ancestral_step(sigmas[i], sigmas[i + 1])
if callback is not None:
callback({'x': x, 'i': i, 'sigma': sigmas[i], 'sigma_hat': sigmas[i], 'denoised': denoised})
+
+ if img_callback: yield from img_callback(x, i)
+
d = to_d(x, sigmas[i], denoised)
# Euler method
dt = sigma_down - sigmas[i]
x = x + d * dt
x = x + torch.randn_like(x) * sigma_up
- return x
+
+ yield from img_callback(x, len(sigmas)-1)
@torch.no_grad()
- def heun_sampling(self, ac, x, S, cond, unconditional_conditioning = None, unconditional_guidance_scale = 1, extra_args=None, callback=None, disable=None, s_churn=0., s_tmin=0., s_tmax=float('inf'), s_noise=1.):
+ def heun_sampling(self, ac, x, S, cond, unconditional_conditioning = None, unconditional_guidance_scale = 1, extra_args=None, callback=None, disable=None, s_churn=0., s_tmin=0., s_tmax=float('inf'), s_noise=1.,
+ img_callback=None):
"""Implements Algorithm 2 (Heun steps) from Karras et al. (2022)."""
extra_args = {} if extra_args is None else extra_args
@@ -855,6 +883,8 @@ class UNet(DDPM):
sigmas = cvd.get_sigmas(S)
x = x*sigmas[0]
+ print(f"Running Heun Sampling with {len(sigmas) - 1} timesteps")
+
s_in = x.new_ones([x.shape[0]]).half()
for i in trange(len(sigmas) - 1, disable=disable):
@@ -876,6 +906,9 @@ class UNet(DDPM):
d = to_d(x, sigma_hat, denoised)
if callback is not None:
callback({'x': x, 'i': i, 'sigma': sigmas[i], 'sigma_hat': sigma_hat, 'denoised': denoised})
+
+ if img_callback: yield from img_callback(x, i)
+
dt = sigmas[i + 1] - sigma_hat
if sigmas[i + 1] == 0:
# Euler method
@@ -895,11 +928,13 @@ class UNet(DDPM):
d_2 = to_d(x_2, sigmas[i + 1], denoised_2)
d_prime = (d + d_2) / 2
x = x + d_prime * dt
- return x
+
+ yield from img_callback(x, len(sigmas)-1)
@torch.no_grad()
- def dpm_2_sampling(self,ac,x, S, cond, unconditional_conditioning = None, unconditional_guidance_scale = 1,extra_args=None, callback=None, disable=None, s_churn=0., s_tmin=0., s_tmax=float('inf'), s_noise=1.):
+ def dpm_2_sampling(self,ac,x, S, cond, unconditional_conditioning = None, unconditional_guidance_scale = 1,extra_args=None, callback=None, disable=None, s_churn=0., s_tmin=0., s_tmax=float('inf'), s_noise=1.,
+ img_callback=None):
"""A sampler inspired by DPM-Solver-2 and Algorithm 2 from Karras et al. (2022)."""
extra_args = {} if extra_args is None else extra_args
@@ -907,6 +942,8 @@ class UNet(DDPM):
sigmas = cvd.get_sigmas(S)
x = x*sigmas[0]
+ print(f"Running DPM2 Sampling with {len(sigmas) - 1} timesteps")
+
s_in = x.new_ones([x.shape[0]]).half()
for i in trange(len(sigmas) - 1, disable=disable):
gamma = min(s_churn / (len(sigmas) - 1), 2 ** 0.5 - 1) if s_tmin <= sigmas[i] <= s_tmax else 0.
@@ -924,7 +961,7 @@ class UNet(DDPM):
e_t_uncond, e_t = (x_in + eps * c_out).chunk(2)
denoised = e_t_uncond + unconditional_guidance_scale * (e_t - e_t_uncond)
-
+ if img_callback: yield from img_callback(x, i)
d = to_d(x, sigma_hat, denoised)
# Midpoint method, where the midpoint is chosen according to a rho=3 Karras schedule
@@ -945,11 +982,13 @@ class UNet(DDPM):
d_2 = to_d(x_2, sigma_mid, denoised_2)
x = x + d_2 * dt_2
- return x
+
+ yield from img_callback(x, len(sigmas)-1)
@torch.no_grad()
- def dpm_2_ancestral_sampling(self,ac,x, S, cond, unconditional_conditioning = None, unconditional_guidance_scale = 1, extra_args=None, callback=None, disable=None):
+ def dpm_2_ancestral_sampling(self,ac,x, S, cond, unconditional_conditioning = None, unconditional_guidance_scale = 1, extra_args=None, callback=None, disable=None,
+ img_callback=None):
"""Ancestral sampling with DPM-Solver inspired second-order steps."""
extra_args = {} if extra_args is None else extra_args
@@ -957,6 +996,8 @@ class UNet(DDPM):
sigmas = cvd.get_sigmas(S)
x = x*sigmas[0]
+ print(f"Running DPM2 Ancestral Sampling with {len(sigmas) - 1} timesteps")
+
s_in = x.new_ones([x.shape[0]]).half()
for i in trange(len(sigmas) - 1, disable=disable):
@@ -973,6 +1014,9 @@ class UNet(DDPM):
sigma_down, sigma_up = get_ancestral_step(sigmas[i], sigmas[i + 1])
if callback is not None:
callback({'x': x, 'i': i, 'sigma': sigmas[i], 'sigma_hat': sigmas[i], 'denoised': denoised})
+
+ if img_callback: yield from img_callback(x, i)
+
d = to_d(x, sigmas[i], denoised)
# Midpoint method, where the midpoint is chosen according to a rho=3 Karras schedule
sigma_mid = ((sigmas[i] ** (1 / 3) + sigma_down ** (1 / 3)) / 2) ** 3
@@ -993,11 +1037,13 @@ class UNet(DDPM):
d_2 = to_d(x_2, sigma_mid, denoised_2)
x = x + d_2 * dt_2
x = x + torch.randn_like(x) * sigma_up
- return x
+
+ yield from img_callback(x, len(sigmas)-1)
@torch.no_grad()
- def lms_sampling(self,ac,x, S, cond, unconditional_conditioning = None, unconditional_guidance_scale = 1, extra_args=None, callback=None, disable=None, order=4):
+ def lms_sampling(self,ac,x, S, cond, unconditional_conditioning = None, unconditional_guidance_scale = 1, extra_args=None, callback=None, disable=None, order=4,
+ img_callback=None):
extra_args = {} if extra_args is None else extra_args
s_in = x.new_ones([x.shape[0]])
@@ -1005,6 +1051,8 @@ class UNet(DDPM):
sigmas = cvd.get_sigmas(S)
x = x*sigmas[0]
+ print(f"Running LMS Sampling with {len(sigmas) - 1} timesteps")
+
ds = []
for i in trange(len(sigmas) - 1, disable=disable):
@@ -1017,6 +1065,7 @@ class UNet(DDPM):
e_t_uncond, e_t = (x_in + eps * c_out).chunk(2)
denoised = e_t_uncond + unconditional_guidance_scale * (e_t - e_t_uncond)
+ if img_callback: yield from img_callback(x, i)
d = to_d(x, sigmas[i], denoised)
ds.append(d)
@@ -1027,4 +1076,5 @@ class UNet(DDPM):
cur_order = min(i + 1, order)
coeffs = [linear_multistep_coeff(cur_order, sigmas.cpu(), i, j) for j in range(cur_order)]
x = x + sum(coeff * d for coeff, d in zip(coeffs, reversed(ds)))
- return x
+
+ yield from img_callback(x, len(sigmas)-1)
diff --git a/optimizedSD/openaimodelSplit.py b/optimizedSD/openaimodelSplit.py
index abc3098..7a32ffe 100644
--- a/optimizedSD/openaimodelSplit.py
+++ b/optimizedSD/openaimodelSplit.py
@@ -13,7 +13,7 @@ from ldm.modules.diffusionmodules.util import (
normalization,
timestep_embedding,
)
-from splitAttention import SpatialTransformer
+from .splitAttention import SpatialTransformer
class AttentionPool2d(nn.Module):

181
ui/sd_internal/runtime.py
View File

@ -1,9 +1,10 @@
import json
import os, re import os, re
import traceback import traceback
import torch import torch
import numpy as np import numpy as np
from omegaconf import OmegaConf from omegaconf import OmegaConf
from PIL import Image from PIL import Image, ImageOps
from tqdm import tqdm, trange from tqdm import tqdm, trange
from itertools import islice from itertools import islice
from einops import rearrange from einops import rearrange
@ -32,10 +33,11 @@ filename_regex = re.compile('[^a-zA-Z0-9]')
from . import Request, Response, Image as ResponseImage from . import Request, Response, Image as ResponseImage
import base64 import base64
from io import BytesIO from io import BytesIO
#from colorama import Fore
# local # local
session_id = str(uuid.uuid4())[-8:]
stop_processing = False stop_processing = False
temp_images = {}
ckpt_file = None ckpt_file = None
gfpgan_file = None gfpgan_file = None
@ -184,23 +186,47 @@ def load_model_real_esrgan(real_esrgan_to_use):
print('loaded ', real_esrgan_to_use, 'to', device, 'precision', precision) print('loaded ', real_esrgan_to_use, 'to', device, 'precision', precision)
def mk_img(req: Request): def mk_img(req: Request):
global modelFS, device try:
yield from do_mk_img(req)
except Exception as e:
print(traceback.format_exc())
gc()
if device != "cpu":
modelFS.to("cpu")
modelCS.to("cpu")
model.model1.to("cpu")
model.model2.to("cpu")
gc()
yield json.dumps({
"status": 'failed',
"detail": str(e)
})
def do_mk_img(req: Request):
global model, modelCS, modelFS, device
global model_gfpgan, model_real_esrgan global model_gfpgan, model_real_esrgan
global stop_processing global stop_processing
stop_processing = False stop_processing = False
res = Response() res = Response()
res.session_id = session_id
res.request = req res.request = req
res.images = [] res.images = []
temp_images.clear()
model.turbo = req.turbo model.turbo = req.turbo
if req.use_cpu: if req.use_cpu:
if device != 'cpu': if device != 'cpu':
device = 'cpu' device = 'cpu'
if model_is_half: if model_is_half:
del model, modelCS, modelFS
load_model_ckpt(ckpt_file, device) load_model_ckpt(ckpt_file, device)
load_model_gfpgan(gfpgan_file) load_model_gfpgan(gfpgan_file)
@ -215,7 +241,8 @@ def mk_img(req: Request):
(req.init_image is None and model_fs_is_half) or \ (req.init_image is None and model_fs_is_half) or \
(req.init_image is not None and not model_fs_is_half and not force_full_precision): (req.init_image is not None and not model_fs_is_half and not force_full_precision):
load_model_ckpt(ckpt_file, device, model.turbo, unet_bs, ('full' if req.use_full_precision else 'autocast'), half_model_fs=(req.init_image is not None and not req.use_full_precision)) del model, modelCS, modelFS
load_model_ckpt(ckpt_file, device, req.turbo, unet_bs, ('full' if req.use_full_precision else 'autocast'), half_model_fs=(req.init_image is not None and not req.use_full_precision))
if prev_device != device: if prev_device != device:
load_model_gfpgan(gfpgan_file) load_model_gfpgan(gfpgan_file)
@ -248,6 +275,7 @@ def mk_img(req: Request):
opt_use_upscale = req.use_upscale opt_use_upscale = req.use_upscale
opt_show_only_filtered = req.show_only_filtered_image opt_show_only_filtered = req.show_only_filtered_image
opt_format = 'png' opt_format = 'png'
opt_sampler_name = req.sampler
print(req.to_string(), '\n device', device) print(req.to_string(), '\n device', device)
@ -265,6 +293,8 @@ def mk_img(req: Request):
else: else:
precision_scope = nullcontext precision_scope = nullcontext
mask = None
if req.init_image is None: if req.init_image is None:
handler = _txt2img handler = _txt2img
@ -284,18 +314,22 @@ def mk_img(req: Request):
init_image = repeat(init_image, '1 ... -> b ...', b=batch_size) init_image = repeat(init_image, '1 ... -> b ...', b=batch_size)
init_latent = modelFS.get_first_stage_encoding(modelFS.encode_first_stage(init_image)) # move to latent space init_latent = modelFS.get_first_stage_encoding(modelFS.encode_first_stage(init_image)) # move to latent space
if device != "cpu": if req.mask is not None:
mem = torch.cuda.memory_allocated() / 1e6 mask = load_mask(req.mask, opt_W, opt_H, init_latent.shape[2], init_latent.shape[3], True).to(device)
modelFS.to("cpu") mask = mask[0][0].unsqueeze(0).repeat(4, 1, 1).unsqueeze(0)
while torch.cuda.memory_allocated() / 1e6 >= mem: mask = repeat(mask, '1 ... -> b ...', b=batch_size)
time.sleep(1)
if device != "cpu" and precision == "autocast":
mask = mask.half()
move_fs_to_cpu()
assert 0. <= opt_strength <= 1., 'can only work with strength in [0.0, 1.0]' assert 0. <= opt_strength <= 1., 'can only work with strength in [0.0, 1.0]'
t_enc = int(opt_strength * opt_ddim_steps) t_enc = int(opt_strength * opt_ddim_steps)
print(f"target t_enc is {t_enc} steps") print(f"target t_enc is {t_enc} steps")
if opt_save_to_disk_path is not None: if opt_save_to_disk_path is not None:
session_out_path = os.path.join(opt_save_to_disk_path, session_id) session_out_path = os.path.join(opt_save_to_disk_path, req.session_id)
os.makedirs(session_out_path, exist_ok=True) os.makedirs(session_out_path, exist_ok=True)
else: else:
session_out_path = None session_out_path = None
@ -326,29 +360,60 @@ def mk_img(req: Request):
else: else:
c = modelCS.get_learned_conditioning(prompts) c = modelCS.get_learned_conditioning(prompts)
modelFS.to(device)
partial_x_samples = None partial_x_samples = None
def img_callback(x_samples, i): def img_callback(x_samples, i):
nonlocal partial_x_samples nonlocal partial_x_samples
partial_x_samples = x_samples partial_x_samples = x_samples
if req.stream_progress_updates:
n_steps = opt_ddim_steps if req.init_image is None else t_enc
progress = {"step": i, "total_steps": n_steps}
if req.stream_image_progress and i % 5 == 0:
partial_images = []
for i in range(batch_size):
x_samples_ddim = modelFS.decode_first_stage(x_samples[i].unsqueeze(0))
x_sample = torch.clamp((x_samples_ddim + 1.0) / 2.0, min=0.0, max=1.0)
x_sample = 255.0 * rearrange(x_sample[0].cpu().numpy(), "c h w -> h w c")
x_sample = x_sample.astype(np.uint8)
img = Image.fromarray(x_sample)
buf = BytesIO()
img.save(buf, format='JPEG')
buf.seek(0)
del img, x_sample, x_samples_ddim
# don't delete x_samples, it is used in the code that called this callback
temp_images[str(req.session_id) + '/' + str(i)] = buf
partial_images.append({'path': f'/image/tmp/{req.session_id}/{i}'})
progress['output'] = partial_images
yield json.dumps(progress)
if stop_processing: if stop_processing:
raise UserInitiatedStop("User requested that we stop processing") raise UserInitiatedStop("User requested that we stop processing")
# run the handler # run the handler
try: try:
if handler == _txt2img: if handler == _txt2img:
x_samples = _txt2img(opt_W, opt_H, opt_n_samples, opt_ddim_steps, opt_scale, None, opt_C, opt_f, opt_ddim_eta, c, uc, opt_seed, img_callback) x_samples = _txt2img(opt_W, opt_H, opt_n_samples, opt_ddim_steps, opt_scale, None, opt_C, opt_f, opt_ddim_eta, c, uc, opt_seed, img_callback, mask, opt_sampler_name)
else: else:
x_samples = _img2img(init_latent, t_enc, batch_size, opt_scale, c, uc, opt_ddim_steps, opt_ddim_eta, opt_seed, img_callback) x_samples = _img2img(init_latent, t_enc, batch_size, opt_scale, c, uc, opt_ddim_steps, opt_ddim_eta, opt_seed, img_callback, mask)
yield from x_samples
x_samples = partial_x_samples
except UserInitiatedStop: except UserInitiatedStop:
if partial_x_samples is None: if partial_x_samples is None:
continue continue
x_samples = partial_x_samples x_samples = partial_x_samples
modelFS.to(device)
print("saving images") print("saving images")
for i in range(batch_size): for i in range(batch_size):
@ -358,6 +423,14 @@ def mk_img(req: Request):
x_sample = x_sample.astype(np.uint8) x_sample = x_sample.astype(np.uint8)
img = Image.fromarray(x_sample) img = Image.fromarray(x_sample)
has_filters = (opt_use_face_correction is not None and opt_use_face_correction.startswith('GFPGAN')) or \
(opt_use_upscale is not None and opt_use_upscale.startswith('RealESRGAN'))
return_orig_img = not has_filters or not opt_show_only_filtered
if stop_processing:
return_orig_img = True
if opt_save_to_disk_path is not None: if opt_save_to_disk_path is not None:
prompt_flattened = filename_regex.sub('_', prompts[0]) prompt_flattened = filename_regex.sub('_', prompts[0])
prompt_flattened = prompt_flattened[:50] prompt_flattened = prompt_flattened[:50]
@ -368,12 +441,12 @@ def mk_img(req: Request):
img_out_path = os.path.join(session_out_path, f"{file_path}.{opt_format}") img_out_path = os.path.join(session_out_path, f"{file_path}.{opt_format}")
meta_out_path = os.path.join(session_out_path, f"{file_path}.txt") meta_out_path = os.path.join(session_out_path, f"{file_path}.txt")
if not opt_show_only_filtered: if return_orig_img:
save_image(img, img_out_path) save_image(img, img_out_path)
save_metadata(meta_out_path, prompts, opt_seed, opt_W, opt_H, opt_ddim_steps, opt_scale, opt_strength, opt_use_face_correction, opt_use_upscale) save_metadata(meta_out_path, prompts, opt_seed, opt_W, opt_H, opt_ddim_steps, opt_scale, opt_strength, opt_use_face_correction, opt_use_upscale, opt_sampler_name)
if not opt_show_only_filtered: if return_orig_img:
img_data = img_to_base64_str(img) img_data = img_to_base64_str(img)
res_image_orig = ResponseImage(data=img_data, seed=opt_seed) res_image_orig = ResponseImage(data=img_data, seed=opt_seed)
res.images.append(res_image_orig) res.images.append(res_image_orig)
@ -381,8 +454,10 @@ def mk_img(req: Request):
if opt_save_to_disk_path is not None: if opt_save_to_disk_path is not None:
res_image_orig.path_abs = img_out_path res_image_orig.path_abs = img_out_path
if (opt_use_face_correction is not None and opt_use_face_correction.startswith('GFPGAN')) or \ del img
(opt_use_upscale is not None and opt_use_upscale.startswith('RealESRGAN')):
if has_filters and not stop_processing:
print('Applying filters..')
gc() gc()
filters_applied = [] filters_applied = []
@ -410,18 +485,19 @@ def mk_img(req: Request):
save_image(filtered_image, filtered_img_out_path) save_image(filtered_image, filtered_img_out_path)
res_image_filtered.path_abs = filtered_img_out_path res_image_filtered.path_abs = filtered_img_out_path
del filtered_image
seeds += str(opt_seed) + "," seeds += str(opt_seed) + ","
opt_seed += 1 opt_seed += 1
if device != "cpu": move_fs_to_cpu()
mem = torch.cuda.memory_allocated() / 1e6 gc()
modelFS.to("cpu") del x_samples, x_samples_ddim, x_sample
while torch.cuda.memory_allocated() / 1e6 >= mem:
time.sleep(1)
del x_samples
print("memory_final = ", torch.cuda.memory_allocated() / 1e6) print("memory_final = ", torch.cuda.memory_allocated() / 1e6)
return res print('Task completed')
yield json.dumps(res.json())
def save_image(img, img_out_path): def save_image(img, img_out_path):
try: try:
@ -429,8 +505,8 @@ def save_image(img, img_out_path):
except: except:
print('could not save the file', traceback.format_exc()) print('could not save the file', traceback.format_exc())
def save_metadata(meta_out_path, prompts, opt_seed, opt_W, opt_H, opt_ddim_steps, opt_scale, opt_prompt_strength, opt_correct_face, opt_upscale): def save_metadata(meta_out_path, prompts, opt_seed, opt_W, opt_H, opt_ddim_steps, opt_scale, opt_prompt_strength, opt_correct_face, opt_upscale, sampler_name):
metadata = f"{prompts[0]}\nWidth: {opt_W}\nHeight: {opt_H}\nSeed: {opt_seed}\nSteps: {opt_ddim_steps}\nGuidance Scale: {opt_scale}\nPrompt Strength: {opt_prompt_strength}\nUse Face Correction: {opt_correct_face}\nUse Upscaling: {opt_upscale}" metadata = f"{prompts[0]}\nWidth: {opt_W}\nHeight: {opt_H}\nSeed: {opt_seed}\nSteps: {opt_ddim_steps}\nGuidance Scale: {opt_scale}\nPrompt Strength: {opt_prompt_strength}\nUse Face Correction: {opt_correct_face}\nUse Upscaling: {opt_upscale}\nSampler: {sampler_name}"
try: try:
with open(meta_out_path, 'w') as f: with open(meta_out_path, 'w') as f:
@ -438,7 +514,7 @@ def save_metadata(meta_out_path, prompts, opt_seed, opt_W, opt_H, opt_ddim_steps
except: except:
print('could not save the file', traceback.format_exc()) print('could not save the file', traceback.format_exc())
def _txt2img(opt_W, opt_H, opt_n_samples, opt_ddim_steps, opt_scale, start_code, opt_C, opt_f, opt_ddim_eta, c, uc, opt_seed, img_callback): def _txt2img(opt_W, opt_H, opt_n_samples, opt_ddim_steps, opt_scale, start_code, opt_C, opt_f, opt_ddim_eta, c, uc, opt_seed, img_callback, mask, sampler_name):
shape = [opt_n_samples, opt_C, opt_H // opt_f, opt_W // opt_f] shape = [opt_n_samples, opt_C, opt_H // opt_f, opt_W // opt_f]
if device != "cpu": if device != "cpu":
@ -458,12 +534,13 @@ def _txt2img(opt_W, opt_H, opt_n_samples, opt_ddim_steps, opt_scale, start_code,
eta=opt_ddim_eta, eta=opt_ddim_eta,
x_T=start_code, x_T=start_code,
img_callback=img_callback, img_callback=img_callback,
sampler = 'plms', mask=mask,
sampler = sampler_name,
) )
return samples_ddim yield from samples_ddim
def _img2img(init_latent, t_enc, batch_size, opt_scale, c, uc, opt_ddim_steps, opt_ddim_eta, opt_seed, img_callback): def _img2img(init_latent, t_enc, batch_size, opt_scale, c, uc, opt_ddim_steps, opt_ddim_eta, opt_seed, img_callback, mask):
# encode (scaled latent) # encode (scaled latent)
z_enc = model.stochastic_encode( z_enc = model.stochastic_encode(
init_latent, init_latent,
@ -472,6 +549,8 @@ def _img2img(init_latent, t_enc, batch_size, opt_scale, c, uc, opt_ddim_steps, o
opt_ddim_eta, opt_ddim_eta,
opt_ddim_steps, opt_ddim_steps,
) )
x_T = None if mask is None else init_latent
# decode it # decode it
samples_ddim = model.sample( samples_ddim = model.sample(
t_enc, t_enc,
@ -480,10 +559,19 @@ def _img2img(init_latent, t_enc, batch_size, opt_scale, c, uc, opt_ddim_steps, o
unconditional_guidance_scale=opt_scale, unconditional_guidance_scale=opt_scale,
unconditional_conditioning=uc, unconditional_conditioning=uc,
img_callback=img_callback, img_callback=img_callback,
mask=mask,
x_T=x_T,
sampler = 'ddim' sampler = 'ddim'
) )
return samples_ddim yield from samples_ddim
def move_fs_to_cpu():
if device != "cpu":
mem = torch.cuda.memory_allocated() / 1e6
modelFS.to("cpu")
while torch.cuda.memory_allocated() / 1e6 >= mem:
time.sleep(1)
def gc(): def gc():
if device == 'cpu': if device == 'cpu':
@ -525,6 +613,31 @@ def load_img(img_str, w0, h0):
image = torch.from_numpy(image) image = torch.from_numpy(image)
return 2.*image - 1. return 2.*image - 1.
def load_mask(mask_str, h0, w0, newH, newW, invert=False):
image = base64_str_to_img(mask_str).convert("RGB")
w, h = image.size
print(f"loaded input mask of size ({w}, {h})")
if invert:
print("inverted")
image = ImageOps.invert(image)
# where_0, where_1 = np.where(image == 0), np.where(image == 255)
# image[where_0], image[where_1] = 255, 0
if h0 is not None and w0 is not None:
h, w = h0, w0
w, h = map(lambda x: x - x % 64, (w, h)) # resize to integer multiple of 64
print(f"New mask size ({w}, {h})")
image = image.resize((newW, newH), resample=Image.Resampling.LANCZOS)
image = np.array(image)
image = image.astype(np.float32) / 255.0
image = image[None].transpose(0, 3, 1, 2)
image = torch.from_numpy(image)
return image
# https://stackoverflow.com/a/61114178 # https://stackoverflow.com/a/61114178
def img_to_base64_str(img): def img_to_base64_str(img):
buffered = BytesIO() buffered = BytesIO()

47
ui/server.py
View File

@ -15,7 +15,8 @@ CONFIG_DIR = os.path.join(SD_UI_DIR, '..', 'scripts')
OUTPUT_DIRNAME = "Stable Diffusion UI" # in the user's home folder OUTPUT_DIRNAME = "Stable Diffusion UI" # in the user's home folder
from fastapi import FastAPI, HTTPException from fastapi import FastAPI, HTTPException
from starlette.responses import FileResponse from fastapi.staticfiles import StaticFiles
from starlette.responses import FileResponse, StreamingResponse
from pydantic import BaseModel from pydantic import BaseModel
import logging import logging
@ -31,6 +32,7 @@ outpath = os.path.join(os.path.expanduser("~"), OUTPUT_DIRNAME)
# defaults from https://huggingface.co/blog/stable_diffusion # defaults from https://huggingface.co/blog/stable_diffusion
class ImageRequest(BaseModel): class ImageRequest(BaseModel):
session_id: str = "session"
prompt: str = "" prompt: str = ""
init_image: str = None # base64 init_image: str = None # base64
mask: str = None # base64 mask: str = None # base64
@ -41,6 +43,7 @@ class ImageRequest(BaseModel):
height: int = 512 height: int = 512
seed: int = 42 seed: int = 42
prompt_strength: float = 0.8 prompt_strength: float = 0.8
sampler: str = None # "ddim", "plms", "heun", "euler", "euler_a", "dpm2", "dpm2_a", "lms"
# allow_nsfw: bool = False # allow_nsfw: bool = False
save_to_disk_path: str = None save_to_disk_path: str = None
turbo: bool = True turbo: bool = True
@ -50,9 +53,14 @@ class ImageRequest(BaseModel):
use_upscale: str = None # or "RealESRGAN_x4plus" or "RealESRGAN_x4plus_anime_6B" use_upscale: str = None # or "RealESRGAN_x4plus" or "RealESRGAN_x4plus_anime_6B"
show_only_filtered_image: bool = False show_only_filtered_image: bool = False
stream_progress_updates: bool = False
stream_image_progress: bool = False
class SetAppConfigRequest(BaseModel): class SetAppConfigRequest(BaseModel):
update_branch: str = "main" update_branch: str = "main"
app.mount('/media', StaticFiles(directory=os.path.join(SD_UI_DIR, 'media/')), name="media")
@app.get('/') @app.get('/')
def read_root(): def read_root():
headers = {"Cache-Control": "no-cache, no-store, must-revalidate", "Pragma": "no-cache", "Expires": "0"} headers = {"Cache-Control": "no-cache, no-store, must-revalidate", "Pragma": "no-cache", "Expires": "0"}
@ -87,6 +95,7 @@ def image(req : ImageRequest):
from sd_internal import runtime from sd_internal import runtime
r = Request() r = Request()
r.session_id = req.session_id
r.prompt = req.prompt r.prompt = req.prompt
r.init_image = req.init_image r.init_image = req.init_image
r.mask = req.mask r.mask = req.mask
@ -97,6 +106,7 @@ def image(req : ImageRequest):
r.height = req.height r.height = req.height
r.seed = req.seed r.seed = req.seed
r.prompt_strength = req.prompt_strength r.prompt_strength = req.prompt_strength
r.sampler = req.sampler
# r.allow_nsfw = req.allow_nsfw # r.allow_nsfw = req.allow_nsfw
r.turbo = req.turbo r.turbo = req.turbo
r.use_cpu = req.use_cpu r.use_cpu = req.use_cpu
@ -106,10 +116,24 @@ def image(req : ImageRequest):
r.use_face_correction = req.use_face_correction r.use_face_correction = req.use_face_correction
r.show_only_filtered_image = req.show_only_filtered_image r.show_only_filtered_image = req.show_only_filtered_image
try: r.stream_progress_updates = True # the underlying implementation only supports streaming
res: Response = runtime.mk_img(r) r.stream_image_progress = req.stream_image_progress
return res.json() try:
if not req.stream_progress_updates:
r.stream_image_progress = False
res = runtime.mk_img(r)
if req.stream_progress_updates:
return StreamingResponse(res, media_type='application/json')
else: # compatibility mode: buffer the streaming responses, and return the last one
last_result = None
for result in res:
last_result = result
return json.loads(last_result)
except Exception as e: except Exception as e:
print(traceback.format_exc()) print(traceback.format_exc())
return HTTPException(status_code=500, detail=str(e)) return HTTPException(status_code=500, detail=str(e))
@ -128,6 +152,13 @@ def stop():
print(traceback.format_exc()) print(traceback.format_exc())
return HTTPException(status_code=500, detail=str(e)) return HTTPException(status_code=500, detail=str(e))
@app.get('/image/tmp/{session_id}/{img_id}')
def get_image(session_id, img_id):
from sd_internal import runtime
buf = runtime.temp_images[session_id + '/' + img_id]
buf.seek(0)
return StreamingResponse(buf, media_type='image/jpeg')
@app.post('/app_config') @app.post('/app_config')
async def setAppConfig(req : SetAppConfigRequest): async def setAppConfig(req : SetAppConfigRequest):
try: try:
@ -173,14 +204,6 @@ def getAppConfig():
print(traceback.format_exc()) print(traceback.format_exc())
return HTTPException(status_code=500, detail=str(e)) return HTTPException(status_code=500, detail=str(e))
@app.get('/media/ding.mp3')
def read_ding():
return FileResponse(os.path.join(SD_UI_DIR, 'media/ding.mp3'))
@app.get('/media/kofi.png')
def read_modifiers():
return FileResponse(os.path.join(SD_UI_DIR, 'media/kofi.png'))
@app.get('/modifiers.json') @app.get('/modifiers.json')
def read_modifiers(): def read_modifiers():
return FileResponse(os.path.join(SD_UI_DIR, 'modifiers.json')) return FileResponse(os.path.join(SD_UI_DIR, 'modifiers.json'))