hydra

{text_encoder → anime/text_encoder}/pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:38a67003cd791d4fc008ae1fd24615b8b168f83cc8e853b746a7ec7bb3d64f42
+oid sha256:ec3883f2cc6b4a8661212175bb7ccefe9f2309888d3e53c9f274e7de60acc88a
 size 492306077

{unet → anime/unet}/diffusion_pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b64c917a491e553494fa0eb452824cc068296273a09436ca0803b06c42046c7d
+oid sha256:d8e4501ae1ba2c9950a2c06cbdd74a3317a4b4efbce54557163d114945760a0a
 size 3438366373

{vae → anime/vae}/diffusion_pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a425a89f2e522790b3975b93ed380814e68ec77a04841dced0832cad70eab929
+oid sha256:40bec3b829f93e173729fd6dca141f4488803311de360d60de1c7d3714953c0f
 size 334712113

embeddings/kkw-NativeAmerican.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ed4b15ad30845744274acde1341783cbfc115f520a41fe3d8457537dfef15a75
+size 108492

handler.py

@@ -12,7 +12,7 @@ from typing import Union
 from concurrent.futures import ThreadPoolExecutor
 import numpy as np
 from PIL import ImageFilter
-
+from transformers import CLIPImageProcessor, CLIPTokenizer, CLIPModel
 
 import torch
 from diffusers import (
@@ -20,6 +20,7 @@ from diffusers import (
     DPMSolverMultistepScheduler,
     DPMSolverSinglestepScheduler,
     EulerAncestralDiscreteScheduler,
+    StableDiffusionPipeline,
     utils,
 )
 from safetensors.torch import load_file
@@ -41,20 +42,66 @@ if device.type != "cuda":
 
 class EndpointHandler:
     LORA_PATHS = {
-        "hairdetailer": str(REPO_DIR / "lora/hairdetailer.safetensors"),
-        "lora_leica": str(REPO_DIR / "lora/lora_leica.safetensors"),
-        "epiNoiseoffset_v2": str(REPO_DIR / "lora/epiNoiseoffset_v2.safetensors"),
-        "MBHU-TT2FRS": str(REPO_DIR / "lora/MBHU-TT2FRS.safetensors"),
-        "ShinyOiledSkin_v20": str(
-            REPO_DIR / "lora/ShinyOiledSkin_v20-LoRA.safetensors"
-        ),
-        "polyhedron_new_skin_v1.1": str(
-            REPO_DIR / "lora/polyhedron_new_skin_v1.1.safetensors"
-        ),
-        "detailed_eye-10": str(REPO_DIR / "lora/detailed_eye-10.safetensors"),
-        "add_detail": str(REPO_DIR / "lora/add_detail.safetensors"),
-        "MuscleGirl_v1": str(REPO_DIR / "lora/MuscleGirl_v1.safetensors"),
-        "flat2": str(REPO_DIR / "lora/flat2.safetensors"),
+        "hairdetailer": [str(REPO_DIR / "lora/hairdetailer.safetensors"), ""],
+        "lora_leica": [str(REPO_DIR / "lora/lora_leica.safetensors"), "leica_style"],
+        "epiNoiseoffset_v2": [str(REPO_DIR / "lora/epiNoiseoffset_v2.safetensors"), ""],
+        "MBHU-TT2FRS": [
+            str(REPO_DIR / "lora/MBHU-TT2FRS.safetensors"),
+            "flat breast, small breast, big breast, fake breast",
+        ],
+        "polyhedron_new_skin_v1.1": [
+            str(REPO_DIR / "lora/polyhedron_new_skin_v1.1.safetensors"),
+            "skin blemish,  detailed skin ",
+        ],
+        "ShinyOiledSkin_v20": [
+            str(REPO_DIR / "lora/ShinyOiledSkin_v20-LoRA.safetensors"),
+            "shiny skin",
+        ],
+        "detailed_eye-10": [str(REPO_DIR / "lora/detailed_eye-10.safetensors"), ""],
+        "add_detail": [str(REPO_DIR / "lora/add_detail.safetensors"), ""],
+        "MuscleGirl_v1": [str(REPO_DIR / "lora/MuscleGirl_v1.safetensors"), "abs"],
+        "nurse_v11-05": [str(REPO_DIR / "lora/nurse_v11-05.safetensors"), "nurse"],
+        "shibari_v20": [str(REPO_DIR / "lora/shibari_v20.safetensors"), "shibari,rope"],
+        "tajnaclub_high_heelsv1.2": [
+            str(REPO_DIR / "lora/tajnaclub_high_heelsv1.2.safetensors"),
+            "high heels",
+        ],
+        "CyberPunkAI": [
+            str(REPO_DIR / "lora/CyberPunkAI.safetensors"),
+            "neon CyberpunkAI",
+        ],
+        "FutaCockCloseUp-v1": [
+            str(REPO_DIR / "lora/FutaCockCloseUp-v1.safetensors"),
+            "huge penis",
+        ],
+        "PovBlowjob-v3": [
+            str(REPO_DIR / "lora/PovBlowjob-v3.safetensors"),
+            "blowjob, deepthroat, kneeling, runny makeup, creampie",
+        ],
+        "dp_from_behind_v0.1b": [
+            str(REPO_DIR / "lora/dp_from_behind_v0.1b.safetensors"),
+            "1girl, 2boys, double penetration, multiple penises",
+        ],
+        "EkuneSideDoggy": [
+            str(REPO_DIR / "lora/EkuneSideDoggy.safetensors"),
+            "sidedoggystyle, doggystyle",
+        ],
+        "qqq-grabbing_from_behind-v2-000006": [
+            str(REPO_DIR / "lora/qqq-grabbing_from_behind-v2-000006.safetensors"),
+            "grabbing from behind, breast grab",
+        ],
+        "ftm-v0": [
+            str(REPO_DIR / "lora/ftm-v0.safetensors"),
+            "big mouth, tongue, long tongue",
+        ],
+        "tgirls_V3_5": [
+            str(REPO_DIR / "lora/tgirls_V3_5.safetensors"),
+            "large penis, penis, erect penis",
+        ],
+        "fapp9": [
+            str(REPO_DIR / "lora/fapp9.safetensors"),
+            "large penis, penis, erect penis",
+        ],
     }
 
     TEXTUAL_INVERSION = [
@@ -63,9 +110,13 @@ class EndpointHandler:
             "token": "easynegative",
         },
         {
-            "weight_name": str(REPO_DIR / "embeddings/badhandv4.pt"),
+            "weight_name": str(REPO_DIR / "embeddings/kkw-NativeAmerican.pt"),
             "token": "badhandv4",
         },
+        {
+            "weight_name": str(REPO_DIR / "embeddings/badhandv4.pt"),
+            "token": "kkw-Afro, kkw-Asian, kkw-Euro ",
+        },
         {
             "weight_name": str(REPO_DIR / "embeddings/bad-artist-anime.pt"),
             "token": "bad-artist-anime",
@@ -95,56 +146,119 @@ class EndpointHandler:
             max_workers=1
         )  # Vous pouvez ajuster max_workers en fonction de vos besoins
 
+        realistic_path = str(REPO_DIR / "realistic/")
+        self.pipe_realistic, self.safety_checker = self.load_realistic(realistic_path)
+
+        anime_path = str(REPO_DIR / "anime/")
+        self.pipe_anime, self.pipe_anime_safety_checker = self.load_anime(anime_path)
+
+        # Load CLipImagePRocessor for NSFW check later
+        self.image_processor = CLIPImageProcessor.from_pretrained(
+            "openai/clip-vit-base-patch16"
+        )
+
+    def load_model_essentials(self, model_path):
+        """common to all models"""
+
         # load the optimized model
-        self.pipe = DiffusionPipeline.from_pretrained(
-            path,
+
+        if "realistic" in model_path:
+            pipe = DiffusionPipeline.from_pretrained(
+                pretrained_model_name_or_path=model_path,
+                custom_pipeline="lpw_stable_diffusion",  # avoid 77 token limit
+                torch_dtype=torch.float16,  # accelerate render
+            )
+
+            safety_checker = pipe.safety_checker.to(device).to(torch.float16)
+        else:
+            safety_checker = None
+
+        pipe = DiffusionPipeline.from_pretrained(
+            pretrained_model_name_or_path=model_path,
             custom_pipeline="lpw_stable_diffusion",  # avoid 77 token limit
             torch_dtype=torch.float16,  # accelerate render
+            safety_checker=None,  # Mode boulardus
         )
-        self.pipe = self.pipe.to(device)
+
+        pipe = pipe.to(device)
+
+        # Disable progress bar
+        pipe.set_progress_bar_config(disable=True)
+
+        # Load negative embeddings to avoid bad hands, etc
+        self.load_embeddings(pipe)
+
+        # boosts performance by another 20%
+        pipe.enable_xformers_memory_efficient_attention()
+        pipe.enable_attention_slicing()  # may need a requirement in the root with xformer
+
+        return pipe, safety_checker
+
+    def load_anime(self, path):
+        """Load anime model"""
+
+        # Init pipe
+        pipe, safety_checker = self.load_model_essentials(path)
 
         # https://stablediffusionapi.com/docs/a1111schedulers/
 
-        # DPM++ 2M SDE Karras
-        # increase step to avoid high contrast num_inference_steps=30
-        # self.pipe.scheduler = DPMSolverMultistepScheduler.from_config(
-        #     self.pipe.scheduler.config,
-        #     use_karras_sigmas=True,
-        #     algorithm_type="sde-dpmsolver++",
-        # )
+        # Euler a
+        pipe.scheduler = EulerAncestralDiscreteScheduler.from_config(
+            pipe.scheduler.config,
+        )
 
-        # DPM++ 2M Karras
-        self.pipe.scheduler = DPMSolverMultistepScheduler.from_config(
-            self.pipe.scheduler.config,
-            use_karras_sigmas=True,
+        # Load loras one time only
+        # Must be replaced once we will know how to hot load/unload
+        # it use the own made load_lora function
+        self.load_selected_loras(
+            pipe,
+            [
+                ["detailed_eye-10", 0.3],
+                ["add_detail", 0.3],
+                ["MuscleGirl_v1", 0.3],
+                ["dp_from_behind_v0.1b", 0.2],
+                ["shibari_v20", 0.03],
+                ["ftm-v0", 0.03],
+                ["PovBlowjob-v3", 0.03],
+            ],
         )
 
-        # Mode boulardus
-        self.pipe.safety_checker = None
+        return pipe, safety_checker
 
-        # Disable progress bar
-        self.pipe.set_progress_bar_config(disable=True)
+    def load_realistic(self, path):
+        """Load realistic model"""
 
-        # Load negative embeddings to avoid bad hands, etc
-        self.load_embeddings()
+        # Init pipe
+        pipe, safety_checker = self.load_model_essentials(path)
 
-        # boosts performance by another 20%
-        self.pipe.enable_xformers_memory_efficient_attention()
-        self.pipe.enable_attention_slicing()
-        # may need a requirement in the root with xformer
+        # https://stablediffusionapi.com/docs/a1111schedulers/
+
+        # DPM++ 2M Karras
+        pipe.scheduler = DPMSolverMultistepScheduler.from_config(
+            pipe.scheduler.config,
+            use_karras_sigmas=True,
+        )
 
         # Load loras one time only
         # Must be replaced once we will know how to hot load/unload
         # it use the own made load_lora function
         self.load_selected_loras(
+            pipe,
             [
-                ["polyhedron_new_skin_v1.1", 0.2],
+                ["polyhedron_new_skin_v1.1", 0.25],
                 ["detailed_eye-10", 0.2],
-                ["add_detail", 0.3],
+                ["add_detail", 0.2],
                 ["MuscleGirl_v1", 0.2],
-            ]
+                ["tgirls_V3_5", 0.02],
+                ["PovBlowjob-v3", 0.02],
+                # ["dp_from_behind_v0.1b", 0.2],
+                ["shibari_v20", 0.02],
+                ["ftm-v0", 0.02],
+            ],
         )
 
+        return pipe, safety_checker
+
     def load_lora(self, pipeline, lora_path, lora_weight=0.5):
         state_dict = load_file(lora_path)
         LORA_PREFIX_UNET = "lora_unet"
@@ -221,19 +335,19 @@ class EndpointHandler:
 
         return pipeline
 
-    def load_embeddings(self):
+    def load_embeddings(self, pipeline):
         """Load textual inversions, avoid bad prompts"""
         for model in EndpointHandler.TEXTUAL_INVERSION:
-            self.pipe.load_textual_inversion(
+            pipeline.load_textual_inversion(
                 ".", weight_name=model["weight_name"], token=model["token"]
             )
 
-    def load_selected_loras(self, selections):
+    def load_selected_loras(self, pipeline, selections):
         """Load Loras models, can lead to marvelous creations"""
         for model_name, weight in selections:
-            lora_path = EndpointHandler.LORA_PATHS[model_name]
+            lora_path = EndpointHandler.LORA_PATHS[model_name][0]
             # self.pipe.load_lora_weights(lora_path)
-            self.load_lora(self.pipe, lora_path, weight)
+            self.load_lora(pipeline, lora_path, weight)
 
     def clean_negative_prompt(self, negative_prompt):
         """Clean negative prompt to remove already used negative prompt handlers"""
@@ -260,6 +374,30 @@ class EndpointHandler:
 
         return negative_prompt
 
+    def check_fields(self, data):
+        """check for fields, if some missing return error"""
+
+        # 1. Verify input arguments
+        required_fields = [
+            "prompt",
+            "negative_prompt",
+            "width",
+            "num_inference_steps",
+            "height",
+            "guidance_scale",
+            "request_id",
+        ]
+
+        missing_fields = [field for field in required_fields if field not in data]
+
+        if missing_fields:
+            return {
+                "flag": "error",
+                "message": f"Missing fields: {', '.join(missing_fields)}",
+            }
+
+        return False
+
     def clean_request_data(self):
         """Clean up the data related to a specific request ID."""
 
@@ -285,11 +423,12 @@ class EndpointHandler:
         latents: Any,
         request_id: str,
         status: str,
+        pipeline: Any,
     ):
         try:
             if status == "progress":
                 # Latents to numpy
-                img_data = self.pipe.decode_latents(latents)
+                img_data = pipeline.decode_latents(latents)
                 img_data = (img_data.squeeze() * 255).astype(np.uint8)
                 img = Image.fromarray(img_data, "RGB")
 
@@ -304,6 +443,7 @@ class EndpointHandler:
             else:
                 # pil object
                 # print(latents)
+
                 img = latents
 
             buffered = BytesIO()
@@ -335,48 +475,42 @@ class EndpointHandler:
 
         if progress >= 100:
             status = "complete"
+
+            # Check if Image is NSFW
+            image_data = base64.b64decode(latest_image)
+            image_io = BytesIO(image_data)
+            is_nsfw = self.check_nsfw(Image.open(image_io))[0]
+            # is_nsfw = "bypass"
         else:
             status = "in-progress"
+            is_nsfw = ""
 
         return {
             "flag": "success",
             "status": status,
             "progress": int(progress),
             "image": latest_image,
+            "is_nsfw": is_nsfw,
         }
 
-    def start_inference(self, data: Dict) -> Dict:
-        """Start a new inference."""
+    def check_nsfw(self, image):
+        """Check if image is NSFW"""
 
-        global device
+        safety_checker_input = self.image_processor(image, return_tensors="pt").to(
+            device
+        )
 
-        # Which Lora do we load ?
-        # selected_models = [
-        #     ("ShinyOiledSkin_v20", 0.3),
-        #     ("MBHU-TT2FRS", 0.5),
-        #     ("hairdetailer", 0.5),
-        #     ("lora_leica", 0.5),
-        #     ("epiNoiseoffset_v2", 0.5),
-        # ]
+        image, has_nsfw_concept = self.safety_checker(
+            images=np.array(image),
+            clip_input=safety_checker_input.pixel_values.to(torch.float16),
+        )
 
-        # 1. Verify input arguments
-        required_fields = [
-            "prompt",
-            "negative_prompt",
-            "width",
-            "num_inference_steps",
-            "height",
-            "guidance_scale",
-            "request_id",
-        ]
+        return has_nsfw_concept
 
-        missing_fields = [field for field in required_fields if field not in data]
+    def start_inference(self, pipeline, data: Dict) -> Dict:
+        """Start a new inference."""
 
-        if missing_fields:
-            return {
-                "flag": "error",
-                "message": f"Missing fields: {', '.join(missing_fields)}",
-            }
+        global device
 
         # Now extract the fields
         prompt = data["prompt"]
@@ -398,18 +532,14 @@ class EndpointHandler:
         # Set the generator seed if provided
         generator = torch.Generator(device="cuda").manual_seed(seed) if seed else None
 
-        # Load the provided Lora models
-        # self.pipe.unload_lora_weights()  # Unload models to avoid lora staking
-        # if loras_model:
-        #     self.load_selected_loras(loras_model)
-
-        # set scale of loras, for now take only first scale of the loaded lora and apply to all until we find the way to apply specified scale
-        # scale = {"scale": loras_model[0][1]} if loras_model else None
+        # set scale of loras (mix all loras and apply common scale, can't be indivual)
+        # scale = 0.25 # seems ok
+        # scale = 0.2
 
         try:
             # 2. Process
             with autocast(device.type):
-                image = self.pipe.text2img(
+                image = pipeline.text2img(
                     prompt=prompt,
                     guidance_scale=guidance_scale,
                     num_inference_steps=num_inference_steps,
@@ -419,15 +549,20 @@ class EndpointHandler:
                     generator=generator,
                     max_embeddings_multiples=5,
                     callback=lambda step, timestep, latents: self.progress_callback(
-                        step, timestep, latents, request_id, "progress"
+                        step, timestep, latents, request_id, "progress", pipeline
                     ),
                     callback_steps=5,
-                    # cross_attention_kwargs={"scale": 0.2},
-                ).images[0]
+                    # cross_attention_kwargs={"scale": 0.02},
+                )
 
             # print(image)
             self.progress_callback(
-                num_inference_steps, 0, image, request_id, "complete"
+                num_inference_steps,
+                0,
+                image.images[0],
+                request_id,
+                "complete",
+                pipeline,
             )
 
             self.inference_in_progress = False
@@ -444,6 +579,7 @@ class EndpointHandler:
 
         action = data.get("action", None)
         request_id = data.get("request_id")
+        genre = data.get("genre")
 
         # Check if the request_id is valid for all actions
         if not request_id:
@@ -458,6 +594,11 @@ class EndpointHandler:
             return self.check_progress(request_id)
 
         elif action == "inference":
+            # Check field before doing anything
+            check_fields = self.check_fields(data)
+            if check_fields:
+                return check_fields
+
             # Check if an inference is already in progress
             if self.inference_in_progress:
                 return {
@@ -472,7 +613,14 @@ class EndpointHandler:
             self.inference_images[request_id] = None
             self.active_request_ids.append(request_id)
 
-            self.executor.submit(self.start_inference, data)
+            # Load model according to genre
+            if genre == "anime":
+                pipe = self.pipe_anime
+            else:
+                pipe = self.pipe_realistic
+
+            self.executor.submit(self.start_inference, pipe, data)
+            # self.start_inference(data)
 
             return {
                 "flag": "success",

lora/ASSv12-000011.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:70a5e09e70701195b0f8e359558cc1a17d26b6bd0c864010964875d7c5e68862
+size 51912025

lora/BarbieCore.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b6d1118c1bc942b248f07492e1ee149d118aae7916c7af9cdad1b58b2ce0c092
+size 151112335

lora/CyberPunkAI.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5a05b7123370fc40abdeaf6913cfec64e3d6a0633d52043e5a375d7a0ddbf7e1
+size 151112443

lora/EkuneSideDoggy.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e6c00be1d16bf804aaa406cc28385e18fb85aa006dd6b84848ffb06d9f04254a
+size 37875655

lora/FutaCockCloseUp-v1.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:926d74e0d7a7bde7ac8cb9b4487222c30d21197675b3791c690ea057be0ea21d
+size 151111728

lora/PovBlowjob-v3.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5ca2bf4d3f7618fea43f1324d67ef220b07b6cc93d824067ad34ecf439cf9d5c
+size 151118749

lora/dp_from_behind_v0.1b.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7daf59a9e7c2fbb0a728fa4cd4dfc019a8009b07eca395e08eec22b9b62d1bfd
+size 9567809

lora/ftm-v0.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:21be0e6c7673eefcb01d14fba3dcecc02104ce01e41c47f83c7f2156486e10d2
+size 151444477

lora/nurse_v11-05.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f98e01d6528f32ef47439c8cf696fb1bce7cd86f78136bfbb4b959690d7fc00e
+size 19045712

lora/qqq-grabbing_from_behind-v2-000006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8a7be0c8e5e418699a251c6dfefdd4a5b52de8d2dd83eb7e18cad820f4d15a98
+size 37877190

lora/rindou_v4.2.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:27fa90277cc0e007de331d57b7414f0b7a4a42a87e2d166e518521c7fd8796e4
+size 9558043

lora/screaming.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:884e863e81b53539f4dac8d89b3154fb2b0996d7a72511cac580cc7596bd4a94
+size 18992918

lora/shibari_v20.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1764d4d9eeebab51ecf295330f983ebd206b20688feb8fd7c2a83f784ff0406d
+size 37864830

lora/tajnaclub_high_heelsv1.2.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9476e2fa9759e0bdb733825692ea2a60dd51075dfff38d7ed235d81f1ce3537e
+size 18999240

lora/tgirls_V3_5.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3ba32950fb1457aaf3876c3c7c2c9af4fd4ea857487d09c9eb9ed4bec9e716ef
+size 38001393

lycoris/fapp9.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:70bfdbb77a5c1d390efb1c62cb9409541994667acdfdb942b5e0bc6d1ddda537
+size 130206396

realistic/feature_extractor/preprocessor_config.json

@@ -0,0 +1,28 @@
+{
+  "crop_size": {
+    "height": 224,
+    "width": 224
+  },
+  "do_center_crop": true,
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "feature_extractor_type": "CLIPFeatureExtractor",
+  "image_mean": [
+    0.48145466,
+    0.4578275,
+    0.40821073
+  ],
+  "image_processor_type": "CLIPFeatureExtractor",
+  "image_std": [
+    0.26862954,
+    0.26130258,
+    0.27577711
+  ],
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "shortest_edge": 224
+  }
+}

realistic/model_index.json

@@ -0,0 +1,33 @@
+{
+  "_class_name": "StableDiffusionPipeline",
+  "_diffusers_version": "0.20.0",
+  "feature_extractor": [
+    "transformers",
+    "CLIPFeatureExtractor"
+  ],
+  "requires_safety_checker": true,
+  "safety_checker": [
+    "stable_diffusion",
+    "StableDiffusionSafetyChecker"
+  ],
+  "scheduler": [
+    "diffusers",
+    "PNDMScheduler"
+  ],
+  "text_encoder": [
+    "transformers",
+    "CLIPTextModel"
+  ],
+  "tokenizer": [
+    "transformers",
+    "CLIPTokenizer"
+  ],
+  "unet": [
+    "diffusers",
+    "UNet2DConditionModel"
+  ],
+  "vae": [
+    "diffusers",
+    "AutoencoderKL"
+  ]
+}

realistic/safety_checker/config.json

@@ -0,0 +1,168 @@
+{
+  "_commit_hash": "cb41f3a270d63d454d385fc2e4f571c487c253c5",
+  "_name_or_path": "CompVis/stable-diffusion-safety-checker",
+  "architectures": [
+    "StableDiffusionSafetyChecker"
+  ],
+  "initializer_factor": 1.0,
+  "logit_scale_init_value": 2.6592,
+  "model_type": "clip",
+  "projection_dim": 768,
+  "text_config": {
+    "_name_or_path": "",
+    "add_cross_attention": false,
+    "architectures": null,
+    "attention_dropout": 0.0,
+    "bad_words_ids": null,
+    "begin_suppress_tokens": null,
+    "bos_token_id": 49406,
+    "chunk_size_feed_forward": 0,
+    "cross_attention_hidden_size": null,
+    "decoder_start_token_id": null,
+    "diversity_penalty": 0.0,
+    "do_sample": false,
+    "dropout": 0.0,
+    "early_stopping": false,
+    "encoder_no_repeat_ngram_size": 0,
+    "eos_token_id": 49407,
+    "exponential_decay_length_penalty": null,
+    "finetuning_task": null,
+    "forced_bos_token_id": null,
+    "forced_eos_token_id": null,
+    "hidden_act": "quick_gelu",
+    "hidden_size": 768,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "initializer_factor": 1.0,
+    "initializer_range": 0.02,
+    "intermediate_size": 3072,
+    "is_decoder": false,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "layer_norm_eps": 1e-05,
+    "length_penalty": 1.0,
+    "max_length": 20,
+    "max_position_embeddings": 77,
+    "min_length": 0,
+    "model_type": "clip_text_model",
+    "no_repeat_ngram_size": 0,
+    "num_attention_heads": 12,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_hidden_layers": 12,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_scores": false,
+    "pad_token_id": 1,
+    "prefix": null,
+    "problem_type": null,
+    "projection_dim": 512,
+    "pruned_heads": {},
+    "remove_invalid_values": false,
+    "repetition_penalty": 1.0,
+    "return_dict": true,
+    "return_dict_in_generate": false,
+    "sep_token_id": null,
+    "suppress_tokens": null,
+    "task_specific_params": null,
+    "temperature": 1.0,
+    "tf_legacy_loss": false,
+    "tie_encoder_decoder": false,
+    "tie_word_embeddings": true,
+    "tokenizer_class": null,
+    "top_k": 50,
+    "top_p": 1.0,
+    "torch_dtype": null,
+    "torchscript": false,
+    "transformers_version": "4.31.0",
+    "typical_p": 1.0,
+    "use_bfloat16": false,
+    "vocab_size": 49408
+  },
+  "torch_dtype": "float32",
+  "transformers_version": null,
+  "vision_config": {
+    "_name_or_path": "",
+    "add_cross_attention": false,
+    "architectures": null,
+    "attention_dropout": 0.0,
+    "bad_words_ids": null,
+    "begin_suppress_tokens": null,
+    "bos_token_id": null,
+    "chunk_size_feed_forward": 0,
+    "cross_attention_hidden_size": null,
+    "decoder_start_token_id": null,
+    "diversity_penalty": 0.0,
+    "do_sample": false,
+    "dropout": 0.0,
+    "early_stopping": false,
+    "encoder_no_repeat_ngram_size": 0,
+    "eos_token_id": null,
+    "exponential_decay_length_penalty": null,
+    "finetuning_task": null,
+    "forced_bos_token_id": null,
+    "forced_eos_token_id": null,
+    "hidden_act": "quick_gelu",
+    "hidden_size": 1024,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "image_size": 224,
+    "initializer_factor": 1.0,
+    "initializer_range": 0.02,
+    "intermediate_size": 4096,
+    "is_decoder": false,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "layer_norm_eps": 1e-05,
+    "length_penalty": 1.0,
+    "max_length": 20,
+    "min_length": 0,
+    "model_type": "clip_vision_model",
+    "no_repeat_ngram_size": 0,
+    "num_attention_heads": 16,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_channels": 3,
+    "num_hidden_layers": 24,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_scores": false,
+    "pad_token_id": null,
+    "patch_size": 14,
+    "prefix": null,
+    "problem_type": null,
+    "projection_dim": 512,
+    "pruned_heads": {},
+    "remove_invalid_values": false,
+    "repetition_penalty": 1.0,
+    "return_dict": true,
+    "return_dict_in_generate": false,
+    "sep_token_id": null,
+    "suppress_tokens": null,
+    "task_specific_params": null,
+    "temperature": 1.0,
+    "tf_legacy_loss": false,
+    "tie_encoder_decoder": false,
+    "tie_word_embeddings": true,
+    "tokenizer_class": null,
+    "top_k": 50,
+    "top_p": 1.0,
+    "torch_dtype": null,
+    "torchscript": false,
+    "transformers_version": "4.31.0",
+    "typical_p": 1.0,
+    "use_bfloat16": false
+  }
+}

realistic/safety_checker/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:753acd54aa6d288d6c0ce9d51468eb28f495fcbaacf0edf755fa5fc7ce678cd9
+size 1216062333

realistic/scheduler/scheduler_config.json

@@ -0,0 +1,15 @@
+{
+  "_class_name": "PNDMScheduler",
+  "_diffusers_version": "0.20.0",
+  "beta_end": 0.012,
+  "beta_schedule": "scaled_linear",
+  "beta_start": 0.00085,
+  "clip_sample": false,
+  "num_train_timesteps": 1000,
+  "prediction_type": "epsilon",
+  "set_alpha_to_one": false,
+  "skip_prk_steps": true,
+  "steps_offset": 1,
+  "timestep_spacing": "leading",
+  "trained_betas": null
+}

realistic/text_encoder/config.json

@@ -0,0 +1,24 @@
+{
+  "architectures": [
+    "CLIPTextModel"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 0,
+  "dropout": 0.0,
+  "eos_token_id": 2,
+  "hidden_act": "quick_gelu",
+  "hidden_size": 768,
+  "initializer_factor": 1.0,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 77,
+  "model_type": "clip_text_model",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "pad_token_id": 1,
+  "projection_dim": 768,
+  "torch_dtype": "float32",
+  "transformers_version": "4.31.0",
+  "vocab_size": 49408
+}

realistic/text_encoder/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:046e1b2cf3561e8ba31e060f94e92ec7b3c8a890bf3571933d9de1a82abda1f6
+size 492306077

realistic/tokenizer/special_tokens_map.json

@@ -0,0 +1,24 @@
+{
+  "bos_token": {
+    "content": "<|startoftext|>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<|endoftext|>",
+  "unk_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  }
+}

realistic/tokenizer/tokenizer_config.json

@@ -0,0 +1,33 @@
+{
+  "add_prefix_space": false,
+  "bos_token": {
+    "__type": "AddedToken",
+    "content": "<|startoftext|>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "clean_up_tokenization_spaces": true,
+  "do_lower_case": true,
+  "eos_token": {
+    "__type": "AddedToken",
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "errors": "replace",
+  "model_max_length": 77,
+  "pad_token": "<|endoftext|>",
+  "tokenizer_class": "CLIPTokenizer",
+  "unk_token": {
+    "__type": "AddedToken",
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  }
+}

realistic/unet/config.json

@@ -0,0 +1,65 @@
+{
+  "_class_name": "UNet2DConditionModel",
+  "_diffusers_version": "0.20.0",
+  "act_fn": "silu",
+  "addition_embed_type": null,
+  "addition_embed_type_num_heads": 64,
+  "addition_time_embed_dim": null,
+  "attention_head_dim": 8,
+  "attention_type": "default",
+  "block_out_channels": [
+    320,
+    640,
+    1280,
+    1280
+  ],
+  "center_input_sample": false,
+  "class_embed_type": null,
+  "class_embeddings_concat": false,
+  "conv_in_kernel": 3,
+  "conv_out_kernel": 3,
+  "cross_attention_dim": 768,
+  "cross_attention_norm": null,
+  "down_block_types": [
+    "CrossAttnDownBlock2D",
+    "CrossAttnDownBlock2D",
+    "CrossAttnDownBlock2D",
+    "DownBlock2D"
+  ],
+  "downsample_padding": 1,
+  "dual_cross_attention": false,
+  "encoder_hid_dim": null,
+  "encoder_hid_dim_type": null,
+  "flip_sin_to_cos": true,
+  "freq_shift": 0,
+  "in_channels": 4,
+  "layers_per_block": 2,
+  "mid_block_only_cross_attention": null,
+  "mid_block_scale_factor": 1,
+  "mid_block_type": "UNetMidBlock2DCrossAttn",
+  "norm_eps": 1e-05,
+  "norm_num_groups": 32,
+  "num_attention_heads": null,
+  "num_class_embeds": null,
+  "only_cross_attention": false,
+  "out_channels": 4,
+  "projection_class_embeddings_input_dim": null,
+  "resnet_out_scale_factor": 1.0,
+  "resnet_skip_time_act": false,
+  "resnet_time_scale_shift": "default",
+  "sample_size": 64,
+  "time_cond_proj_dim": null,
+  "time_embedding_act_fn": null,
+  "time_embedding_dim": null,
+  "time_embedding_type": "positional",
+  "timestep_post_act": null,
+  "transformer_layers_per_block": 1,
+  "up_block_types": [
+    "UpBlock2D",
+    "CrossAttnUpBlock2D",
+    "CrossAttnUpBlock2D",
+    "CrossAttnUpBlock2D"
+  ],
+  "upcast_attention": false,
+  "use_linear_projection": false
+}

realistic/unet/diffusion_pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4ef2d71d90c97c9d0c98a68b5e07b8a7b859f27753509d7cc1bd2e3949343a1c
+size 3438366373

realistic/vae/config.json

@@ -0,0 +1,31 @@
+{
+  "_class_name": "AutoencoderKL",
+  "_diffusers_version": "0.20.0",
+  "act_fn": "silu",
+  "block_out_channels": [
+    128,
+    256,
+    512,
+    512
+  ],
+  "down_block_types": [
+    "DownEncoderBlock2D",
+    "DownEncoderBlock2D",
+    "DownEncoderBlock2D",
+    "DownEncoderBlock2D"
+  ],
+  "force_upcast": true,
+  "in_channels": 3,
+  "latent_channels": 4,
+  "layers_per_block": 2,
+  "norm_num_groups": 32,
+  "out_channels": 3,
+  "sample_size": 512,
+  "scaling_factor": 0.18215,
+  "up_block_types": [
+    "UpDecoderBlock2D",
+    "UpDecoderBlock2D",
+    "UpDecoderBlock2D",
+    "UpDecoderBlock2D"
+  ]
+}

realistic/vae/diffusion_pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1b01618945554d9840701d3453d4a9fe3db0db090164a5ed6305641306285b6f
+size 334712113