Upload 42 files

app.py

@@ -0,0 +1,359 @@
+import spaces
+import gradio as gr
+import torch
+from diffusers import StableDiffusionInpaintPipeline, StableDiffusionImg2ImgPipeline
+from PIL import Image
+import random
+import numpy as np
+import torch
+import os
+import json
+from datetime import datetime
+
+from pipeline_rf_inversionfree_edit import RectifiedFlowPipeline as RectifiedFlowEditPipeline
+
+
+pipe_edit = RectifiedFlowEditPipeline.from_pretrained("XCLIU/2_rectified_flow_from_sd_1_5", torch_dtype=torch.float32)
+pipe_edit.to("cuda")
+
+# Function to process the image
+@spaces.GPU(duration=10)
+def process_image(
+    image_layers, prompt, edit_prompt, seed, randomize_seed, num_inference_steps,
+    max_steps, learning_rate, max_source_steps, optimization_steps, true_cfg, mask_input
+):
+    image_with_mask = {
+        "image": image_layers["background"],
+        "mask": image_layers["layers"][0] if mask_input is None else mask_input
+    }
+    
+    # Set seed
+    if randomize_seed or seed is None:
+        seed = random.randint(0, 2**32 - 1)
+    generator = torch.Generator("cuda").manual_seed(int(seed))
+
+    # Unpack image and mask
+    if image_with_mask is None:
+        return None, f"❌ Please upload an image and create a mask."
+    image = image_with_mask["image"]
+    mask = image_with_mask["mask"]
+
+    if image is None or mask is None:
+        return None, f"❌ Please ensure both image and mask are provided."
+
+    # Convert images to RGB
+    image = image.convert("RGB")
+    mask = mask.split()[-1]  # Convert mask to grayscale
+
+    if not edit_prompt:
+        return None, f"❌ Please provide a prompt for editing."
+    if not prompt:
+        prompt = ""
+    # Resize the mask to match the image
+    # mask = mask.resize(image.size)
+    with torch.autocast("cuda"):
+        # Placeholder for using advanced parameters in the future
+        # Adjust parameters according to advanced settings if applicable
+        result = pipe_edit(
+            prompt=prompt,
+            edit_prompt=edit_prompt,
+            input_image=image.resize((512, 512)),
+            mask_image=mask.resize((512, 512)),
+            negative_prompt="",
+            num_inference_steps=num_inference_steps,
+            guidance_scale=true_cfg,
+            generator=generator,
+            # save_masked_image=False,
+            # output_path="",
+            learning_rate=learning_rate,
+            max_steps=max_steps,
+            optimization_steps=optimization_steps,
+            full_source_steps=max_source_steps,
+        ).images[0]
+    return result, f"✅ Editing completed with seed {seed}."
+
+# Design the Gradio interface
+with gr.Blocks() as demo:
+    gr.Markdown(
+        """
+        <style>
+            body {background-color: #f5f5f5; color: #333333;}
+            h1 {text-align: center; font-family: 'Helvetica', sans-serif; margin-bottom: 10px;}
+            h2 {text-align: center; color: #666666; font-weight: normal; margin-bottom: 30px;}
+            .gradio-container {max-width: 800px; margin: auto;}
+            .footer {text-align: center; margin-top: 20px; color: #999999; font-size: 12px;}
+        </style>
+        """
+    )
+    gr.Markdown("<h1>🍲 FlowChef 🍲</h1>")
+    gr.Markdown("<h2>Inversion/Gradient/Training-free Steering of <u>InstaFlow (SDv1.5) for Image Editing</u></h2>")
+    gr.Markdown("<h3><p><a href='https://flowchef.github.io/'>Project Page</a> | <a href='#'>Paper</a></p> (Steering Rectified Flow Models in the Vector Field for Controlled Image Generation)</h3>")
+    gr.Markdown("<h3>💡 We recommend going through our <a href='#'>tutorial introduction</a> before getting started!</h3>")
+    gr.Markdown("<h3>⚡ For better performance, check out our demo on <a href='https://huggingface.co/spaces/FlowChef/FlowChef-Flux1-dev'>Flux</a>!</h3>")
+
+    # Store current state
+    current_input_image = None
+    current_mask = None 
+    current_output_image = None
+    current_params = {}
+
+    # Images at the top
+    with gr.Row():
+        with gr.Column():
+            image_input = gr.ImageMask(
+                # source="upload",
+                # tool="sketch",
+                type="pil",
+                label="Input Image and Mask",
+                image_mode="RGBA",
+                height=512,
+                width=512,
+            )
+        with gr.Column():
+            output_image = gr.Image(label="Output Image")
+
+    # All options below
+    with gr.Column():
+        prompt = gr.Textbox(
+            label="Prompt",
+            placeholder="Describe what should appear in the masked area..."
+        )
+        edit_prompt = gr.Textbox(
+            label="Editing Prompt",
+            placeholder="Describe how you want to edit the image..."
+        )
+        with gr.Row():
+            seed = gr.Number(label="Seed (Optional)", value=None)
+            randomize_seed = gr.Checkbox(label="Randomize Seed", value=True)
+        num_inference_steps = gr.Slider(
+            label="Inference Steps", minimum=10, maximum=100, value=50
+        )
+        # Advanced settings in an accordion
+        with gr.Accordion("Advanced Settings", open=False):
+            max_steps = gr.Slider(label="Max Steps", minimum=10, maximum=100, value=50)
+            learning_rate = gr.Slider(label="Learning Rate", minimum=0.1, maximum=1.0, value=0.5)
+            true_cfg = gr.Slider(label="Guidance Scale", minimum=1, maximum=20, value=2)
+            max_source_steps = gr.Slider(label="Max Source Steps", minimum=1, maximum=200, value=40)
+            optimization_steps = gr.Slider(label="Optimization Steps", minimum=1, maximum=10, value=1)
+            mask_input = gr.Image(
+                type="pil",
+                label="Optional Mask",
+                image_mode="RGBA",
+            )
+        with gr.Row():
+            run_button = gr.Button("Run", variant="primary")
+            # save_button = gr.Button("Save Data", variant="secondary")
+
+    # def update_visibility(selected_mode):
+    #     if selected_mode == "Inpainting":
+    #         return gr.update(visible=True), gr.update(visible=False)
+    #     else:
+    #         return gr.update(visible=True), gr.update(visible=True)
+
+    # mode.change(
+    #     update_visibility,
+    #     inputs=mode,
+    #     outputs=[prompt, edit_prompt],
+    # )
+
+    def run_and_update_status(
+        image_with_mask, prompt, edit_prompt, seed, randomize_seed, num_inference_steps,
+        max_steps, learning_rate, max_source_steps, optimization_steps, true_cfg, mask_input
+    ):
+        result_image, result_status = process_image(
+            image_with_mask, prompt, edit_prompt, seed, randomize_seed, num_inference_steps,
+            max_steps, learning_rate, max_source_steps, optimization_steps, true_cfg, mask_input
+        )
+        
+        # Store current state
+        global current_input_image, current_mask, current_output_image, current_params
+
+        current_input_image = image_with_mask["background"] if image_with_mask else None
+        current_mask = mask_input if mask_input is not None else (image_with_mask["layers"][0] if image_with_mask else None)
+        current_output_image = result_image
+        current_params = {
+            "prompt": prompt,
+            "edit_prompt": edit_prompt,
+            "seed": seed,
+            "randomize_seed": randomize_seed,
+            "num_inference_steps": num_inference_steps,
+            "max_steps": max_steps,
+            "learning_rate": learning_rate,
+            "max_source_steps": max_source_steps,
+            "optimization_steps": optimization_steps,
+            "true_cfg": true_cfg,
+        }
+        
+        return result_image
+
+    def save_data():
+        if not os.path.exists("saved_results"):
+            os.makedirs("saved_results")
+            
+        timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
+        save_dir = os.path.join("saved_results", timestamp)
+        os.makedirs(save_dir)
+        
+        # Save images
+        if current_input_image:
+            current_input_image.save(os.path.join(save_dir, "input.png"))
+        if current_mask:
+            current_mask.save(os.path.join(save_dir, "mask.png"))
+        if current_output_image:
+            current_output_image.save(os.path.join(save_dir, "output.png"))
+            
+        # Save parameters
+        with open(os.path.join(save_dir, "parameters.json"), "w") as f:
+            json.dump(current_params, f, indent=4)
+            
+        return f"✅ Data saved in {save_dir}"
+
+    run_button.click(
+        fn=run_and_update_status,
+        inputs=[
+            image_input,
+            prompt,
+            edit_prompt,
+            seed,
+            randomize_seed,
+            num_inference_steps,
+            max_steps,
+            learning_rate,
+            max_source_steps,
+            optimization_steps,
+            true_cfg,
+            mask_input
+        ],
+        outputs=output_image,
+    )
+
+    # save_button.click(fn=save_data)
+
+    gr.Markdown(
+        "<div class='footer'>Developed with ❤️ using InstaFlow (Stable Diffusion v1.5) and Gradio by <a href='https://maitreyapatel.com'>Maitreya Patel</a></div>"
+    )
+
+    def load_example_image_with_mask(image_path):
+        # Load the image
+        image = Image.open(image_path)
+        # Create an empty mask of the same size
+        mask = Image.new('L', image.size, 0)
+        return {"background": image, "layers": [mask], "composite": image}
+
+    examples_dir = "assets"
+    volcano_dict = load_example_image_with_mask(os.path.join(examples_dir, "vulcano.jpg"))
+    dog_dict = load_example_image_with_mask(os.path.join(examples_dir, "dog.webp"))
+
+    gr.Examples(
+        examples=[
+            [
+                "./saved_results/20241129_154837/input.png",  # image with mask
+                "./saved_results/20241129_154837/mask.png",
+                "./saved_results/20241129_154837/output.png",
+                "a cat",  # prompt
+                "a tiger",  # edit_prompt 
+                0,  # seed
+                True,  # randomize_seed
+                50,  # num_inference_steps
+                50,  # max_steps
+                0.5,  # learning_rate
+                20,  # max_source_steps
+                5,  # optimization_steps
+                2,  # true_cfg
+            ],
+            [
+                "./saved_results/20241129_195331/input.png",  # image with mask
+                "./saved_results/20241129_195331/mask.png",
+                "./saved_results/20241129_195331/output.png",
+                "a cat",  # prompt
+                "a silver sculpture of cat",  # edit_prompt 
+                0,  # seed
+                True,  # randomize_seed
+                50,  # num_inference_steps
+                50,  # max_steps
+                0.5,  # learning_rate
+                20,  # max_source_steps
+                5,  # optimization_steps
+                2,  # true_cfg
+            ],
+            [
+                "./saved_results/20241129_160439/input.png",  # image with mask
+                "./saved_results/20241129_160439/mask.png",
+                "./saved_results/20241129_160439/output.png",
+                "a dog",  # prompt
+                "a lion",  # edit_prompt 
+                0,  # seed
+                True,  # randomize_seed
+                50,  # num_inference_steps
+                20,  # max_steps
+                0.5,  # learning_rate
+                20,  # max_source_steps
+                5,  # optimization_steps
+                4,  # true_cfg
+            ],
+            [
+                "./saved_results/20241129_161118/input.png",  # image with mask
+                "./saved_results/20241129_161118/mask.png",
+                "./saved_results/20241129_161118/output.png",
+                "two birds sitting on a branch",  # prompt
+                "two origami birds sitting on a branch",  # edit_prompt 
+                0,  # seed
+                True,  # randomize_seed
+                50,  # num_inference_steps
+                50,  # max_steps
+                0.5,  # learning_rate
+                30,  # max_source_steps
+                2,  # optimization_steps
+                2,  # true_cfg
+            ],
+            [
+                "./saved_results/20241129_161602/input.png",  # image with mask
+                "./saved_results/20241129_161602/mask.png",
+                "./saved_results/20241129_161602/output.png",
+                "a woman with long hair sitting in the sand at sunset",  # prompt
+                "a woman with short hair sitting in the sand at sunset",  # edit_prompt 
+                0,  # seed
+                True,  # randomize_seed
+                50,  # num_inference_steps
+                30,  # max_steps
+                0.5,  # learning_rate
+                20,  # max_source_steps
+                2,  # optimization_steps
+                2,  # true_cfg
+            ],
+            [
+                "./saved_results/20241129_160150/input.png",  # image with mask
+                "./saved_results/20241129_160150/mask.png",
+                "./saved_results/20241129_160150/output.png",
+                "A cute rabbit with big eyes",  # prompt
+                "A cute pig with big eyes",  # edit_prompt 
+                0,  # seed
+                True,  # randomize_seed
+                50,  # num_inference_steps
+                40,  # max_steps
+                0.5,  # learning_rate
+                20,  # max_source_steps
+                5,  # optimization_steps
+                4,  # true_cfg
+            ],
+        ],
+        inputs=[
+            image_input,
+            mask_input,
+            output_image,
+            prompt,
+            edit_prompt,
+            seed,
+            randomize_seed,
+            num_inference_steps,
+            max_steps,
+            learning_rate,
+            max_source_steps,
+            optimization_steps,
+            true_cfg,
+        ],
+        # outputs=[output_image],
+        # fn=run_and_update_status,
+        # cache_examples=True,
+    )
+demo.launch()

pipeline_rf_inversionfree_edit.py

@@ -0,0 +1,699 @@
+# Copyright 2023 The HuggingFace Team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import inspect
+from typing import Any, Callable, Dict, List, Optional, Union
+
+import torch
+from packaging import version
+from transformers import CLIPImageProcessor, CLIPTextModel, CLIPTokenizer
+
+from diffusers.configuration_utils import FrozenDict
+from diffusers.image_processor import VaeImageProcessor
+from diffusers.loaders import FromSingleFileMixin, LoraLoaderMixin, TextualInversionLoaderMixin
+from diffusers.models import AutoencoderKL, UNet2DConditionModel
+from diffusers.models.lora import adjust_lora_scale_text_encoder
+from diffusers.schedulers import KarrasDiffusionSchedulers
+from diffusers.utils import (
+    deprecate,
+    logging,
+    replace_example_docstring,
+)
+from diffusers.utils.torch_utils import randn_tensor
+from diffusers.pipelines.pipeline_utils import DiffusionPipeline
+from diffusers.pipelines.stable_diffusion import StableDiffusionPipelineOutput
+from diffusers.pipelines.stable_diffusion.safety_checker import StableDiffusionSafetyChecker
+import kornia.augmentation as K
+
+import os
+import torch
+import cv2
+
+import numpy as np
+from torchvision import transforms as TF
+from tqdm import tqdm
+
+import sys
+sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), '..', '..')))
+sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), '..', '..')))
+
+logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
+
+def retrieve_latents(
+    encoder_output: torch.Tensor, generator: Optional[torch.Generator] = None, sample_mode: str = "sample"
+):
+    if hasattr(encoder_output, "latent_dist") and sample_mode == "sample":
+        return encoder_output.latent_dist.sample(generator)
+    elif hasattr(encoder_output, "latent_dist") and sample_mode == "argmax":
+        return encoder_output.latent_dist.mode()
+    elif hasattr(encoder_output, "latents"):
+        return encoder_output.latents
+    else:
+        raise AttributeError("Could not access latents of provided encoder_output")
+
+
+def rescale_noise_cfg(noise_cfg, noise_pred_text, guidance_rescale=0.0):
+    """
+    Rescale `noise_cfg` according to `guidance_rescale`. Based on findings of [Common Diffusion Noise Schedules and
+    Sample Steps are Flawed](https://arxiv.org/pdf/2305.08891.pdf). See Section 3.4
+    """
+    std_text = noise_pred_text.std(dim=list(range(1, noise_pred_text.ndim)), keepdim=True)
+    std_cfg = noise_cfg.std(dim=list(range(1, noise_cfg.ndim)), keepdim=True)
+    # rescale the results from guidance (fixes overexposure)
+    noise_pred_rescaled = noise_cfg * (std_text / std_cfg)
+    # mix with the original results from guidance by factor guidance_rescale to avoid "plain looking" images
+    noise_cfg = guidance_rescale * noise_pred_rescaled + (1 - guidance_rescale) * noise_cfg
+    return noise_cfg
+
+
+class RectifiedFlowPipeline(DiffusionPipeline, TextualInversionLoaderMixin, LoraLoaderMixin, FromSingleFileMixin):
+    r"""
+    Pipeline for text-to-image generation using Rectified Flow and Euler discretization.
+    This customized pipeline is based on StableDiffusionPipeline from the official Diffusers library (0.21.4)
+
+    This model inherits from [`DiffusionPipeline`]. Check the superclass documentation for the generic methods
+    implemented for all pipelines (downloading, saving, running on a particular device, etc.).
+
+    The pipeline also inherits the following loading methods:
+        - [`~loaders.TextualInversionLoaderMixin.load_textual_inversion`] for loading textual inversion embeddings
+        - [`~loaders.LoraLoaderMixin.load_lora_weights`] for loading LoRA weights
+        - [`~loaders.LoraLoaderMixin.save_lora_weights`] for saving LoRA weights
+        - [`~loaders.FromSingleFileMixin.from_single_file`] for loading `.ckpt` files
+
+    Args:
+        vae ([`AutoencoderKL`]):
+            Variational Auto-Encoder (VAE) model to encode and decode images to and from latent representations.
+        text_encoder ([`~transformers.CLIPTextModel`]):
+            Frozen text-encoder ([clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14)).
+        tokenizer ([`~transformers.CLIPTokenizer`]):
+            A `CLIPTokenizer` to tokenize text.
+        unet ([`UNet2DConditionModel`]):
+            A `UNet2DConditionModel` to denoise the encoded image latents.
+        scheduler ([`SchedulerMixin`]):
+            A scheduler to be used in combination with `unet` to denoise the encoded image latents. Can be one of
+            [`DDIMScheduler`], [`LMSDiscreteScheduler`], or [`PNDMScheduler`].
+        safety_checker ([`StableDiffusionSafetyChecker`]):
+            Classification module that estimates whether generated images could be considered offensive or harmful.
+            Please refer to the [model card](https://huggingface.co/runwayml/stable-diffusion-v1-5) for more details
+            about a model's potential harms.
+        feature_extractor ([`~transformers.CLIPImageProcessor`]):
+            A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
+    """
+    model_cpu_offload_seq = "text_encoder->unet->vae"
+    _optional_components = ["safety_checker", "feature_extractor"]
+    _exclude_from_cpu_offload = ["safety_checker"]
+
+    def __init__(
+        self,
+        vae: AutoencoderKL,
+        text_encoder: CLIPTextModel,
+        tokenizer: CLIPTokenizer,
+        unet: UNet2DConditionModel,
+        scheduler: KarrasDiffusionSchedulers,
+        safety_checker: StableDiffusionSafetyChecker,
+        feature_extractor: CLIPImageProcessor,
+        requires_safety_checker: bool = True,
+    ):
+        super().__init__()
+
+        if hasattr(scheduler.config, "steps_offset") and scheduler.config.steps_offset != 1:
+            deprecation_message = (
+                f"The configuration file of this scheduler: {scheduler} is outdated. `steps_offset`"
+                f" should be set to 1 instead of {scheduler.config.steps_offset}. Please make sure "
+                "to update the config accordingly as leaving `steps_offset` might led to incorrect results"
+                " in future versions. If you have downloaded this checkpoint from the Hugging Face Hub,"
+                " it would be very nice if you could open a Pull request for the `scheduler/scheduler_config.json`"
+                " file"
+            )
+            deprecate("steps_offset!=1", "1.0.0", deprecation_message, standard_warn=False)
+            new_config = dict(scheduler.config)
+            new_config["steps_offset"] = 1
+            scheduler._internal_dict = FrozenDict(new_config)
+
+        if hasattr(scheduler.config, "clip_sample") and scheduler.config.clip_sample is True:
+            deprecation_message = (
+                f"The configuration file of this scheduler: {scheduler} has not set the configuration `clip_sample`."
+                " `clip_sample` should be set to False in the configuration file. Please make sure to update the"
+                " config accordingly as not setting `clip_sample` in the config might lead to incorrect results in"
+                " future versions. If you have downloaded this checkpoint from the Hugging Face Hub, it would be very"
+                " nice if you could open a Pull request for the `scheduler/scheduler_config.json` file"
+            )
+            deprecate("clip_sample not set", "1.0.0", deprecation_message, standard_warn=False)
+            new_config = dict(scheduler.config)
+            new_config["clip_sample"] = False
+            scheduler._internal_dict = FrozenDict(new_config)
+
+        if safety_checker is None and requires_safety_checker:
+            logger.warning(
+                f"You have disabled the safety checker for {self.__class__} by passing `safety_checker=None`. Ensure"
+                " that you abide to the conditions of the Stable Diffusion license and do not expose unfiltered"
+                " results in services or applications open to the public. Both the diffusers team and Hugging Face"
+                " strongly recommend to keep the safety filter enabled in all public facing circumstances, disabling"
+                " it only for use-cases that involve analyzing network behavior or auditing its results. For more"
+                " information, please have a look at https://github.com/huggingface/diffusers/pull/254 ."
+            )
+
+        if safety_checker is not None and feature_extractor is None:
+            raise ValueError(
+                "Make sure to define a feature extractor when loading {self.__class__} if you want to use the safety"
+                " checker. If you do not want to use the safety checker, you can pass `'safety_checker=None'` instead."
+            )
+
+        is_unet_version_less_0_9_0 = hasattr(unet.config, "_diffusers_version") and version.parse(
+            version.parse(unet.config._diffusers_version).base_version
+        ) < version.parse("0.9.0.dev0")
+        is_unet_sample_size_less_64 = hasattr(unet.config, "sample_size") and unet.config.sample_size < 64
+        if is_unet_version_less_0_9_0 and is_unet_sample_size_less_64:
+            deprecation_message = (
+                "The configuration file of the unet has set the default `sample_size` to smaller than"
+                " 64 which seems highly unlikely. If your checkpoint is a fine-tuned version of any of the"
+                " following: \n- CompVis/stable-diffusion-v1-4 \n- CompVis/stable-diffusion-v1-3 \n-"
+                " CompVis/stable-diffusion-v1-2 \n- CompVis/stable-diffusion-v1-1 \n- runwayml/stable-diffusion-v1-5"
+                " \n- runwayml/stable-diffusion-inpainting \n you should change 'sample_size' to 64 in the"
+                " configuration file. Please make sure to update the config accordingly as leaving `sample_size=32`"
+                " in the config might lead to incorrect results in future versions. If you have downloaded this"
+                " checkpoint from the Hugging Face Hub, it would be very nice if you could open a Pull request for"
+                " the `unet/config.json` file"
+            )
+            deprecate("sample_size<64", "1.0.0", deprecation_message, standard_warn=False)
+            new_config = dict(unet.config)
+            new_config["sample_size"] = 64
+            unet._internal_dict = FrozenDict(new_config)
+
+        self.register_modules(
+            vae=vae,
+            text_encoder=text_encoder,
+            tokenizer=tokenizer,
+            unet=unet,
+            scheduler=scheduler,
+            safety_checker=safety_checker,
+            feature_extractor=feature_extractor,
+        )
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor)
+        self.register_to_config(requires_safety_checker=requires_safety_checker)
+
+    def enable_vae_slicing(self):
+        r"""
+        Enable sliced VAE decoding. When this option is enabled, the VAE will split the input tensor in slices to
+        compute decoding in several steps. This is useful to save some memory and allow larger batch sizes.
+        """
+        self.vae.enable_slicing()
+
+    def disable_vae_slicing(self):
+        r"""
+        Disable sliced VAE decoding. If `enable_vae_slicing` was previously enabled, this method will go back to
+        computing decoding in one step.
+        """
+        self.vae.disable_slicing()
+
+    def enable_vae_tiling(self):
+        r"""
+        Enable tiled VAE decoding. When this option is enabled, the VAE will split the input tensor into tiles to
+        compute decoding and encoding in several steps. This is useful for saving a large amount of memory and to allow
+        processing larger images.
+        """
+        self.vae.enable_tiling()
+
+    def disable_vae_tiling(self):
+        r"""
+        Disable tiled VAE decoding. If `enable_vae_tiling` was previously enabled, this method will go back to
+        computing decoding in one step.
+        """
+        self.vae.disable_tiling()
+
+    def _encode_prompt(
+        self,
+        prompt,
+        device,
+        num_images_per_prompt,
+        do_classifier_free_guidance,
+        negative_prompt=None,
+        prompt_embeds: Optional[torch.FloatTensor] = None,
+        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        lora_scale: Optional[float] = None,
+    ):
+        deprecation_message = "`_encode_prompt()` is deprecated and it will be removed in a future version. Use `encode_prompt()` instead. Also, be aware that the output format changed from a concatenated tensor to a tuple."
+        deprecate("_encode_prompt()", "1.0.0", deprecation_message, standard_warn=False)
+
+        prompt_embeds_tuple = self.encode_prompt(
+            prompt=prompt,
+            device=device,
+            num_images_per_prompt=num_images_per_prompt,
+            do_classifier_free_guidance=do_classifier_free_guidance,
+            negative_prompt=negative_prompt,
+            prompt_embeds=prompt_embeds,
+            negative_prompt_embeds=negative_prompt_embeds,
+            lora_scale=lora_scale,
+        )
+
+        # concatenate for backwards comp
+        prompt_embeds = torch.cat([prompt_embeds_tuple[1], prompt_embeds_tuple[0]])
+
+        return prompt_embeds
+
+    def encode_prompt(
+        self,
+        prompt,
+        device,
+        num_images_per_prompt,
+        do_classifier_free_guidance,
+        negative_prompt=None,
+        prompt_embeds: Optional[torch.FloatTensor] = None,
+        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        lora_scale: Optional[float] = None,
+    ):
+        r"""
+        Encodes the prompt into text encoder hidden states.
+
+        Args:
+            prompt (`str` or `List[str]`, *optional*):
+                prompt to be encoded
+            device: (`torch.device`):
+                torch device
+            num_images_per_prompt (`int`):
+                number of images that should be generated per prompt
+            do_classifier_free_guidance (`bool`):
+                whether to use classifier free guidance or not
+            negative_prompt (`str` or `List[str]`, *optional*):
+                The prompt or prompts not to guide the image generation. If not defined, one has to pass
+                `negative_prompt_embeds` instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is
+                less than `1`).
+            prompt_embeds (`torch.FloatTensor`, *optional*):
+                Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
+                provided, text embeddings will be generated from `prompt` input argument.
+            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+                Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
+                weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
+                argument.
+            lora_scale (`float`, *optional*):
+                A lora scale that will be applied to all LoRA layers of the text encoder if LoRA layers are loaded.
+        """
+        # set lora scale so that monkey patched LoRA
+        # function of text encoder can correctly access it
+        if lora_scale is not None and isinstance(self, LoraLoaderMixin):
+            self._lora_scale = lora_scale
+
+            # dynamically adjust the LoRA scale
+            adjust_lora_scale_text_encoder(self.text_encoder, lora_scale)
+
+        if prompt is not None and isinstance(prompt, str):
+            batch_size = 1
+        elif prompt is not None and isinstance(prompt, list):
+            batch_size = len(prompt)
+        else:
+            batch_size = prompt_embeds.shape[0]
+
+        if prompt_embeds is None:
+            # textual inversion: procecss multi-vector tokens if necessary
+            if isinstance(self, TextualInversionLoaderMixin):
+                prompt = self.maybe_convert_prompt(prompt, self.tokenizer)
+
+            text_inputs = self.tokenizer(
+                prompt,
+                padding="max_length",
+                max_length=self.tokenizer.model_max_length,
+                truncation=True,
+                return_tensors="pt",
+            )
+            text_input_ids = text_inputs.input_ids
+            untruncated_ids = self.tokenizer(prompt, padding="longest", return_tensors="pt").input_ids
+
+            if untruncated_ids.shape[-1] >= text_input_ids.shape[-1] and not torch.equal(
+                text_input_ids, untruncated_ids
+            ):
+                removed_text = self.tokenizer.batch_decode(
+                    untruncated_ids[:, self.tokenizer.model_max_length - 1 : -1]
+                )
+                logger.warning(
+                    "The following part of your input was truncated because CLIP can only handle sequences up to"
+                    f" {self.tokenizer.model_max_length} tokens: {removed_text}"
+                )
+
+            if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask:
+                attention_mask = text_inputs.attention_mask.to(device)
+            else:
+                attention_mask = None
+
+            prompt_embeds = self.text_encoder(
+                text_input_ids.to(device),
+                attention_mask=attention_mask,
+            )
+            prompt_embeds = prompt_embeds[0]
+
+        if self.text_encoder is not None:
+            prompt_embeds_dtype = self.text_encoder.dtype
+        elif self.unet is not None:
+            prompt_embeds_dtype = self.unet.dtype
+        else:
+            prompt_embeds_dtype = prompt_embeds.dtype
+
+        prompt_embeds = prompt_embeds.to(dtype=prompt_embeds_dtype, device=device)
+
+        bs_embed, seq_len, _ = prompt_embeds.shape
+        # duplicate text embeddings for each generation per prompt, using mps friendly method
+        prompt_embeds = prompt_embeds.repeat(1, num_images_per_prompt, 1)
+        prompt_embeds = prompt_embeds.view(bs_embed * num_images_per_prompt, seq_len, -1)
+
+        # get unconditional embeddings for classifier free guidance
+        if do_classifier_free_guidance and negative_prompt_embeds is None:
+            uncond_tokens: List[str]
+            if negative_prompt is None:
+                uncond_tokens = [""] * batch_size
+            elif prompt is not None and type(prompt) is not type(negative_prompt):
+                raise TypeError(
+                    f"`negative_prompt` should be the same type to `prompt`, but got {type(negative_prompt)} !="
+                    f" {type(prompt)}."
+                )
+            elif isinstance(negative_prompt, str):
+                uncond_tokens = [negative_prompt]
+            elif batch_size != len(negative_prompt):
+                raise ValueError(
+                    f"`negative_prompt`: {negative_prompt} has batch size {len(negative_prompt)}, but `prompt`:"
+                    f" {prompt} has batch size {batch_size}. Please make sure that passed `negative_prompt` matches"
+                    " the batch size of `prompt`."
+                )
+            else:
+                uncond_tokens = negative_prompt
+
+            # textual inversion: procecss multi-vector tokens if necessary
+            if isinstance(self, TextualInversionLoaderMixin):
+                uncond_tokens = self.maybe_convert_prompt(uncond_tokens, self.tokenizer)
+
+            max_length = prompt_embeds.shape[1]
+            uncond_input = self.tokenizer(
+                uncond_tokens,
+                padding="max_length",
+                max_length=max_length,
+                truncation=True,
+                return_tensors="pt",
+            )
+
+            if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask:
+                attention_mask = uncond_input.attention_mask.to(device)
+            else:
+                attention_mask = None
+
+            negative_prompt_embeds = self.text_encoder(
+                uncond_input.input_ids.to(device),
+                attention_mask=attention_mask,
+            )
+            negative_prompt_embeds = negative_prompt_embeds[0]
+
+        if do_classifier_free_guidance:
+            # duplicate unconditional embeddings for each generation per prompt, using mps friendly method
+            seq_len = negative_prompt_embeds.shape[1]
+
+            negative_prompt_embeds = negative_prompt_embeds.to(dtype=prompt_embeds_dtype, device=device)
+
+            negative_prompt_embeds = negative_prompt_embeds.repeat(1, num_images_per_prompt, 1)
+            negative_prompt_embeds = negative_prompt_embeds.view(batch_size * num_images_per_prompt, seq_len, -1)
+
+        return prompt_embeds, negative_prompt_embeds
+
+    def run_safety_checker(self, image, device, dtype):
+        if self.safety_checker is None:
+            has_nsfw_concept = None
+        else:
+            if torch.is_tensor(image):
+                feature_extractor_input = self.image_processor.postprocess(image, output_type="pil")
+            else:
+                feature_extractor_input = self.image_processor.numpy_to_pil(image)
+            safety_checker_input = self.feature_extractor(feature_extractor_input, return_tensors="pt").to(device)
+            image, has_nsfw_concept = self.safety_checker(
+                images=image, clip_input=safety_checker_input.pixel_values.to(dtype)
+            )
+        return image, has_nsfw_concept
+
+    def decode_latents(self, latents):
+        deprecation_message = "The decode_latents method is deprecated and will be removed in 1.0.0. Please use VaeImageProcessor.postprocess(...) instead"
+        deprecate("decode_latents", "1.0.0", deprecation_message, standard_warn=False)
+
+        latents = 1 / self.vae.config.scaling_factor * latents
+        image = self.vae.decode(latents, return_dict=False)[0]
+        image = (image / 2 + 0.5).clamp(0, 1)
+        # we always cast to float32 as this does not cause significant overhead and is compatible with bfloat16
+        image = image.cpu().permute(0, 2, 3, 1).float().numpy()
+        return image
+
+    def prepare_extra_step_kwargs(self, generator, eta):
+        # prepare extra kwargs for the scheduler step, since not all schedulers have the same signature
+        # eta (η) is only used with the DDIMScheduler, it will be ignored for other schedulers.
+        # eta corresponds to η in DDIM paper: https://arxiv.org/abs/2010.02502
+        # and should be between [0, 1]
+
+        accepts_eta = "eta" in set(inspect.signature(self.scheduler.step).parameters.keys())
+        extra_step_kwargs = {}
+        if accepts_eta:
+            extra_step_kwargs["eta"] = eta
+
+        # check if the scheduler accepts generator
+        accepts_generator = "generator" in set(inspect.signature(self.scheduler.step).parameters.keys())
+        if accepts_generator:
+            extra_step_kwargs["generator"] = generator
+        return extra_step_kwargs
+
+    def check_inputs(
+        self,
+        prompt,
+        height,
+        width,
+        callback_steps,
+        negative_prompt=None,
+        prompt_embeds=None,
+        negative_prompt_embeds=None,
+    ):
+        if height % 8 != 0 or width % 8 != 0:
+            raise ValueError(f"`height` and `width` have to be divisible by 8 but are {height} and {width}.")
+
+        if (callback_steps is None) or (
+            callback_steps is not None and (not isinstance(callback_steps, int) or callback_steps <= 0)
+        ):
+            raise ValueError(
+                f"`callback_steps` has to be a positive integer but is {callback_steps} of type"
+                f" {type(callback_steps)}."
+            )
+
+        if prompt is not None and prompt_embeds is not None:
+            raise ValueError(
+                f"Cannot forward both `prompt`: {prompt} and `prompt_embeds`: {prompt_embeds}. Please make sure to"
+                " only forward one of the two."
+            )
+        elif prompt is None and prompt_embeds is None:
+            raise ValueError(
+                "Provide either `prompt` or `prompt_embeds`. Cannot leave both `prompt` and `prompt_embeds` undefined."
+            )
+        elif prompt is not None and (not isinstance(prompt, str) and not isinstance(prompt, list)):
+            raise ValueError(f"`prompt` has to be of type `str` or `list` but is {type(prompt)}")
+
+        if negative_prompt is not None and negative_prompt_embeds is not None:
+            raise ValueError(
+                f"Cannot forward both `negative_prompt`: {negative_prompt} and `negative_prompt_embeds`:"
+                f" {negative_prompt_embeds}. Please make sure to only forward one of the two."
+            )
+
+        if prompt_embeds is not None and negative_prompt_embeds is not None:
+            if prompt_embeds.shape != negative_prompt_embeds.shape:
+                raise ValueError(
+                    "`prompt_embeds` and `negative_prompt_embeds` must have the same shape when passed directly, but"
+                    f" got: `prompt_embeds` {prompt_embeds.shape} != `negative_prompt_embeds`"
+                    f" {negative_prompt_embeds.shape}."
+                )
+
+    def prepare_latents(self, batch_size, num_channels_latents, height, width, dtype, device, generator, latents=None):
+        shape = (batch_size, num_channels_latents, height // self.vae_scale_factor, width // self.vae_scale_factor)
+        if isinstance(generator, list) and len(generator) != batch_size:
+            raise ValueError(
+                f"You have passed a list of generators of length {len(generator)}, but requested an effective batch"
+                f" size of {batch_size}. Make sure the batch size matches the length of the generators."
+            )
+
+        if latents is None:
+            latents = randn_tensor(shape, generator=generator, device=device, dtype=dtype)
+        else:
+            latents = latents.to(device)
+
+        # scale the initial noise by the standard deviation required by the scheduler
+        latents = latents * self.scheduler.init_noise_sigma
+        return latents
+
+    @torch.no_grad()
+    def __call__(
+        self,
+        prompt: Union[str, List[str]] = None,
+        edit_prompt: Union[str, List[str]] = None,
+        height: Optional[int] = None,
+        width: Optional[int] = None,
+        num_inference_steps: int = 50,
+        guidance_scale: float = 7.5,
+        negative_prompt: Optional[Union[str, List[str]]] = None,
+        num_images_per_prompt: Optional[int] = 1,
+        eta: float = 0.0,
+        generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
+        latents: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.FloatTensor] = None,
+        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        output_type: Optional[str] = "pil",
+        return_dict: bool = True,
+        callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
+        callback_steps: int = 1,
+        cross_attention_kwargs: Optional[Dict[str, Any]] = None,
+        guidance_rescale: float = 0.0,
+        optimization_steps: int = 1,
+        learning_rate: float = 0.05,
+        max_steps: int = 50,
+        input_image = None,
+        mask_image = None,
+        full_source_steps: int = 10,
+    ):
+        # 0. Default height and width to unet
+        height = height or self.unet.config.sample_size * self.vae_scale_factor
+        width = width or self.unet.config.sample_size * self.vae_scale_factor
+
+        # 1. Check inputs
+        self.check_inputs(prompt, height, width, callback_steps, negative_prompt, prompt_embeds, negative_prompt_embeds)
+
+        # 2. Define call parameters
+        batch_size = 1 if prompt is None else (1 if isinstance(prompt, str) else len(prompt))
+        device = self._execution_device
+        do_classifier_free_guidance = guidance_scale > 1.0
+
+        # 3. Encode input prompt
+        target_prompt_embeds, negative_prompt_embeds = self.encode_prompt(
+            edit_prompt, device, num_images_per_prompt, do_classifier_free_guidance, negative_prompt=prompt,
+            prompt_embeds=None,
+            lora_scale=cross_attention_kwargs.get("scale", None) if cross_attention_kwargs is not None else None,
+        )
+        if do_classifier_free_guidance:
+            target_prompt_embeds = torch.cat([negative_prompt_embeds, target_prompt_embeds])
+
+        # 4. Prepare timesteps
+        timesteps = [(1. - i/num_inference_steps) * 1000. for i in range(num_inference_steps)]
+
+        # 5. Prepare latent variables
+        num_channels_latents = self.unet.config.in_channels
+
+        # 6. Prepare extra step kwargs
+        extra_step_kwargs = self.prepare_extra_step_kwargs(generator, eta)
+
+
+        image = self.image_processor.preprocess(input_image)
+        image = image.to(device=device, dtype=self.unet.dtype)
+
+
+        gt_latents = retrieve_latents(self.vae.encode(image), generator=generator) * self.vae.config.scaling_factor
+
+
+        random_latents = self.prepare_latents(
+            batch_size * num_images_per_prompt, num_channels_latents, height, width,
+            target_prompt_embeds.dtype, device, generator
+        )
+        latents = random_latents.clone().detach()
+
+        mask = TF.ToTensor()(mask_image).to(device)
+        if mask.shape[0] == 3:
+            mask = mask[0]
+        mask = mask.float()
+        mask = mask.unsqueeze(0)#.unsqueeze(0)
+
+        mask = torch.nn.functional.interpolate(mask, size=(height // self.vae_scale_factor, width // self.vae_scale_factor), mode='bilinear', align_corners=False)
+        mask = ~mask.expand_as(latents).bool()
+
+
+        self.unet.eval()
+        self.vae.eval()
+        self.vae.eval()
+
+        latents = random_latents.clone().detach()
+        with self.progress_bar(total=num_inference_steps) as progress_bar:
+            for i, t in enumerate(timesteps):
+                latents, _ = self.perform_denoising_step(
+                    latents, t, target_prompt_embeds, do_classifier_free_guidance, guidance_scale,
+                    device, i, optimization_steps, learning_rate,
+                    max_steps, timesteps, gt_latents, full_source_steps, mask=mask
+                )
+
+                if callback is not None and i % callback_steps == 0:
+                    callback(i // getattr(self.scheduler, "order", 1), t, latents)
+
+                progress_bar.update()
+
+        # 10. Post-processing
+        image = self.post_process_image(latents, output_type)
+
+        # 11. Offload all models
+        self.maybe_free_model_hooks()
+
+        if not return_dict:
+            return (image, None)
+
+        return StableDiffusionPipelineOutput(images=image, nsfw_content_detected=None)
+
+    def load_and_preprocess_image(self, image_path, custom_image_processor, device):
+        from diffusers.utils import load_image
+        img = load_image(image_path)
+        img = img.resize((512, 512))
+        return custom_image_processor(img).unsqueeze(0).to(device)
+
+    def perform_denoising_step(self, latents, t, prompt_embeds, do_classifier_free_guidance, guidance_scale,
+                               device, step, optimization_steps, learning_rate,
+                                max_steps, timesteps, gt_latents, full_source_steps, mask=None):
+        latent_model_input = torch.cat([latents] * 2) if do_classifier_free_guidance else latents
+        vec_t = torch.ones((latent_model_input.shape[0],), device=latents.device) * t 
+        v_pred = self.unet(latent_model_input, vec_t, encoder_hidden_states=prompt_embeds).sample
+
+        if do_classifier_free_guidance:
+            v_pred_neg, v_pred_text = v_pred.chunk(2)
+            v_pred = v_pred_text + ~mask*(v_pred_text - v_pred_neg) * guidance_scale
+        else:
+            mask = torch.ones_like(v_pred, dtype=torch.float32)
+
+        if step < full_source_steps:
+            mask = torch.ones_like(v_pred, dtype=torch.float32)
+
+        if step <= max_steps:
+            latents = self.optimize_latents(latents, v_pred_neg, t,
+                                            device, optimization_steps, learning_rate, gt_latents, mask)
+            
+        return latents + (1.0 / len(timesteps)) * v_pred, mask
+
+    def optimize_latents(self, latents, v_pred, t, device, optimization_steps, learning_rate,
+                         gt_latents, mask):
+        with torch.enable_grad():
+            latent_copy = latents.clone().detach()
+            latents = torch.autograd.Variable(latents, requires_grad=True)
+            optimizer = torch.optim.Adam([latents], lr=learning_rate)
+
+            # latents_new = latents + new_parameter
+            for _ in range(optimization_steps):
+                latents_p = latents + t/1000 * v_pred
+
+                loss = (0.00001*torch.nn.functional.mse_loss(latents_p, gt_latents, reduction='none')*mask).mean()
+                loss.backward()
+                optimizer.step()
+                optimizer.zero_grad()
+
+        return latents
+
+    def decode_latents(self, latents):
+        image = self.vae.decode(latents / self.vae.config.scaling_factor, return_dict=False)[0]
+        return self.image_processor.postprocess(image, output_type="pt")[0]
+
+    def post_process_image(self, latents, output_type):
+        if output_type == "latent":
+            return latents
+
+        image = self.vae.decode(latents / self.vae.config.scaling_factor, return_dict=False)[0]
+        do_denormalize = [True] * image.shape[0]
+        return self.image_processor.postprocess(image, output_type=output_type, do_denormalize=do_denormalize)

requirements.txt

@@ -0,0 +1,11 @@
+spaces
+diffusers
+gradio==5.6.0
+numpy
+Pillow
+torch
+torchvision
+transformers
+accelerate
+xformers
+sentencepiece

saved_results/20241129_041830/parameters.json

@@ -0,0 +1,14 @@
+{
+    "mode": "Editing",
+    "prompt": "a dog",
+    "edit_prompt": "a tiger",
+    "seed": 0,
+    "randomize_seed": true,
+    "num_inference_steps": 50,
+    "max_steps": 50,
+    "learning_rate": 1,
+    "max_source_steps": 20,
+    "optimization_steps": 5,
+    "true_cfg": 2,
+    "inverseproblem": false
+}

saved_results/20241129_154837/parameters.json

@@ -0,0 +1,13 @@
+{
+    "prompt": "a cat",
+    "edit_prompt": "a tiger",
+    "seed": 0,
+    "randomize_seed": true,
+    "num_inference_steps": 50,
+    "max_steps": 50,
+    "learning_rate": 0.5,
+    "max_source_steps": 20,
+    "optimization_steps": 5,
+    "true_cfg": 2,
+    "inverseproblem": false
+}

saved_results/20241129_160150/parameters.json

@@ -0,0 +1,13 @@
+{
+    "prompt": "A cute rabbit with big eyes",
+    "edit_prompt": "A cute pig with big eyes",
+    "seed": 0,
+    "randomize_seed": true,
+    "num_inference_steps": 50,
+    "max_steps": 40,
+    "learning_rate": 0.5,
+    "max_source_steps": 20,
+    "optimization_steps": 5,
+    "true_cfg": 4,
+    "inverseproblem": false
+}

saved_results/20241129_160359/parameters.json

@@ -0,0 +1,13 @@
+{
+    "prompt": "a dog",
+    "edit_prompt": "a lion",
+    "seed": 0,
+    "randomize_seed": true,
+    "num_inference_steps": 50,
+    "max_steps": 40,
+    "learning_rate": 0.5,
+    "max_source_steps": 20,
+    "optimization_steps": 5,
+    "true_cfg": 4,
+    "inverseproblem": false
+}

saved_results/20241129_160439/parameters.json

@@ -0,0 +1,13 @@
+{
+    "prompt": "a dog",
+    "edit_prompt": "a lion",
+    "seed": 0,
+    "randomize_seed": true,
+    "num_inference_steps": 50,
+    "max_steps": 20,
+    "learning_rate": 0.5,
+    "max_source_steps": 20,
+    "optimization_steps": 5,
+    "true_cfg": 4,
+    "inverseproblem": false
+}

saved_results/20241129_161118/parameters.json

@@ -0,0 +1,13 @@
+{
+    "prompt": "two birds sitting on a branch",
+    "edit_prompt": "two origami birds sitting on a branch",
+    "seed": 0,
+    "randomize_seed": true,
+    "num_inference_steps": 50,
+    "max_steps": 50,
+    "learning_rate": 0.5,
+    "max_source_steps": 30,
+    "optimization_steps": 2,
+    "true_cfg": 2,
+    "inverseproblem": false
+}

saved_results/20241129_161547/parameters.json

@@ -0,0 +1,13 @@
+{
+    "prompt": " a woman with long hair sitting in the sand at sunset",
+    "edit_prompt": "a woman with short hair sitting in the sand at sunset",
+    "seed": 0,
+    "randomize_seed": true,
+    "num_inference_steps": 50,
+    "max_steps": 50,
+    "learning_rate": 0.5,
+    "max_source_steps": 20,
+    "optimization_steps": 2,
+    "true_cfg": 2,
+    "inverseproblem": false
+}

saved_results/20241129_161602/parameters.json

@@ -0,0 +1,13 @@
+{
+    "prompt": " a woman with long hair sitting in the sand at sunset",
+    "edit_prompt": "a woman with short hair sitting in the sand at sunset",
+    "seed": 0,
+    "randomize_seed": true,
+    "num_inference_steps": 50,
+    "max_steps": 30,
+    "learning_rate": 0.5,
+    "max_source_steps": 20,
+    "optimization_steps": 2,
+    "true_cfg": 2,
+    "inverseproblem": false
+}

saved_results/20241129_195331/parameters.json

@@ -0,0 +1,12 @@
+{
+    "prompt": "a cat",
+    "edit_prompt": "a silver sculpture of cat",
+    "seed": 0,
+    "randomize_seed": true,
+    "num_inference_steps": 50,
+    "max_steps": 50,
+    "learning_rate": 0.5,
+    "max_source_steps": 20,
+    "optimization_steps": 5,
+    "true_cfg": 2,
+}