from diffusers import StableDiffusionControlNetPipeline, ControlNetModel, StableDiffusionPipeline import torch import cv2 import numpy as np from transformers import pipeline import gradio as gr from PIL import Image from diffusers.utils import load_image import os, random, gc, re, json, time, shutil, glob import PIL.Image import tqdm from controlnet_aux import OpenposeDetector from accelerate import Accelerator from huggingface_hub import HfApi, list_models, InferenceClient, ModelCard, RepoCard, upload_folder, hf_hub_download, HfFileSystem HfApi=HfApi() HF_TOKEN=os.getenv("HF_TOKEN") HF_HUB_DISABLE_TELEMETRY=1 DO_NOT_TRACK=1 HF_HUB_ENABLE_HF_TRANSFER=0 accelerator = Accelerator(cpu=True) InferenceClient=InferenceClient() models =[ "runwayml/stable-diffusion-v1-5", "prompthero/openjourney-v4", "CompVis/stable-diffusion-v1-4", "stabilityai/stable-diffusion-2-1", "stablediffusionapi/edge-of-realism", "MirageML/fantasy-scene", "wavymulder/lomo-diffusion", "sd-dreambooth-library/fashion", "DucHaiten/DucHaitenDreamWorld", "VegaKH/Ultraskin", "kandinsky-community/kandinsky-2-1", "MirageML/lowpoly-cyberpunk", "thehive/everyjourney-sdxl-0.9-finetuned", "plasmo/woolitize-768sd1-5", "plasmo/food-crit", "johnslegers/epic-diffusion-v1.1", "Fictiverse/ElRisitas", "robotjung/SemiRealMix", "herpritts/FFXIV-Style", "prompthero/linkedin-diffusion", "RayHell/popupBook-diffusion", "MirageML/lowpoly-world", "deadman44/SD_Photoreal_Merged_Models", "johnslegers/epic-diffusion", "tilake/China-Chic-illustration", "wavymulder/modelshoot", "prompthero/openjourney-lora", "Fictiverse/Stable_Diffusion_VoxelArt_Model", "darkstorm2150/Protogen_v2.2_Official_Release", "hassanblend/HassanBlend1.5.1.2", "hassanblend/hassanblend1.4", "nitrosocke/redshift-diffusion", "prompthero/openjourney-v2", "nitrosocke/Arcane-Diffusion", "Lykon/DreamShaper", "wavymulder/Analog-Diffusion", "nitrosocke/mo-di-diffusion", "dreamlike-art/dreamlike-diffusion-1.0", "dreamlike-art/dreamlike-photoreal-2.0", "digiplay/RealismEngine_v1", "digiplay/AIGEN_v1.4_diffusers", "stablediffusionapi/dreamshaper-v6", "p1atdev/liminal-space-diffusion", "nadanainone/gigaschizonegs", "lckidwell/album-cover-style", "axolotron/ice-cream-animals", "perion/ai-avatar", "digiplay/GhostMix", "ThePioneer/MISA", "TheLastBen/froggy-style-v21-768", "FloydianSound/Nixeu_Diffusion_v1-5", "kakaobrain/karlo-v1-alpha-image-variations", "digiplay/PotoPhotoRealism_v1", "ConsistentFactor/Aurora-By_Consistent_Factor", "rim0/quadruped_mechas", "Akumetsu971/SD_Samurai_Anime_Model", "Bojaxxx/Fantastic-Mr-Fox-Diffusion", "sd-dreambooth-library/original-character-cyclps", ] loris=[] apol=[] def smdls(models): models=models mtlst=HfApi.list_models(filter="diffusers:StableDiffusionPipeline",limit=500,full=True,) if mtlst: for nea in mtlst: vmh=""+str(nea.id)+"" models.append(vmh) return models def sldls(loris): loris=loris ltlst=HfApi.list_models(filter="stable-diffusion",search="lora",limit=500,full=True,) if ltlst: for noa in ltlst: lmh=""+str(noa.id)+"" loris.append(lmh) return loris def chdr(apol,prompt,modil,los,stips,fnamo,gaul): try: type="SD_controlnet" tre='./tmpo/'+fnamo+'.json' tra='./tmpo/'+fnamo+'_0.png' trm='./tmpo/'+fnamo+'_1.png' trv='./tmpo/'+fnamo+'_pose.png' trh='./tmpo/'+fnamo+'_canny.png' trg='./tmpo/'+fnamo+'_cann_im.png' trq='./tmpo/'+fnamo+'_tilage.png' flng=["yssup", "sllab", "stsaerb", "sinep", "selppin", "ssa", "tnuc", "mub", "kcoc", "kcid", "anigav", "dekan", "edun", "slatineg", "xes", "nrop", "stit", "ttub", "bojwolb", "noitartenep", "kcuf", "kcus", "kcil", "elttil", "gnuoy", "thgit", "lrig", "etitep", "dlihc", "yxes"] flng=[itm[::-1] for itm in flng] ptn = r"\b" + r"\b|\b".join(flng) + r"\b" if re.search(ptn, prompt, re.IGNORECASE): print("onon buddy") else: dobj={'img_name':fnamo,'model':modil,'lora':los,'prompt':prompt,'steps':stips,'type':type} with open(tre, 'w') as f: json.dump(dobj, f) HfApi.upload_folder(repo_id="JoPmt/hf_community_images",folder_path="./tmpo",repo_type="dataset",path_in_repo="./",token=HF_TOKEN) dobj={'img_name':fnamo,'model':modil,'lora':los,'prompt':prompt,'steps':stips,'type':type,'haed':gaul,} try: for pxn in glob.glob('./tmpo/*.png'): os.remove(pxn) except: print("lou") with open(tre, 'w') as f: json.dump(dobj, f) HfApi.upload_folder(repo_id="JoPmt/Tst_datast_imgs",folder_path="./tmpo",repo_type="dataset",path_in_repo="./",token=HF_TOKEN) try: for pgn in glob.glob('./tmpo/*.png'): os.remove(pgn) for jgn in glob.glob('./tmpo/*.json'): os.remove(jgn) del tre del tra del trm del trv del trh del trg del trq except: print("cant") except: print("failed to umake obj") def crll(dnk): lix="" lotr=HfApi.list_files_info(repo_id=""+dnk+"",repo_type="model") for flre in list(lotr): fllr=[] gar=re.match(r'.+(\.pt|\.ckpt|\.bin|\.safetensors)$', flre.path) yir=re.search(r'[^/]+$', flre.path) if gar: fllr.append(""+str(yir.group(0))+"") lix=""+fllr[-1]+"" else: lix="" return lix def plax(gaul,req: gr.Request): gaul=str(req.headers) return gaul def plex(prompt,mput,neg_prompt,modil,stips,scaly,csal,csbl,nut,wei,hei,los,loca,gaul,progress=gr.Progress(track_tqdm=True)): gc.collect() adi="" ldi="" openpose = OpenposeDetector.from_pretrained("lllyasviel/ControlNet") controlnet = [ ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-openpose", torch_dtype=torch.float32), ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-canny", torch_dtype=torch.float32), ] try: crda=ModelCard.load(""+modil+"") card=ModelCard.load(""+modil+"").data.to_dict().get("instance_prompt") cerd=ModelCard.load(""+modil+"").data.to_dict().get("custom_prompt") cird=ModelCard.load(""+modil+"").data.to_dict().get("lora_prompt") mtch=re.search(r'(?:(?<=trigger words:)|(?<=trigger:)|(?<=You could use)|(?<=You should use))\s*(.*?)\s*(?=to trigger)', crda.text, re.IGNORECASE) moch=re.search(r'(?:(?<=trigger words:)|(?<=trigger:)|(?<=You could use)|(?<=You should use))\s*([^.]*)', crda.text, re.IGNORECASE) if moch: adi+=""+str(moch.group(1))+", " else: print("no floff trigger") if mtch: adi+=""+str(mtch.group(1))+", " else: print("no fluff trigger") if card: adi+=""+str(card)+", " else: print("no instance") if cerd: adi+=""+str(cerd)+", " else: print("no custom") if cird: adi+=""+str(cird)+", " else: print("no lora") except: print("no card") try: pope = accelerator.prepare(StableDiffusionPipeline.from_pretrained(""+modil+"", use_safetensors=False,torch_dtype=torch.float32, safety_checker=None)) pipe = accelerator.prepare(StableDiffusionControlNetPipeline.from_pretrained(""+modil+"", use_safetensors=False,controlnet=controlnet,torch_dtype=torch.float32,safety_checker=None)) except: gc.collect() pope = accelerator.prepare(StableDiffusionPipeline.from_pretrained(""+modil+"", use_safetensors=True,torch_dtype=torch.float32, safety_checker=None)) pipe = accelerator.prepare(StableDiffusionControlNetPipeline.from_pretrained(""+modil+"", use_safetensors=True,controlnet=controlnet,torch_dtype=torch.float32,safety_checker=None)) if los: try: lrda=ModelCard.load(""+los+"") lard=ModelCard.load(""+los+"").data.to_dict().get("instance_prompt") lerd=ModelCard.load(""+los+"").data.to_dict().get("custom_prompt") lird=ModelCard.load(""+los+"").data.to_dict().get("stable-diffusion") ltch=re.search(r'(?:(?<=trigger words:)|(?<=trigger:)|(?<=You could use)|(?<=You should use))\s*(.*?)\s*(?=to trigger)', lrda.text, re.IGNORECASE) loch=re.search(r'(?:(?<=trigger words:)|(?<=trigger:)|(?<=You could use)|(?<=You should use))\s*([^.]*)', lrda.text, re.IGNORECASE) if loch and lird: ldi+=""+str(loch.group(1))+", " else: print("no lloff trigger") if ltch and lird: ldi+=""+str(ltch.group(1))+", " else: print("no lluff trigger") if lard and lird: ldi+=""+str(lard)+", " else: print("no instance") ldi+="" if lerd and lird: ldi+=""+str(lerd)+", " else: print("no custom") ldi+="" except: print("no trigger") try: pope.load_lora_weights(""+los+"", weight_name=""+str(crll(los))+"",) pope.fuse_lora(fuse_unet=True,fuse_text_encoder=False) except: print("no can do") else: los="" pope.unet.to(memory_format=torch.channels_last) pope = accelerator.prepare(pope.to("cpu")) pipe.unet.to(memory_format=torch.channels_last) pipe = accelerator.prepare(pipe.to("cpu")) gc.collect() apol=[] height=hei width=wei prompt=""+str(adi)+""+str(ldi)+""+prompt+"" negative_prompt=""+neg_prompt+"" lora_scale=loca if nut == 0: nm = random.randint(1, 2147483616) while nm % 32 != 0: nm = random.randint(1, 2147483616) else: nm=nut generator = torch.Generator(device="cpu").manual_seed(nm) tilage = pope(prompt,num_inference_steps=5,height=height,width=width,generator=generator,cross_attention_kwargs={"scale": lora_scale}).images[0] cannyimage = np.array(tilage) low_threshold = 100 high_threshold = 200 fnamo=""+str(int(time.time()))+"" cannyimage = cv2.Canny(cannyimage, low_threshold, high_threshold) cammyimage=Image.fromarray(cannyimage).save('./tmpo/'+fnamo+'_canny.png', 'PNG') zero_start = cannyimage.shape[1] // 4 zero_end = zero_start + cannyimage.shape[1] // 2 cannyimage[:, zero_start:zero_end] = 0 cannyimage = cannyimage[:, :, None] cannyimage = np.concatenate([cannyimage, cannyimage, cannyimage], axis=2) canny_image = Image.fromarray(cannyimage) pose_image = load_image(mput).resize((512, 512)) openpose_image = openpose(pose_image) images = [openpose_image, canny_image] omage=pipe([prompt]*2,images,num_inference_steps=stips,generator=generator,negative_prompt=[neg_prompt]*2,controlnet_conditioning_scale=[csal, csbl]) for i, imge in enumerate(omage["images"]): apol.append(imge) imge.save('./tmpo/'+fnamo+'_'+str(i)+'.png', 'PNG') apol.append(openpose_image) apol.append(cammyimage) apol.append(canny_image) apol.append(tilage) openpose_image.save('./tmpo/'+fnamo+'_pose.png', 'PNG') canny_image.save('./tmpo/'+fnamo+'_cann_im.png', 'PNG') tilage.save('./tmpo/'+fnamo+'_tilage.png', 'PNG') chdr(apol,prompt,modil,los,stips,fnamo,gaul) return apol def aip(ill,api_name="/run"): return def pit(ill,api_name="/predict"): return with gr.Blocks(theme=random.choice([gr.themes.Monochrome(),gr.themes.Base.from_hub("gradio/seafoam"),gr.themes.Base.from_hub("freddyaboulton/dracula_revamped"),gr.themes.Glass(),gr.themes.Base(),]),analytics_enabled=False) as iface: ##iface.description="Running on cpu, very slow! by JoPmt." out=gr.Gallery(label="Generated Output Image", columns=1) inut=gr.Textbox(label="Prompt") mput=gr.Image(type="filepath") gaul=gr.Textbox(visible=False) inot=gr.Dropdown(choices=smdls(models),value=random.choice(models), type="value") btn=gr.Button("GENERATE") with gr.Accordion("Advanced Settings", open=False): inlt=gr.Dropdown(choices=sldls(loris),value=None, type="value") inet=gr.Textbox(label="Negative_prompt", value="low quality, bad quality,") inyt=gr.Slider(label="Num inference steps",minimum=1,step=1,maximum=30,value=20) inat=gr.Slider(label="Guidance_scale",minimum=1,step=1,maximum=20,value=7) csal=gr.Slider(label="condition_scale_canny", value=0.5, minimum=0.1, step=0.1, maximum=1) csbl=gr.Slider(label="condition_scale_pose", value=0.5, minimum=0.1, step=0.1, maximum=1) loca=gr.Slider(label="Lora scale",minimum=0.1,step=0.1,maximum=0.9,value=0.5) indt=gr.Slider(label="Manual seed (leave 0 for random)",minimum=0,step=32,maximum=2147483616,value=0) inwt=gr.Slider(label="Width",minimum=512,step=32,maximum=1024,value=512) inht=gr.Slider(label="Height",minimum=512,step=32,maximum=1024,value=512) btn.click(fn=plax,inputs=gaul,outputs=gaul).then(fn=plex, outputs=[out], inputs=[inut,mput,inet,inot,inyt,inat,csal,csbl,indt,inwt,inht,inlt,loca,gaul]) iface.queue(max_size=1,api_open=False) iface.launch(max_threads=20,inline=False,show_api=False)