Merge branch 'master' into enable-override-hypernet

CODEOWNERS

 *       @AUTOMATIC1111
-/localizations/ar_AR.json   @xmodar @blackneoo
-/localizations/de_DE.json   @LunixWasTaken
-/localizations/es_ES.json   @innovaciones
-/localizations/fr_FR.json   @tumbly
-/localizations/it_IT.json   @EugenioBuffo
-/localizations/ja_JP.json   @yuuki76
-/localizations/ko_KR.json   @36DB
-/localizations/pt_BR.json   @M-art-ucci
-/localizations/ru_RU.json   @kabachuha
-/localizations/tr_TR.json   @camenduru
-/localizations/zh_CN.json   @dtlnor @bgluminous
-/localizations/zh_TW.json   @benlisquare
+
+# if you were managing a localization and were removed from this file, this is because
+# the intended way to do localizations now is via extensions. See:
+# https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Developing-extensions
+# Make a repo with your localization and since you are still listed as a collaborator
+# you can add it to the wiki page yourself. This change is because some people complained
+# the git commit log is cluttered with things unrelated to almost everyone and
+# because I believe this is the best overall for the project to handle localizations almost
+# entirely without my oversight.
+
+

README.md

@@ -155,14 +155,15 @@ The documentation was moved from this README over to the project's [wiki](https:
 - Swin2SR - https://github.com/mv-lab/swin2sr
 - LDSR - https://github.com/Hafiidz/latent-diffusion
 - Ideas for optimizations - https://github.com/basujindal/stable-diffusion
-- Doggettx - Cross Attention layer optimization - https://github.com/Doggettx/stable-diffusion, original idea for prompt editing.
-- InvokeAI, lstein - Cross Attention layer optimization - https://github.com/invoke-ai/InvokeAI (originally http://github.com/lstein/stable-diffusion)
-- Rinon Gal - Textual Inversion - https://github.com/rinongal/textual_inversion (we're not using his code, but we are using his ideas).
+- Cross Attention layer optimization - Doggettx - https://github.com/Doggettx/stable-diffusion, original idea for prompt editing.
+- Cross Attention layer optimization - InvokeAI, lstein - https://github.com/invoke-ai/InvokeAI (originally http://github.com/lstein/stable-diffusion)
+- Textual Inversion - Rinon Gal - https://github.com/rinongal/textual_inversion (we're not using his code, but we are using his ideas).
 - Idea for SD upscale - https://github.com/jquesnelle/txt2imghd
 - Noise generation for outpainting mk2 - https://github.com/parlance-zz/g-diffuser-bot
 - CLIP interrogator idea and borrowing some code - https://github.com/pharmapsychotic/clip-interrogator
 - Idea for Composable Diffusion - https://github.com/energy-based-model/Compositional-Visual-Generation-with-Composable-Diffusion-Models-PyTorch
 - xformers - https://github.com/facebookresearch/xformers
 - DeepDanbooru - interrogator for anime diffusers https://github.com/KichangKim/DeepDanbooru
+- Security advice - RyotaK
 - Initial Gradio script - posted on 4chan by an Anonymous user. Thank you Anonymous user.
 - (You)

javascript/edit-attention.js

 addEventListener('keydown', (event) => {
 	let target = event.originalTarget || event.composedPath()[0];
-	if (!target.hasAttribute("placeholder")) return;
-	if (!target.placeholder.toLowerCase().includes("prompt")) return;
+	if (!target.matches("#toprow textarea.gr-text-input[placeholder]")) return;
 	if (! (event.metaKey || event.ctrlKey)) return;
 
 

javascript/progressbar.js

@@ -3,8 +3,21 @@ global_progressbars = {}
 galleries = {}
 galleryObservers = {}
 
+// this tracks laumnches of window.setTimeout for progressbar to prevent starting a new timeout when the previous is still running
+timeoutIds = {}
+
 function check_progressbar(id_part, id_progressbar, id_progressbar_span, id_skip, id_interrupt, id_preview, id_gallery){
-    var progressbar = gradioApp().getElementById(id_progressbar)
+    // gradio 3.8's enlightened approach allows them to create two nested div elements inside each other with same id
+    // every time you use gr.HTML(elem_id='xxx'), so we handle this here
+    var progressbar = gradioApp().querySelector("#"+id_progressbar+" #"+id_progressbar)
+    var progressbarParent
+    if(progressbar){
+        progressbarParent = gradioApp().querySelector("#"+id_progressbar)
+    } else{
+        progressbar = gradioApp().getElementById(id_progressbar)
+        progressbarParent = null
+    }
+
     var skip = id_skip ? gradioApp().getElementById(id_skip) : null
     var interrupt = gradioApp().getElementById(id_interrupt)
     
@@ -26,18 +39,26 @@ function check_progressbar(id_part, id_progressbar, id_progressbar_span, id_skip
 	    global_progressbars[id_progressbar] = progressbar
 
         var mutationObserver = new MutationObserver(function(m){
+            if(timeoutIds[id_part]) return;
+
             preview = gradioApp().getElementById(id_preview)
             gallery = gradioApp().getElementById(id_gallery)
 
             if(preview != null && gallery != null){
                 preview.style.width = gallery.clientWidth + "px"
                 preview.style.height = gallery.clientHeight + "px"
+                if(progressbarParent) progressbar.style.width = progressbarParent.clientWidth + "px"
 
 				//only watch gallery if there is a generation process going on
                 check_gallery(id_gallery);
 
                 var progressDiv = gradioApp().querySelectorAll('#' + id_progressbar_span).length > 0;
-                if(!progressDiv){
+                if(progressDiv){
+                    timeoutIds[id_part] = window.setTimeout(function() {
+                        timeoutIds[id_part] = null
+                        requestMoreProgress(id_part, id_progressbar_span, id_skip, id_interrupt)
+                    }, 500)
+                } else{
                     if (skip) {
                         skip.style.display = "none"
                     }
@@ -49,11 +70,8 @@ function check_progressbar(id_part, id_progressbar, id_progressbar_span, id_skip
                         galleries[id_gallery] = null;
                     }
                 }
-
-
             }
 
-            window.setTimeout(function() { requestMoreProgress(id_part, id_progressbar_span, id_skip, id_interrupt) }, 500)
         });
         mutationObserver.observe( progressbar, { childList:true, subtree:true })
 	}

javascript/ui.js

@@ -208,4 +208,6 @@ function update_token_counter(button_id) {
 function restart_reload(){
     document.body.innerHTML='<h1 style="font-family:monospace;margin-top:20%;color:lightgray;text-align:center;">Reloading...</h1>';
     setTimeout(function(){location.reload()},2000)
+
+    return []
 }

launch.py

@@ -142,7 +142,7 @@ def prepare_enviroment():
 
     stable_diffusion_commit_hash = os.environ.get('STABLE_DIFFUSION_COMMIT_HASH', "69ae4b35e0a0f6ee1af8bb9a5d0016ccb27e36dc")
     taming_transformers_commit_hash = os.environ.get('TAMING_TRANSFORMERS_COMMIT_HASH', "24268930bf1dce879235a7fddd0b2355b84d7ea6")
-    k_diffusion_commit_hash = os.environ.get('K_DIFFUSION_COMMIT_HASH', "f4e99857772fc3a126ba886aadf795a332774878")
+    k_diffusion_commit_hash = os.environ.get('K_DIFFUSION_COMMIT_HASH', "60e5042ca0da89c14d1dd59d73883280f8fce991")
     codeformer_commit_hash = os.environ.get('CODEFORMER_COMMIT_HASH', "c5b4593074ba6214284d6acd5f1719b6c5d739af")
     blip_commit_hash = os.environ.get('BLIP_COMMIT_HASH', "48211a1594f1321b00f14c9f7a5b4813144b2fb9")
 
@@ -238,12 +238,15 @@ def tests(argv):
     proc.kill()
 
 
-def start_webui():
-    print(f"Launching Web UI with arguments: {' '.join(sys.argv[1:])}")
+def start():
+    print(f"Launching {'API server' if '--nowebui' in sys.argv else 'Web UI'} with arguments: {' '.join(sys.argv[1:])}")
     import webui
+    if '--nowebui' in sys.argv:
+        webui.api_only()
+    else:
         webui.webui()
 
 
 if __name__ == "__main__":
     prepare_enviroment()
-    start_webui()
+    start()

localizations/de_DE.json

@@ -70,7 +70,7 @@
     "None": "Nichts",
     "Prompt matrix": "Promptmatrix",
     "Prompts from file or textbox": "Prompts aus Datei oder Textfeld",
-    "X/Y plot": "X/Y Graf",
+    "X/Y plot": "X/Y Graph",
     "Put variable parts at start of prompt": "Variable teile am start des Prompt setzen",
     "Iterate seed every line": "Iterate seed every line",
     "List of prompt inputs": "List of prompt inputs",

modules/api/api.py

+import base64
+import io
 import time
 import uvicorn
+from threading import Lock
 from gradio.processing_utils import encode_pil_to_base64, decode_base64_to_file, decode_base64_to_image
-from fastapi import APIRouter, Depends, HTTPException
+from fastapi import APIRouter, Depends, FastAPI, HTTPException
 import modules.shared as shared
-from modules import devices
 from modules.api.models import *
 from modules.processing import StableDiffusionProcessingTxt2Img, StableDiffusionProcessingImg2Img, process_images
 from modules.sd_samplers import all_samplers
 from modules.extras import run_extras, run_pnginfo
-
+from PIL import PngImagePlugin
+from modules.sd_models import checkpoints_list
+from modules.realesrgan_model import get_realesrgan_models
+from typing import List
 
 def upscaler_to_index(name: str):
     try:
@@ -29,8 +34,26 @@ def setUpscalers(req: dict):
     return reqDict
 
 
+def encode_pil_to_base64(image):
+    with io.BytesIO() as output_bytes:
+
+        # Copy any text-only metadata
+        use_metadata = False
+        metadata = PngImagePlugin.PngInfo()
+        for key, value in image.info.items():
+            if isinstance(key, str) and isinstance(value, str):
+                metadata.add_text(key, value)
+                use_metadata = True
+
+        image.save(
+            output_bytes, "PNG", pnginfo=(metadata if use_metadata else None)
+        )
+        bytes_data = output_bytes.getvalue()
+    return base64.b64encode(bytes_data)
+
+
 class Api:
-    def __init__(self, app, queue_lock):
+    def __init__(self, app: FastAPI, queue_lock: Lock):
         self.router = APIRouter()
         self.app = app
         self.queue_lock = queue_lock
@@ -40,7 +63,20 @@ class Api:
         self.app.add_api_route("/sdapi/v1/extra-batch-images", self.extras_batch_images_api, methods=["POST"], response_model=ExtrasBatchImagesResponse)
         self.app.add_api_route("/sdapi/v1/png-info", self.pnginfoapi, methods=["POST"], response_model=PNGInfoResponse)
         self.app.add_api_route("/sdapi/v1/progress", self.progressapi, methods=["GET"], response_model=ProgressResponse)
+        self.app.add_api_route("/sdapi/v1/interrogate", self.interrogateapi, methods=["POST"])
         self.app.add_api_route("/sdapi/v1/interrupt", self.interruptapi, methods=["POST"])
+        self.app.add_api_route("/sdapi/v1/options", self.get_config, methods=["GET"], response_model=OptionsModel)
+        self.app.add_api_route("/sdapi/v1/options", self.set_config, methods=["POST"])
+        self.app.add_api_route("/sdapi/v1/cmd-flags", self.get_cmd_flags, methods=["GET"], response_model=FlagsModel)
+        self.app.add_api_route("/sdapi/v1/samplers", self.get_samplers, methods=["GET"], response_model=List[SamplerItem])
+        self.app.add_api_route("/sdapi/v1/upscalers", self.get_upscalers, methods=["GET"], response_model=List[UpscalerItem])
+        self.app.add_api_route("/sdapi/v1/sd-models", self.get_sd_models, methods=["GET"], response_model=List[SDModelItem])
+        self.app.add_api_route("/sdapi/v1/hypernetworks", self.get_hypernetworks, methods=["GET"], response_model=List[HypernetworkItem])
+        self.app.add_api_route("/sdapi/v1/face-restorers", self.get_face_restorers, methods=["GET"], response_model=List[FaceRestorerItem])
+        self.app.add_api_route("/sdapi/v1/realesrgan-models", self.get_realesrgan_models, methods=["GET"], response_model=List[RealesrganItem])
+        self.app.add_api_route("/sdapi/v1/prompt-styles", self.get_promp_styles, methods=["GET"], response_model=List[PromptStyleItem])
+        self.app.add_api_route("/sdapi/v1/artist-categories", self.get_artists_categories, methods=["GET"], response_model=List[str])
+        self.app.add_api_route("/sdapi/v1/artists", self.get_artists, methods=["GET"], response_model=List[ArtistItem])
 
     def text2imgapi(self, txt2imgreq: StableDiffusionTxt2ImgProcessingAPI):
         sampler_index = sampler_to_index(txt2imgreq.sampler_index)
@@ -171,17 +207,96 @@ class Api:
 
         progress = min(progress, 1)
 
+        shared.state.set_current_image()
+
         current_image = None
         if shared.state.current_image and not req.skip_current_image:
             current_image = encode_pil_to_base64(shared.state.current_image)
 
         return ProgressResponse(progress=progress, eta_relative=eta_relative, state=shared.state.dict(), current_image=current_image)
 
+    def interrogateapi(self, interrogatereq: InterrogateRequest):
+        image_b64 = interrogatereq.image
+        if image_b64 is None:
+            raise HTTPException(status_code=404, detail="Image not found") 
+
+        img = self.__base64_to_image(image_b64)
+
+        # Override object param
+        with self.queue_lock:
+            processed = shared.interrogator.interrogate(img)
+        
+        return InterrogateResponse(caption=processed)
+
     def interruptapi(self):
         shared.state.interrupt()
 
         return {}
 
+    def get_config(self):
+        options = {}
+        for key in shared.opts.data.keys():
+            metadata = shared.opts.data_labels.get(key)
+            if(metadata is not None):
+                options.update({key: shared.opts.data.get(key, shared.opts.data_labels.get(key).default)})
+            else:
+                options.update({key: shared.opts.data.get(key, None)})
+
+        return options
+
+    def set_config(self, req: OptionsModel):
+        # currently req has all options fields even if you send a dict like { "send_seed": false }, which means it will
+        # overwrite all options with default values.
+        raise RuntimeError('Setting options via API is not supported')
+
+        reqDict = vars(req)
+        for o in reqDict:
+            setattr(shared.opts, o, reqDict[o])
+
+        shared.opts.save(shared.config_filename)
+        return
+
+    def get_cmd_flags(self):
+        return vars(shared.cmd_opts)
+
+    def get_samplers(self):
+        return [{"name":sampler[0], "aliases":sampler[2], "options":sampler[3]} for sampler in all_samplers]
+
+    def get_upscalers(self):
+        upscalers = []
+
+        for upscaler in shared.sd_upscalers:
+            u = upscaler.scaler
+            upscalers.append({"name":u.name, "model_name":u.model_name, "model_path":u.model_path, "model_url":u.model_url})
+
+        return upscalers
+
+    def get_sd_models(self):
+        return [{"title":x.title, "model_name":x.model_name, "hash":x.hash, "filename": x.filename, "config": x.config} for x in checkpoints_list.values()]
+
+    def get_hypernetworks(self):
+        return [{"name": name, "path": shared.hypernetworks[name]} for name in shared.hypernetworks]
+
+    def get_face_restorers(self):
+        return [{"name":x.name(), "cmd_dir": getattr(x, "cmd_dir", None)} for x in shared.face_restorers]
+
+    def get_realesrgan_models(self):
+        return [{"name":x.name,"path":x.data_path, "scale":x.scale} for x in get_realesrgan_models(None)]
+
+    def get_promp_styles(self):
+        styleList = []
+        for k in shared.prompt_styles.styles:
+            style = shared.prompt_styles.styles[k]
+            styleList.append({"name":style[0], "prompt": style[1], "negative_prompr": style[2]})
+
+        return styleList
+
+    def get_artists_categories(self):
+        return shared.artist_db.cats
+
+    def get_artists(self):
+        return [{"name":x[0], "score":x[1], "category":x[2]} for x in shared.artist_db.artists]
+
     def launch(self, server_name, port):
         self.app.include_router(self.router)
         uvicorn.run(self.app, host=server_name, port=port)

modules/api/models.py

 import inspect
-from click import prompt
 from pydantic import BaseModel, Field, create_model
 from typing import Any, Optional
 from typing_extensions import Literal
 from inflection import underscore
 from modules.processing import StableDiffusionProcessingTxt2Img, StableDiffusionProcessingImg2Img
-from modules.shared import sd_upscalers
+from modules.shared import sd_upscalers, opts, parser
+from typing import Dict, List
 
 API_NOT_ALLOWED = [
     "self",
@@ -65,6 +65,7 @@ class PydanticModelGenerator:
 
         self._model_name = model_name
         self._class_data = merge_class_params(class_instance)
+
         self._model_def = [
             ModelDef(
                 field=underscore(k),
@@ -109,12 +110,12 @@ StableDiffusionImg2ImgProcessingAPI = PydanticModelGenerator(
 ).generate_model()
 
 class TextToImageResponse(BaseModel):
-    images: list[str] = Field(default=None, title="Image", description="The generated image in base64 format.")
+    images: List[str] = Field(default=None, title="Image", description="The generated image in base64 format.")
     parameters: dict
     info: str
 
 class ImageToImageResponse(BaseModel):
-    images: list[str] = Field(default=None, title="Image", description="The generated image in base64 format.")
+    images: List[str] = Field(default=None, title="Image", description="The generated image in base64 format.")
     parameters: dict
     info: str
 
@@ -131,6 +132,7 @@ class ExtrasBaseRequest(BaseModel):
     upscaler_1: str = Field(default="None", title="Main upscaler", description=f"The name of the main upscaler to use, it has to be one of this list: {' , '.join([x.name for x in sd_upscalers])}")
     upscaler_2: str = Field(default="None", title="Secondary upscaler", description=f"The name of the secondary upscaler to use, it has to be one of this list: {' , '.join([x.name for x in sd_upscalers])}")
     extras_upscaler_2_visibility: float = Field(default=0, title="Secondary upscaler visibility", ge=0, le=1, allow_inf_nan=False, description="Sets the visibility of secondary upscaler, values should be between 0 and 1.")
+    upscale_first: bool = Field(default=False, title="Upscale first", description="Should the upscaler run before restoring faces?")
 
 class ExtraBaseResponse(BaseModel):
     html_info: str = Field(title="HTML info", description="A series of HTML tags containing the process info.")
@@ -146,10 +148,10 @@ class FileData(BaseModel):
     name: str = Field(title="File name")
 
 class ExtrasBatchImagesRequest(ExtrasBaseRequest):
-    imageList: list[FileData] = Field(title="Images", description="List of images to work on. Must be Base64 strings")
+    imageList: List[FileData] = Field(title="Images", description="List of images to work on. Must be Base64 strings")
 
 class ExtrasBatchImagesResponse(ExtraBaseResponse):
-    images: list[str] = Field(title="Images", description="The generated images in base64 format.")
+    images: List[str] = Field(title="Images", description="The generated images in base64 format.")
 
 class PNGInfoRequest(BaseModel):
     image: str = Field(title="Image", description="The base64 encoded PNG image")
@@ -165,3 +167,75 @@ class ProgressResponse(BaseModel):
     eta_relative: float = Field(title="ETA in secs")
     state: dict = Field(title="State", description="The current state snapshot")
     current_image: str = Field(default=None, title="Current image", description="The current image in base64 format. opts.show_progress_every_n_steps is required for this to work.")
+
+class InterrogateRequest(BaseModel):
+    image: str = Field(default="", title="Image", description="Image to work on, must be a Base64 string containing the image's data.")
+
+class InterrogateResponse(BaseModel):
+    caption: str = Field(default=None, title="Caption", description="The generated caption for the image.")
+
+fields = {}
+for key, value in opts.data.items():
+    metadata = opts.data_labels.get(key)
+    optType = opts.typemap.get(type(value), type(value))
+
+    if (metadata is not None):
+        fields.update({key: (Optional[optType], Field(
+            default=metadata.default ,description=metadata.label))})
+    else:
+        fields.update({key: (Optional[optType], Field())})
+
+OptionsModel = create_model("Options", **fields)
+
+flags = {}
+_options = vars(parser)['_option_string_actions']
+for key in _options:
+    if(_options[key].dest != 'help'):
+        flag = _options[key]
+        _type = str
+        if _options[key].default is not None: _type = type(_options[key].default)
+        flags.update({flag.dest: (_type,Field(default=flag.default, description=flag.help))})
+
+FlagsModel = create_model("Flags", **flags)
+
+class SamplerItem(BaseModel):
+    name: str = Field(title="Name")
+    aliases: List[str] = Field(title="Aliases")
+    options: Dict[str, str] = Field(title="Options")
+
+class UpscalerItem(BaseModel):
+    name: str = Field(title="Name")
+    model_name: Optional[str] = Field(title="Model Name")
+    model_path: Optional[str] = Field(title="Path")
+    model_url: Optional[str] = Field(title="URL")
+
+class SDModelItem(BaseModel):
+    title: str = Field(title="Title")
+    model_name: str = Field(title="Model Name")
+    hash: str = Field(title="Hash")
+    filename: str = Field(title="Filename")
+    config: str = Field(title="Config file")
+
+class HypernetworkItem(BaseModel):
+    name: str = Field(title="Name")
+    path: Optional[str] = Field(title="Path")
+
+class FaceRestorerItem(BaseModel):
+    name: str = Field(title="Name")
+    cmd_dir: Optional[str] = Field(title="Path")
+
+class RealesrganItem(BaseModel):
+    name: str = Field(title="Name")
+    path: Optional[str] = Field(title="Path")
+    scale: Optional[int] = Field(title="Scale")
+
+class PromptStyleItem(BaseModel):
+    name: str = Field(title="Name")
+    prompt: Optional[str] = Field(title="Prompt")
+    negative_prompt: Optional[str] = Field(title="Negative Prompt")
+
+class ArtistItem(BaseModel):
+    name: str = Field(title="Name")
+    score: float = Field(title="Score")
+    category: str = Field(title="Category")
+

modules/esrgan_model.py

@@ -50,6 +50,7 @@ def mod2normal(state_dict):
 def resrgan2normal(state_dict, nb=23):
     # this code is copied from https://github.com/victorca25/iNNfer
     if "conv_first.weight" in state_dict and "body.0.rdb1.conv1.weight" in state_dict:
+        re8x = 0
         crt_net = {}
         items = []
         for k, v in state_dict.items():
@@ -75,10 +76,18 @@ def resrgan2normal(state_dict, nb=23):
         crt_net['model.3.bias'] = state_dict['conv_up1.bias']
         crt_net['model.6.weight'] = state_dict['conv_up2.weight']
         crt_net['model.6.bias'] = state_dict['conv_up2.bias']
-        crt_net['model.8.weight'] = state_dict['conv_hr.weight']
-        crt_net['model.8.bias'] = state_dict['conv_hr.bias']
-        crt_net['model.10.weight'] = state_dict['conv_last.weight']
-        crt_net['model.10.bias'] = state_dict['conv_last.bias']
+
+        if 'conv_up3.weight' in state_dict:
+            # modification supporting: https://github.com/ai-forever/Real-ESRGAN/blob/main/RealESRGAN/rrdbnet_arch.py
+            re8x = 3
+            crt_net['model.9.weight'] = state_dict['conv_up3.weight']
+            crt_net['model.9.bias'] = state_dict['conv_up3.bias']
+
+        crt_net[f'model.{8+re8x}.weight'] = state_dict['conv_hr.weight']
+        crt_net[f'model.{8+re8x}.bias'] = state_dict['conv_hr.bias']
+        crt_net[f'model.{10+re8x}.weight'] = state_dict['conv_last.weight']
+        crt_net[f'model.{10+re8x}.bias'] = state_dict['conv_last.bias']
+
         state_dict = crt_net
     return state_dict
 

modules/extensions.py

@@ -34,8 +34,11 @@ class Extension:
         if repo is None or repo.bare:
             self.remote = None
         else:
+            try:
                 self.remote = next(repo.remote().urls, None)
                 self.status = 'unknown'
+            except Exception:
+                self.remote = None
 
     def list_files(self, subdir, extension):
         from modules import scripts

modules/extras.py

@@ -136,12 +136,13 @@ def run_extras(extras_mode, resize_mode, image, image_folder, input_dir, output_
 
     def run_upscalers_blend(params: List[UpscaleParams], image: Image.Image, info: str) -> Tuple[Image.Image, str]:
         blended_result: Image.Image = None
+        image_hash: str = hash(np.array(image.getdata()).tobytes())
         for upscaler in params:
             upscale_args = (upscaler.upscaler_idx, upscaling_resize, resize_mode,
                             upscaling_resize_w, upscaling_resize_h, upscaling_crop)
-            cache_key = LruCache.Key(image_hash=hash(np.array(image.getdata()).tobytes()),
+            cache_key = LruCache.Key(image_hash=image_hash,
                                      info_hash=hash(info),
-                                     args_hash=hash((upscale_args, upscale_first)))
+                                     args_hash=hash(upscale_args))
             cached_entry = cached_images.get(cache_key)
             if cached_entry is None:
                 res = upscale(image, *upscale_args)

modules/hypernetworks/ui.py

@@ -9,7 +9,7 @@ from modules import devices, sd_hijack, shared
 from modules.hypernetworks import hypernetwork
 
 not_available = ["hardswish", "multiheadattention"]
-keys = ["linear"] + list(x for x in hypernetwork.HypernetworkModule.activation_dict.keys() if x not in not_available)
+keys = list(x for x in hypernetwork.HypernetworkModule.activation_dict.keys() if x not in not_available)
 
 def create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure=None, activation_func=None, weight_init=None, add_layer_norm=False, use_dropout=False):
     # Remove illegal characters from name.

modules/img2img.py

@@ -81,6 +81,7 @@ def img2img(mode: int, prompt: str, negative_prompt: str, prompt_style: str, pro
         mask = None
 
     # Use the EXIF orientation of photos taken by smartphones.
+    if image is not None:
         image = ImageOps.exif_transpose(image) 
 
     assert 0. <= denoising_strength <= 1., 'can only work with strength in [0.0, 1.0]'
@@ -137,6 +138,8 @@ def img2img(mode: int, prompt: str, negative_prompt: str, prompt_style: str, pro
         if processed is None:
             processed = process_images(p)
 
+    p.close()
+
     shared.total_tqdm.clear()
 
     generation_info_js = processed.js()

modules/ldsr_model_arch.py

@@ -101,8 +101,8 @@ class LDSR:
         down_sample_rate = target_scale / 4
         wd = width_og * down_sample_rate
         hd = height_og * down_sample_rate
-        width_downsampled_pre = int(wd)
-        height_downsampled_pre = int(hd)
+        width_downsampled_pre = int(np.ceil(wd))
+        height_downsampled_pre = int(np.ceil(hd))
 
         if down_sample_rate != 1:
             print(
@@ -110,7 +110,12 @@ class LDSR:
             im_og = im_og.resize((width_downsampled_pre, height_downsampled_pre), Image.LANCZOS)
         else:
             print(f"Down sample rate is 1 from {target_scale} / 4 (Not downsampling)")
-        logs = self.run(model["model"], im_og, diffusion_steps, eta)
+        
+        # pad width and height to multiples of 64, pads with the edge values of image to avoid artifacts
+        pad_w, pad_h = np.max(((2, 2), np.ceil(np.array(im_og.size) / 64).astype(int)), axis=0) * 64 - im_og.size
+        im_padded = Image.fromarray(np.pad(np.array(im_og), ((0, pad_h), (0, pad_w), (0, 0)), mode='edge'))
+        
+        logs = self.run(model["model"], im_padded, diffusion_steps, eta)
 
         sample = logs["sample"]
         sample = sample.detach().cpu()
@@ -120,6 +125,9 @@ class LDSR:
         sample = np.transpose(sample, (0, 2, 3, 1))
         a = Image.fromarray(sample[0])
 
+        # remove padding
+        a = a.crop((0, 0) + tuple(np.array(im_og.size) * 4))
+
         del model
         gc.collect()
         torch.cuda.empty_cache()

modules/localization.py

@@ -3,6 +3,7 @@ import os
 import sys
 import traceback
 
+
 localizations = {}
 
 
@@ -16,6 +17,11 @@ def list_localizations(dirname):
 
         localizations[fn] = os.path.join(dirname, file)
 
+    from modules import scripts
+    for file in scripts.list_scripts("localizations", ".json"):
+        fn, ext = os.path.splitext(file.filename)
+        localizations[fn] = file.path
+
 
 def localization_js(current_localization_name):
     fn = localizations.get(current_localization_name, None)

modules/masking.py

@@ -49,7 +49,7 @@ def expand_crop_region(crop_region, processing_width, processing_height, image_w
     ratio_processing = processing_width / processing_height
 
     if ratio_crop_region > ratio_processing:
-        desired_height = (x2 - x1) * ratio_processing
+        desired_height = (x2 - x1) / ratio_processing
         desired_height_diff = int(desired_height - (y2-y1))
         y1 -= desired_height_diff//2
         y2 += desired_height_diff - desired_height_diff//2

modules/modelloader.py

@@ -85,6 +85,9 @@ def cleanup_models():
     src_path = os.path.join(root_path, "ESRGAN")
     dest_path = os.path.join(models_path, "ESRGAN")
     move_files(src_path, dest_path)
+    src_path = os.path.join(models_path, "BSRGAN")
+    dest_path = os.path.join(models_path, "ESRGAN")
+    move_files(src_path, dest_path, ".pth")
     src_path = os.path.join(root_path, "gfpgan")
     dest_path = os.path.join(models_path, "GFPGAN")
     move_files(src_path, dest_path)

modules/processing.py

@@ -134,11 +134,7 @@ class StableDiffusionProcessing():
             # Dummy zero conditioning if we're not using inpainting model.
             # Still takes up a bit of memory, but no encoder call.
             # Pretty sure we can just make this a 1x1 image since its not going to be used besides its batch size.
-            return torch.zeros(
-                x.shape[0], 5, 1, 1, 
-                dtype=x.dtype, 
-                device=x.device
-            )
+            return x.new_zeros(x.shape[0], 5, 1, 1)
 
         height = height or self.height
         width = width or self.width
@@ -156,11 +152,7 @@ class StableDiffusionProcessing():
     def img2img_image_conditioning(self, source_image, latent_image, image_mask = None):
         if self.sampler.conditioning_key not in {'hybrid', 'concat'}:
             # Dummy zero conditioning if we're not using inpainting model.
-            return torch.zeros(
-                latent_image.shape[0], 5, 1, 1,
-                dtype=latent_image.dtype,
-                device=latent_image.device
-            )
+            return latent_image.new_zeros(latent_image.shape[0], 5, 1, 1)
 
         # Handle the different mask inputs
         if image_mask is not None:
@@ -174,11 +166,11 @@ class StableDiffusionProcessing():
                 # Inpainting model uses a discretized mask as input, so we round to either 1.0 or 0.0
                 conditioning_mask = torch.round(conditioning_mask)
         else:
-            conditioning_mask = torch.ones(1, 1, *source_image.shape[-2:])
+            conditioning_mask = source_image.new_ones(1, 1, *source_image.shape[-2:])
 
         # Create another latent image, this time with a masked version of the original input.
         # Smoothly interpolate between the masked and unmasked latent conditioning image using a parameter.
-        conditioning_mask = conditioning_mask.to(source_image.device)
+        conditioning_mask = conditioning_mask.to(source_image.device).to(source_image.dtype)
         conditioning_image = torch.lerp(
             source_image,
             source_image * (1.0 - conditioning_mask),
@@ -199,9 +191,13 @@ class StableDiffusionProcessing():
     def init(self, all_prompts, all_seeds, all_subseeds):
         pass
 
-    def sample(self, conditioning, unconditional_conditioning, seeds, subseeds, subseed_strength):
+    def sample(self, conditioning, unconditional_conditioning, seeds, subseeds, subseed_strength, prompts):
         raise NotImplementedError()
 
+    def close(self):
+        self.sd_model = None
+        self.sampler = None
+
 
 class Processed:
     def __init__(self, p: StableDiffusionProcessing, images_list, seed=-1, info="", subseed=None, all_prompts=None, all_seeds=None, all_subseeds=None, index_of_first_image=0, infotexts=None):
@@ -507,6 +503,9 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed:
             if len(prompts) == 0:
                 break
 
+            if p.scripts is not None:
+                p.scripts.process_batch(p, batch_number=n, prompts=prompts, seeds=seeds, subseeds=subseeds)
+
             with devices.autocast():
                 uc = prompt_parser.get_learned_conditioning(shared.sd_model, len(prompts) * [p.negative_prompt], p.steps)
                 c = prompt_parser.get_multicond_learned_conditioning(shared.sd_model, prompts, p.steps)
@@ -519,7 +518,7 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed:
                 shared.state.job = f"Batch {n+1} out of {p.n_iter}"
 
             with devices.autocast():
-                samples_ddim = p.sample(conditioning=c, unconditional_conditioning=uc, seeds=seeds, subseeds=subseeds, subseed_strength=p.subseed_strength)
+                samples_ddim = p.sample(conditioning=c, unconditional_conditioning=uc, seeds=seeds, subseeds=subseeds, subseed_strength=p.subseed_strength, prompts=prompts)
 
             samples_ddim = samples_ddim.to(devices.dtype_vae)
             x_samples_ddim = decode_first_stage(p.sd_model, samples_ddim)
@@ -599,9 +598,6 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed:
     if p.scripts is not None:
         p.scripts.postprocess(p, res)
 
-    p.sd_model = None
-    p.sampler = None
-
     return res
 
 
@@ -650,7 +646,7 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing):
             self.truncate_x = int(self.firstphase_width - firstphase_width_truncated) // opt_f
             self.truncate_y = int(self.firstphase_height - firstphase_height_truncated) // opt_f
 
-    def sample(self, conditioning, unconditional_conditioning, seeds, subseeds, subseed_strength):
+    def sample(self, conditioning, unconditional_conditioning, seeds, subseeds, subseed_strength, prompts):
         self.sampler = sd_samplers.create_sampler_with_index(sd_samplers.samplers, self.sampler_index, self.sd_model)
 
         if not self.enable_hr:
@@ -663,9 +659,28 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing):
 
         samples = samples[:, :, self.truncate_y//2:samples.shape[2]-self.truncate_y//2, self.truncate_x//2:samples.shape[3]-self.truncate_x//2]
 
+        """saves image before applying hires fix, if enabled in options; takes as an arguyment either an image or batch with latent space images"""
+        def save_intermediate(image, index):
+            if not opts.save or self.do_not_save_samples or not opts.save_images_before_highres_fix:
+                return
+
+            if not isinstance(image, Image.Image):
+                image = sd_samplers.sample_to_image(image, index)
+
+            images.save_image(image, self.outpath_samples, "", seeds[index], prompts[index], opts.samples_format, suffix="-before-highres-fix")
+
         if opts.use_scale_latent_for_hires_fix:
+            for i in range(samples.shape[0]):
+                save_intermediate(samples, i)
+
             samples = torch.nn.functional.interpolate(samples, size=(self.height // opt_f, self.width // opt_f), mode="bilinear")
 
+            # Avoid making the inpainting conditioning unless necessary as 
+            # this does need some extra compute to decode / encode the image again.
+            if getattr(self, "inpainting_mask_weight", shared.opts.inpainting_mask_weight) < 1.0:
+                image_conditioning = self.img2img_image_conditioning(decode_first_stage(self.sd_model, samples), samples)
+            else:
+                image_conditioning = self.txt2img_image_conditioning(samples)
         else:
             decoded_samples = decode_first_stage(self.sd_model, samples)
             lowres_samples = torch.clamp((decoded_samples + 1.0) / 2.0, min=0.0, max=1.0)
@@ -675,6 +690,9 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing):
                 x_sample = 255. * np.moveaxis(x_sample.cpu().numpy(), 0, 2)
                 x_sample = x_sample.astype(np.uint8)
                 image = Image.fromarray(x_sample)
+
+                save_intermediate(image, i)
+
                 image = images.resize_image(0, image, self.width, self.height)
                 image = np.array(image).astype(np.float32) / 255.0
                 image = np.moveaxis(image, 2, 0)
@@ -686,14 +704,14 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing):
 
             samples = self.sd_model.get_first_stage_encoding(self.sd_model.encode_first_stage(decoded_samples))
 
+            image_conditioning = self.img2img_image_conditioning(decoded_samples, samples)
+
         shared.state.nextjob()
 
         self.sampler = sd_samplers.create_sampler_with_index(sd_samplers.samplers, self.sampler_index, self.sd_model)
 
         noise = create_random_tensors(samples.shape[1:], seeds=seeds, subseeds=subseeds, subseed_strength=subseed_strength, seed_resize_from_h=self.seed_resize_from_h, seed_resize_from_w=self.seed_resize_from_w, p=self)
 
-        image_conditioning = self.txt2img_image_conditioning(x)
-
         # GC now before running the next img2img to prevent running out of memory
         x = None
         devices.torch_gc()
@@ -832,8 +850,7 @@ class StableDiffusionProcessingImg2Img(StableDiffusionProcessing):
 
         self.image_conditioning = self.img2img_image_conditioning(image, self.init_latent, self.image_mask)
 
-
-    def sample(self, conditioning, unconditional_conditioning, seeds, subseeds, subseed_strength):
+    def sample(self, conditioning, unconditional_conditioning, seeds, subseeds, subseed_strength, prompts):
         x = create_random_tensors([opt_C, self.height // opt_f, self.width // opt_f], seeds=seeds, subseeds=subseeds, subseed_strength=self.subseed_strength, seed_resize_from_h=self.seed_resize_from_h, seed_resize_from_w=self.seed_resize_from_w, p=self)
 
         samples = self.sampler.sample_img2img(self, self.init_latent, x, conditioning, unconditional_conditioning, image_conditioning=self.image_conditioning)

modules/safe.py

@@ -23,11 +23,18 @@ def encode(*args):
 
 
 class RestrictedUnpickler(pickle.Unpickler):
+    extra_handler = None
+
     def persistent_load(self, saved_id):
         assert saved_id[0] == 'storage'
         return TypedStorage()
 
     def find_class(self, module, name):
+        if self.extra_handler is not None:
+            res = self.extra_handler(module, name)
+            if res is not None:
+                return res
+
         if module == 'collections' and name == 'OrderedDict':
             return getattr(collections, name)
         if module == 'torch._utils' and name in ['_rebuild_tensor_v2', '_rebuild_parameter']:
@@ -52,7 +59,7 @@ class RestrictedUnpickler(pickle.Unpickler):
             return set
 
         # Forbid everything else.
-        raise pickle.UnpicklingError(f"global '{module}/{name}' is forbidden")
+        raise Exception(f"global '{module}/{name}' is forbidden")
 
 
 allowed_zip_names = ["archive/data.pkl", "archive/version"]
@@ -69,7 +76,7 @@ def check_zip_filenames(filename, names):
         raise Exception(f"bad file inside {filename}: {name}")
 
 
-def check_pt(filename):
+def check_pt(filename, extra_handler):
     try:
 
         # new pytorch format is a zip file
@@ -78,6 +85,7 @@ def check_pt(filename):
 
             with z.open('archive/data.pkl') as file:
                 unpickler = RestrictedUnpickler(file)
+                unpickler.extra_handler = extra_handler
                 unpickler.load()
 
     except zipfile.BadZipfile:
@@ -85,16 +93,42 @@ def check_pt(filename):
         # if it's not a zip file, it's an olf pytorch format, with five objects written to pickle
         with open(filename, "rb") as file:
             unpickler = RestrictedUnpickler(file)
+            unpickler.extra_handler = extra_handler
             for i in range(5):
                 unpickler.load()
 
 
 def load(filename, *args, **kwargs):
+    return load_with_extra(filename, *args, **kwargs)
+
+
+def load_with_extra(filename, extra_handler=None, *args, **kwargs):
+    """
+    this functon is intended to be used by extensions that want to load models with
+    some extra classes in them that the usual unpickler would find suspicious.
+
+    Use the extra_handler argument to specify a function that takes module and field name as text,
+    and returns that field's value:
+
+    ```python
+    def extra(module, name):
+        if module == 'collections' and name == 'OrderedDict':
+            return collections.OrderedDict
+
+        return None
+
+    safe.load_with_extra('model.pt', extra_handler=extra)
+    ```
+
+    The alternative to this is just to use safe.unsafe_torch_load('model.pt'), which as the name implies is
+    definitely unsafe.
+    """
+
     from modules import shared
 
     try:
         if not shared.cmd_opts.disable_safe_unpickle:
-            check_pt(filename)
+            check_pt(filename, extra_handler)
 
     except pickle.UnpicklingError:
         print(f"Error verifying pickled file from {filename}:", file=sys.stderr)

modules/script_callbacks.py

@@ -2,6 +2,7 @@ import sys
 import traceback
 from collections import namedtuple
 import inspect
+from typing import Optional
 
 from fastapi import FastAPI
 from gradio import Blocks
@@ -26,25 +27,42 @@ class ImageSaveParams:
         """dictionary with parameters for image's PNG info data; infotext will have the key 'parameters'"""
 
 
+class CFGDenoiserParams:
+    def __init__(self, x, image_cond, sigma, sampling_step, total_sampling_steps):
+        self.x = x
+        """Latent image representation in the process of being denoised"""
+        
+        self.image_cond = image_cond
+        """Conditioning image"""
+        
+        self.sigma = sigma
+        """Current sigma noise step value"""
+        
+        self.sampling_step = sampling_step
+        """Current Sampling step number"""
+        
+        self.total_sampling_steps = total_sampling_steps
+        """Total number of sampling steps planned"""
+
+
 ScriptCallback = namedtuple("ScriptCallback", ["script", "callback"])
-callbacks_app_started = []
-callbacks_model_loaded = []
-callbacks_ui_tabs = []
-callbacks_ui_settings = []
-callbacks_before_image_saved = []
-callbacks_image_saved = []
+callback_map = dict(
+    callbacks_app_started=[],
+    callbacks_model_loaded=[],
+    callbacks_ui_tabs=[],
+    callbacks_ui_settings=[],
+    callbacks_before_image_saved=[],
+    callbacks_image_saved=[],
+    callbacks_cfg_denoiser=[]
+)
 
 
 def clear_callbacks():
-    callbacks_model_loaded.clear()
-    callbacks_ui_tabs.clear()
-    callbacks_ui_settings.clear()
-    callbacks_before_image_saved.clear()
-    callbacks_image_saved.clear()
+    for callback_list in callback_map.values():
+        callback_list.clear()
 
-
-def app_started_callback(demo: Blocks, app: FastAPI):
-    for c in callbacks_app_started:
+def app_started_callback(demo: Optional[Blocks], app: FastAPI):
+    for c in callback_map['callbacks_app_started']:
         try:
             c.callback(demo, app)
         except Exception:
@@ -52,7 +70,7 @@ def app_started_callback(demo: Blocks, app: FastAPI):
 
 
 def model_loaded_callback(sd_model):
-    for c in callbacks_model_loaded:
+    for c in callback_map['callbacks_model_loaded']:
         try:
             c.callback(sd_model)
         except Exception:
@@ -62,7 +80,7 @@ def model_loaded_callback(sd_model):
 def ui_tabs_callback():
     res = []
     
-    for c in callbacks_ui_tabs:
+    for c in callback_map['callbacks_ui_tabs']:
         try:
             res += c.callback() or []
         except Exception:
@@ -72,7 +90,7 @@ def ui_tabs_callback():
 
 
 def ui_settings_callback():
-    for c in callbacks_ui_settings:
+    for c in callback_map['callbacks_ui_settings']:
         try:
             c.callback()
         except Exception:
@@ -80,7 +98,7 @@ def ui_settings_callback():
 
 
 def before_image_saved_callback(params: ImageSaveParams):
-    for c in callbacks_before_image_saved:
+    for c in callback_map['callbacks_before_image_saved']:
         try:
             c.callback(params)
         except Exception:
@@ -88,13 +106,21 @@ def before_image_saved_callback(params: ImageSaveParams):
 
 
 def image_saved_callback(params: ImageSaveParams):
-    for c in callbacks_image_saved:
+    for c in callback_map['callbacks_image_saved']:
         try:
             c.callback(params)
         except Exception:
             report_exception(c, 'image_saved_callback')
 
 
+def cfg_denoiser_callback(params: CFGDenoiserParams):
+    for c in callback_map['callbacks_cfg_denoiser']:
+        try:
+            c.callback(params)
+        except Exception:
+            report_exception(c, 'cfg_denoiser_callback')
+
+
 def add_callback(callbacks, fun):
     stack = [x for x in inspect.stack() if x.filename != __file__]
     filename = stack[0].filename if len(stack) > 0 else 'unknown file'
@@ -102,16 +128,32 @@ def add_callback(callbacks, fun):
     callbacks.append(ScriptCallback(filename, fun))
 
     
+def remove_current_script_callbacks():
+    stack = [x for x in inspect.stack() if x.filename != __file__]
+    filename = stack[0].filename if len(stack) > 0 else 'unknown file'
+    if filename == 'unknown file':
+        return
+    for callback_list in callback_map.values():
+        for callback_to_remove in [cb for cb in callback_list if cb.script == filename]:
+            callback_list.remove(callback_to_remove)
+
+
+def remove_callbacks_for_function(callback_func):
+    for callback_list in callback_map.values():
+        for callback_to_remove in [cb for cb in callback_list if cb.callback == callback_func]:
+            callback_list.remove(callback_to_remove)
+
+
 def on_app_started(callback):
     """register a function to be called when the webui started, the gradio `Block` component and
     fastapi `FastAPI` object are passed as the arguments"""
-    add_callback(callbacks_app_started, callback)
+    add_callback(callback_map['callbacks_app_started'], callback)
 
 
 def on_model_loaded(callback):
     """register a function to be called when the stable diffusion model is created; the model is
     passed as an argument"""
-    add_callback(callbacks_model_loaded, callback)
+    add_callback(callback_map['callbacks_model_loaded'], callback)
 
 
 def on_ui_tabs(callback):
@@ -124,13 +166,13 @@ def on_ui_tabs(callback):
     title is tab text displayed to user in the UI
     elem_id is HTML id for the tab
     """
-    add_callback(callbacks_ui_tabs, callback)
+    add_callback(callback_map['callbacks_ui_tabs'], callback)
 
 
 def on_ui_settings(callback):
     """register a function to be called before UI settings are populated; add your settings
     by using shared.opts.add_option(shared.OptionInfo(...)) """
-    add_callback(callbacks_ui_settings, callback)
+    add_callback(callback_map['callbacks_ui_settings'], callback)
 
 
 def on_before_image_saved(callback):
@@ -138,7 +180,7 @@ def on_before_image_saved(callback):
     The callback is called with one argument:
         - params: ImageSaveParams - parameters the image is to be saved with. You can change fields in this object.
     """
-    add_callback(callbacks_before_image_saved, callback)
+    add_callback(callback_map['callbacks_before_image_saved'], callback)
 
 
 def on_image_saved(callback):
@@ -146,4 +188,12 @@ def on_image_saved(callback):
     The callback is called with one argument:
         - params: ImageSaveParams - parameters the image was saved with. Changing fields in this object does nothing.
     """
-    add_callback(callbacks_image_saved, callback)
+    add_callback(callback_map['callbacks_image_saved'], callback)
+
+
+def on_cfg_denoiser(callback):
+    """register a function to be called in the kdiffussion cfg_denoiser method after building the inner model inputs.
+    The callback is called with one argument:
+        - params: CFGDenoiserParams - parameters to be passed to the inner model and sampling state details.
+    """
+    add_callback(callback_map['callbacks_cfg_denoiser'], callback)

modules/scripts.py

@@ -3,7 +3,6 @@ import sys
 import traceback
 from collections import namedtuple
 
-import modules.ui as ui
 import gradio as gr
 
 from modules.processing import StableDiffusionProcessing
@@ -18,6 +17,9 @@ class Script:
     args_to = None
     alwayson = False
 
+    """A gr.Group component that has all script's UI inside it"""
+    group = None
+
     infotext_fields = None
     """if set in ui(), this is a list of pairs of gradio component + text; the text will be used when
     parsing infotext to set the value for the component; see ui.py's txt2img_paste_fields for an example
@@ -70,6 +72,19 @@ class Script:
 
         pass
 
+    def process_batch(self, p, *args, **kwargs):
+        """
+        Same as process(), but called for every batch.
+
+        **kwargs will have those items:
+          - batch_number - index of current batch, from 0 to number of batches-1
+          - prompts - list of prompts for current batch; you can change contents of this list but changing the number of entries will likely break things
+          - seeds - list of seeds for current batch
+          - subseeds - list of subseeds for current batch
+        """
+
+        pass
+
     def postprocess(self, p, processed, *args):
         """
         This function is called after processing ends for AlwaysVisible scripts.
@@ -218,8 +233,6 @@ class ScriptRunner:
 
             for control in controls:
                 control.custom_script_source = os.path.basename(script.filename)
-                if not script.alwayson:
-                    control.visible = False
 
             if script.infotext_fields is not None:
                 self.infotext_fields += script.infotext_fields
@@ -229,40 +242,41 @@ class ScriptRunner:
             script.args_to = len(inputs)
 
         for script in self.alwayson_scripts:
-            with gr.Group():
+            with gr.Group() as group:
                 create_script_ui(script, inputs, inputs_alwayson)
 
+            script.group = group
+
         dropdown = gr.Dropdown(label="Script", elem_id="script_list", choices=["None"] + self.titles, value="None", type="index")
         dropdown.save_to_config = True
         inputs[0] = dropdown
 
         for script in self.selectable_scripts:
+            with gr.Group(visible=False) as group:
                 create_script_ui(script, inputs, inputs_alwayson)
 
+            script.group = group
+
         def select_script(script_index):
-            if 0 < script_index <= len(self.selectable_scripts):
-                script = self.selectable_scripts[script_index-1]
-                args_from = script.args_from
-                args_to = script.args_to
-            else:
-                args_from = 0
-                args_to = 0
+            selected_script = self.selectable_scripts[script_index - 1] if script_index>0 else None
 
-            return [ui.gr_show(True if i == 0 else args_from <= i < args_to or is_alwayson) for i, is_alwayson in enumerate(inputs_alwayson)]
+            return [gr.update(visible=selected_script == s) for s in self.selectable_scripts]
 
         def init_field(title):
+            """called when an initial value is set from ui-config.json to show script's UI components"""
+
             if title == 'None':
                 return
+
             script_index = self.titles.index(title)
-            script = self.selectable_scripts[script_index]
-            for i in range(script.args_from, script.args_to):
-                inputs[i].visible = True
+            self.selectable_scripts[script_index].group.visible = True
 
         dropdown.init_field = init_field
+
         dropdown.change(
             fn=select_script,
             inputs=[dropdown],
-            outputs=inputs
+            outputs=[script.group for script in self.selectable_scripts]
         )
 
         return inputs
@@ -294,6 +308,15 @@ class ScriptRunner:
                 print(f"Error running process: {script.filename}", file=sys.stderr)
                 print(traceback.format_exc(), file=sys.stderr)
 
+    def process_batch(self, p, **kwargs):
+        for script in self.alwayson_scripts:
+            try:
+                script_args = p.script_args[script.args_from:script.args_to]
+                script.process_batch(p, *script_args, **kwargs)
+            except Exception:
+                print(f"Error running process_batch: {script.filename}", file=sys.stderr)
+                print(traceback.format_exc(), file=sys.stderr)
+
     def postprocess(self, p, processed):
         for script in self.alwayson_scripts:
             try:

modules/sd_models.py

@@ -9,7 +9,7 @@ from omegaconf import OmegaConf
 
 from ldm.util import instantiate_from_config
 
-from modules import shared, modelloader, devices, script_callbacks
+from modules import shared, modelloader, devices, script_callbacks, sd_vae
 from modules.paths import models_path
 from modules.sd_hijack_inpainting import do_inpainting_hijack, should_hijack_inpainting
 
@@ -159,13 +159,16 @@ def get_state_dict_from_checkpoint(pl_sd):
     return pl_sd
 
 
-vae_ignore_keys = {"model_ema.decay", "model_ema.num_updates"}
-
-
-def load_model_weights(model, checkpoint_info):
+def load_model_weights(model, checkpoint_info, vae_file="auto"):
     checkpoint_file = checkpoint_info.filename
     sd_model_hash = checkpoint_info.hash
 
+    if shared.opts.sd_checkpoint_cache > 0 and hasattr(model, "sd_checkpoint_info"):
+        sd_vae.restore_base_vae(model)
+        checkpoints_loaded[model.sd_checkpoint_info] = model.state_dict().copy()
+
+    vae_file = sd_vae.resolve_vae(checkpoint_file, vae_file=vae_file)
+
     if checkpoint_info not in checkpoints_loaded:
         print(f"Loading weights [{sd_model_hash}] from {checkpoint_file}")
 
@@ -182,37 +185,36 @@ def load_model_weights(model, checkpoint_info):
             model.to(memory_format=torch.channels_last)
 
         if not shared.cmd_opts.no_half:
+            vae = model.first_stage_model
+
+            # with --no-half-vae, remove VAE from model when doing half() to prevent its weights from being converted to float16
+            if shared.cmd_opts.no_half_vae:
+                model.first_stage_model = None
+
             model.half()
+            model.first_stage_model = vae
 
         devices.dtype = torch.float32 if shared.cmd_opts.no_half else torch.float16
         devices.dtype_vae = torch.float32 if shared.cmd_opts.no_half or shared.cmd_opts.no_half_vae else torch.float16
 
-        vae_file = os.path.splitext(checkpoint_file)[0] + ".vae.pt"
-
-        if not os.path.exists(vae_file) and shared.cmd_opts.vae_path is not None:
-            vae_file = shared.cmd_opts.vae_path
-
-        if os.path.exists(vae_file):
-            print(f"Loading VAE weights from: {vae_file}")
-            vae_ckpt = torch.load(vae_file, map_location=shared.weight_load_location)
-            vae_dict = {k: v for k, v in vae_ckpt["state_dict"].items() if k[0:4] != "loss" and k not in vae_ignore_keys}
-            model.first_stage_model.load_state_dict(vae_dict)
-
         model.first_stage_model.to(devices.dtype_vae)
 
+    else:
+        vae_name = sd_vae.get_filename(vae_file) if vae_file else None
+        vae_message = f" with {vae_name} VAE" if vae_name else ""
+        print(f"Loading weights [{sd_model_hash}]{vae_message} from cache")
+        model.load_state_dict(checkpoints_loaded[checkpoint_info])
+
     if shared.opts.sd_checkpoint_cache > 0:
-            checkpoints_loaded[checkpoint_info] = model.state_dict().copy()
         while len(checkpoints_loaded) > shared.opts.sd_checkpoint_cache:
             checkpoints_loaded.popitem(last=False)  # LRU
-    else:
-        print(f"Loading weights [{sd_model_hash}] from cache")
-        checkpoints_loaded.move_to_end(checkpoint_info)
-        model.load_state_dict(checkpoints_loaded[checkpoint_info])
 
     model.sd_model_hash = sd_model_hash
     model.sd_model_checkpoint = checkpoint_file
     model.sd_checkpoint_info = checkpoint_info
 
+    sd_vae.load_vae(model, vae_file)
+
 
 def load_model(checkpoint_info=None):
     from modules import lowvram, sd_hijack

modules/sd_samplers.py

@@ -12,6 +12,7 @@ from modules import prompt_parser, devices, processing, images
 
 from modules.shared import opts, cmd_opts, state
 import modules.shared as shared
+from modules.script_callbacks import CFGDenoiserParams, cfg_denoiser_callback
 
 
 SamplerData = namedtuple('SamplerData', ['name', 'constructor', 'aliases', 'options'])
@@ -23,11 +24,15 @@ samplers_k_diffusion = [
     ('Heun', 'sample_heun', ['k_heun'], {}),
     ('DPM2', 'sample_dpm_2', ['k_dpm_2'], {}),
     ('DPM2 a', 'sample_dpm_2_ancestral', ['k_dpm_2_a'], {}),
+    ('DPM++ 2S a', 'sample_dpmpp_2s_ancestral', ['k_dpmpp_2s_a'], {}),
+    ('DPM++ 2M', 'sample_dpmpp_2m', ['k_dpmpp_2m'], {}),
     ('DPM fast', 'sample_dpm_fast', ['k_dpm_fast'], {}),
     ('DPM adaptive', 'sample_dpm_adaptive', ['k_dpm_ad'], {}),
     ('LMS Karras', 'sample_lms', ['k_lms_ka'], {'scheduler': 'karras'}),
     ('DPM2 Karras', 'sample_dpm_2', ['k_dpm_2_ka'], {'scheduler': 'karras'}),
     ('DPM2 a Karras', 'sample_dpm_2_ancestral', ['k_dpm_2_a_ka'], {'scheduler': 'karras'}),
+    ('DPM++ 2S a Karras', 'sample_dpmpp_2s_ancestral', ['k_dpmpp_2s_a_ka'], {'scheduler': 'karras'}),
+    ('DPM++ 2M Karras', 'sample_dpmpp_2m', ['k_dpmpp_2m_ka'], {'scheduler': 'karras'}),
 ]
 
 samplers_data_k_diffusion = [
@@ -92,8 +97,8 @@ def single_sample_to_image(sample):
     return Image.fromarray(x_sample)
 
 
-def sample_to_image(samples):
-    return single_sample_to_image(samples[0])
+def sample_to_image(samples, index=0):
+    return single_sample_to_image(samples[index])
 
 
 def samples_to_image_grid(samples):
@@ -280,6 +285,12 @@ class CFGDenoiser(torch.nn.Module):
         image_cond_in = torch.cat([torch.stack([image_cond[i] for _ in range(n)]) for i, n in enumerate(repeats)] + [image_cond])
         sigma_in = torch.cat([torch.stack([sigma[i] for _ in range(n)]) for i, n in enumerate(repeats)] + [sigma])
 
+        denoiser_params = CFGDenoiserParams(x_in, image_cond_in, sigma_in, state.sampling_step, state.sampling_steps)
+        cfg_denoiser_callback(denoiser_params)
+        x_in = denoiser_params.x
+        image_cond_in = denoiser_params.image_cond
+        sigma_in = denoiser_params.sigma
+
         if tensor.shape[1] == uncond.shape[1]:
             cond_in = torch.cat([tensor, uncond])
 

modules/sd_vae.py

+import torch
+import os
+from collections import namedtuple
+from modules import shared, devices, script_callbacks
+from modules.paths import models_path
+import glob
+
+
+model_dir = "Stable-diffusion"
+model_path = os.path.abspath(os.path.join(models_path, model_dir))
+vae_dir = "VAE"
+vae_path = os.path.abspath(os.path.join(models_path, vae_dir))
+
+
+vae_ignore_keys = {"model_ema.decay", "model_ema.num_updates"}
+
+
+default_vae_dict = {"auto": "auto", "None": "None"}
+default_vae_list = ["auto", "None"]
+
+
+default_vae_values = [default_vae_dict[x] for x in default_vae_list]
+vae_dict = dict(default_vae_dict)
+vae_list = list(default_vae_list)
+first_load = True
+
+
+base_vae = None
+loaded_vae_file = None
+checkpoint_info = None
+
+
+def get_base_vae(model):
+    if base_vae is not None and checkpoint_info == model.sd_checkpoint_info and model:
+        return base_vae
+    return None
+
+
+def store_base_vae(model):
+    global base_vae, checkpoint_info
+    if checkpoint_info != model.sd_checkpoint_info:
+        base_vae = model.first_stage_model.state_dict().copy()
+        checkpoint_info = model.sd_checkpoint_info
+
+
+def delete_base_vae():
+    global base_vae, checkpoint_info
+    base_vae = None
+    checkpoint_info = None
+
+
+def restore_base_vae(model):
+    global base_vae, checkpoint_info
+    if base_vae is not None and checkpoint_info == model.sd_checkpoint_info:
+        load_vae_dict(model, base_vae)
+    delete_base_vae()
+
+
+def get_filename(filepath):
+    return os.path.splitext(os.path.basename(filepath))[0]
+
+
+def refresh_vae_list(vae_path=vae_path, model_path=model_path):
+    global vae_dict, vae_list
+    res = {}
+    candidates = [
+        *glob.iglob(os.path.join(model_path, '**/*.vae.ckpt'), recursive=True),
+        *glob.iglob(os.path.join(model_path, '**/*.vae.pt'), recursive=True),
+        *glob.iglob(os.path.join(vae_path, '**/*.ckpt'), recursive=True),
+        *glob.iglob(os.path.join(vae_path, '**/*.pt'), recursive=True)
+    ]
+    if shared.cmd_opts.vae_path is not None and os.path.isfile(shared.cmd_opts.vae_path):
+        candidates.append(shared.cmd_opts.vae_path)
+    for filepath in candidates:
+        name = get_filename(filepath)
+        res[name] = filepath
+    vae_list.clear()
+    vae_list.extend(default_vae_list)
+    vae_list.extend(list(res.keys()))
+    vae_dict.clear()
+    vae_dict.update(res)
+    vae_dict.update(default_vae_dict)
+    return vae_list
+
+
+def resolve_vae(checkpoint_file, vae_file="auto"):
+    global first_load, vae_dict, vae_list
+
+    # if vae_file argument is provided, it takes priority, but not saved
+    if vae_file and vae_file not in default_vae_list:
+        if not os.path.isfile(vae_file):
+            vae_file = "auto"
+            print("VAE provided as function argument doesn't exist")
+    # for the first load, if vae-path is provided, it takes priority, saved, and failure is reported
+    if first_load and shared.cmd_opts.vae_path is not None:
+        if os.path.isfile(shared.cmd_opts.vae_path):
+            vae_file = shared.cmd_opts.vae_path
+            shared.opts.data['sd_vae'] = get_filename(vae_file)
+        else:
+            print("VAE provided as command line argument doesn't exist")
+    # else, we load from settings
+    if vae_file == "auto" and shared.opts.sd_vae is not None:
+        # if saved VAE settings isn't recognized, fallback to auto
+        vae_file = vae_dict.get(shared.opts.sd_vae, "auto")
+        # if VAE selected but not found, fallback to auto
+        if vae_file not in default_vae_values and not os.path.isfile(vae_file):
+            vae_file = "auto"
+            print("Selected VAE doesn't exist")
+    # vae-path cmd arg takes priority for auto
+    if vae_file == "auto" and shared.cmd_opts.vae_path is not None:
+        if os.path.isfile(shared.cmd_opts.vae_path):
+            vae_file = shared.cmd_opts.vae_path
+            print("Using VAE provided as command line argument")
+    # if still not found, try look for ".vae.pt" beside model
+    model_path = os.path.splitext(checkpoint_file)[0]
+    if vae_file == "auto":
+        vae_file_try = model_path + ".vae.pt"
+        if os.path.isfile(vae_file_try):
+            vae_file = vae_file_try
+            print("Using VAE found beside selected model")
+    # if still not found, try look for ".vae.ckpt" beside model
+    if vae_file == "auto":
+        vae_file_try = model_path + ".vae.ckpt"
+        if os.path.isfile(vae_file_try):
+            vae_file = vae_file_try
+            print("Using VAE found beside selected model")
+    # No more fallbacks for auto
+    if vae_file == "auto":
+        vae_file = None
+    # Last check, just because
+    if vae_file and not os.path.exists(vae_file):
+        vae_file = None
+
+    return vae_file
+
+
+def load_vae(model, vae_file=None):
+    global first_load, vae_dict, vae_list, loaded_vae_file
+    # save_settings = False
+
+    if vae_file:
+        print(f"Loading VAE weights from: {vae_file}")
+        vae_ckpt = torch.load(vae_file, map_location=shared.weight_load_location)
+        vae_dict_1 = {k: v for k, v in vae_ckpt["state_dict"].items() if k[0:4] != "loss" and k not in vae_ignore_keys}
+        load_vae_dict(model, vae_dict_1)
+
+        # If vae used is not in dict, update it
+        # It will be removed on refresh though
+        vae_opt = get_filename(vae_file)
+        if vae_opt not in vae_dict:
+            vae_dict[vae_opt] = vae_file
+            vae_list.append(vae_opt)
+
+    loaded_vae_file = vae_file
+
+    """
+    # Save current VAE to VAE settings, maybe? will it work?
+    if save_settings:
+        if vae_file is None:
+            vae_opt = "None"
+
+        # shared.opts.sd_vae = vae_opt
+    """
+
+    first_load = False
+
+
+# don't call this from outside
+def load_vae_dict(model, vae_dict_1=None):
+    if vae_dict_1:
+        store_base_vae(model)
+        model.first_stage_model.load_state_dict(vae_dict_1)
+    else:
+        restore_base_vae()
+    model.first_stage_model.to(devices.dtype_vae)
+
+
+def reload_vae_weights(sd_model=None, vae_file="auto"):
+    from modules import lowvram, devices, sd_hijack
+
+    if not sd_model:
+        sd_model = shared.sd_model
+
+    checkpoint_info = sd_model.sd_checkpoint_info
+    checkpoint_file = checkpoint_info.filename
+    vae_file = resolve_vae(checkpoint_file, vae_file=vae_file)
+
+    if loaded_vae_file == vae_file:
+        return
+
+    if shared.cmd_opts.lowvram or shared.cmd_opts.medvram:
+        lowvram.send_everything_to_cpu()
+    else:
+        sd_model.to(devices.cpu)
+
+    sd_hijack.model_hijack.undo_hijack(sd_model)
+
+    load_vae(sd_model, vae_file)
+
+    sd_hijack.model_hijack.hijack(sd_model)
+    script_callbacks.model_loaded_callback(sd_model)
+
+    if not shared.cmd_opts.lowvram and not shared.cmd_opts.medvram:
+        sd_model.to(devices.device)
+
+    print(f"VAE Weights loaded.")
+    return sd_model

modules/txt2img.py

@@ -47,6 +47,8 @@ def txt2img(prompt: str, negative_prompt: str, prompt_style: str, prompt_style2:
     if processed is None:
         processed = process_images(p)
 
+    p.close()
+
     shared.total_tqdm.clear()
 
     generation_info_js = processed.js()

requirements.txt

@@ -4,7 +4,7 @@ fairscale==0.4.4
 fonts
 font-roboto
 gfpgan
-gradio==3.5
+gradio==3.8
 invisible-watermark
 numpy
 omegaconf

requirements_versions.txt

@@ -2,7 +2,7 @@ transformers==4.19.2
 diffusers==0.3.0
 basicsr==1.4.2
 gfpgan==1.3.8
-gradio==3.5
+gradio==3.8
 numpy==1.23.3
 Pillow==9.2.0
 realesrgan==0.3.0

scripts/custom_code.py

@@ -14,7 +14,7 @@ class Script(scripts.Script):
         return cmd_opts.allow_code
 
     def ui(self, is_img2img):
-        code = gr.Textbox(label="Python code", visible=False, lines=1)
+        code = gr.Textbox(label="Python code", lines=1)
 
         return [code]