InvokeAI 2.2.4.5__py3-none-any.whl

backend/modules/create_cmd_parser.py

@@ -0,0 +1,55 @@
+import argparse
+import os
+from ldm.invoke.args import PRECISION_CHOICES
+
+
+def create_cmd_parser():
+    parser = argparse.ArgumentParser(description="InvokeAI web UI")
+    parser.add_argument(
+        "--host",
+        type=str,
+        help="The host to serve on",
+        default="localhost",
+    )
+    parser.add_argument("--port", type=int, help="The port to serve on", default=9090)
+    parser.add_argument(
+        "--cors",
+        nargs="*",
+        type=str,
+        help="Additional allowed origins, comma-separated",
+    )
+    parser.add_argument(
+        "--embedding_path",
+        type=str,
+        help="Path to a pre-trained embedding manager checkpoint - can only be set on command line",
+    )
+    # TODO: Can't get flask to serve images from any dir (saving to the dir does work when specified)
+    # parser.add_argument(
+    #     "--output_dir",
+    #     default="outputs/",
+    #     type=str,
+    #     help="Directory for output images",
+    # )
+    parser.add_argument(
+        "-v",
+        "--verbose",
+        action="store_true",
+        help="Enables verbose logging",
+    )
+    parser.add_argument(
+        "--precision",
+        dest="precision",
+        type=str,
+        choices=PRECISION_CHOICES,
+        metavar="PRECISION",
+        help=f'Set model precision. Defaults to auto selected based on device. Options: {", ".join(PRECISION_CHOICES)}',
+        default="auto",
+    )
+    parser.add_argument(
+        '--free_gpu_mem',
+        dest='free_gpu_mem',
+        action='store_true',
+        help='Force free gpu memory before final decoding',
+    )
+
+    return parser

backend/modules/get_canvas_generation_mode.py

@@ -0,0 +1,117 @@
+from PIL import Image, ImageChops
+from PIL.Image import Image as ImageType
+from typing import Union, Literal
+
+# https://stackoverflow.com/questions/43864101/python-pil-check-if-image-is-transparent
+def check_for_any_transparency(img: Union[ImageType, str]) -> bool:
+    if type(img) is str:
+        img = Image.open(str)
+
+    if img.info.get("transparency", None) is not None:
+        return True
+    if img.mode == "P":
+        transparent = img.info.get("transparency", -1)
+        for _, index in img.getcolors():
+            if index == transparent:
+                return True
+    elif img.mode == "RGBA":
+        extrema = img.getextrema()
+        if extrema[3][0] < 255:
+            return True
+    return False
+
+
+def get_canvas_generation_mode(
+    init_img: Union[ImageType, str], init_mask: Union[ImageType, str]
+) -> Literal["txt2img", "outpainting", "inpainting", "img2img",]:
+    if type(init_img) is str:
+        init_img = Image.open(init_img)
+
+    if type(init_mask) is str:
+        init_mask = Image.open(init_mask)
+
+    init_img = init_img.convert("RGBA")
+
+    # Get alpha from init_img
+    init_img_alpha = init_img.split()[-1]
+    init_img_alpha_mask = init_img_alpha.convert("L")
+    init_img_has_transparency = check_for_any_transparency(init_img)
+
+    if init_img_has_transparency:
+        init_img_is_fully_transparent = (
+            True if init_img_alpha_mask.getbbox() is None else False
+        )
+
+    """
+    Mask images are white in areas where no change should be made, black where changes
+    should be made.
+    """
+
+    # Fit the mask to init_img's size and convert it to greyscale
+    init_mask = init_mask.resize(init_img.size).convert("L")
+
+    """
+    PIL.Image.getbbox() returns the bounding box of non-zero areas of the image, so we first
+    invert the mask image so that masked areas are white and other areas black == zero.
+    getbbox() now tells us if the are any masked areas.
+    """
+    init_mask_bbox = ImageChops.invert(init_mask).getbbox()
+    init_mask_exists = False if init_mask_bbox is None else True
+
+    if init_img_has_transparency:
+        if init_img_is_fully_transparent:
+            return "txt2img"
+        else:
+            return "outpainting"
+    else:
+        if init_mask_exists:
+            return "inpainting"
+        else:
+            return "img2img"
+
+
+def main():
+    # Testing
+    init_img_opaque = "test_images/init-img_opaque.png"
+    init_img_partial_transparency = "test_images/init-img_partial_transparency.png"
+    init_img_full_transparency = "test_images/init-img_full_transparency.png"
+    init_mask_no_mask = "test_images/init-mask_no_mask.png"
+    init_mask_has_mask = "test_images/init-mask_has_mask.png"
+
+    print(
+        "OPAQUE IMAGE, NO MASK, expect img2img, got ",
+        get_canvas_generation_mode(init_img_opaque, init_mask_no_mask),
+    )
+
+    print(
+        "IMAGE WITH TRANSPARENCY, NO MASK, expect outpainting, got ",
+        get_canvas_generation_mode(
+            init_img_partial_transparency, init_mask_no_mask
+        ),
+    )
+
+    print(
+        "FULLY TRANSPARENT IMAGE NO MASK, expect txt2img, got ",
+        get_canvas_generation_mode(init_img_full_transparency, init_mask_no_mask),
+    )
+
+    print(
+        "OPAQUE IMAGE, WITH MASK, expect inpainting, got ",
+        get_canvas_generation_mode(init_img_opaque, init_mask_has_mask),
+    )
+
+    print(
+        "IMAGE WITH TRANSPARENCY, WITH MASK, expect outpainting, got ",
+        get_canvas_generation_mode(
+            init_img_partial_transparency, init_mask_has_mask
+        ),
+    )
+
+    print(
+        "FULLY TRANSPARENT IMAGE WITH MASK, expect txt2img, got ",
+        get_canvas_generation_mode(init_img_full_transparency, init_mask_has_mask),
+    )
+
+
+if __name__ == "__main__":
+    main()

backend/modules/parameters.py

@@ -0,0 +1,71 @@
+from backend.modules.parse_seed_weights import parse_seed_weights
+import argparse
+
+SAMPLER_CHOICES = [
+    "ddim",
+    "k_dpm_2_a",
+    "k_dpm_2",
+    "k_dpmpp_2_a",
+    "k_dpmpp_2",
+    "k_euler_a",
+    "k_euler",
+    "k_heun",
+    "k_lms",
+    "plms",
+]
+
+
+def parameters_to_command(params):
+    """
+    Converts dict of parameters into a `invoke.py` REPL command.
+    """
+
+    switches = list()
+
+    if "prompt" in params:
+        switches.append(f'"{params["prompt"]}"')
+    if "steps" in params:
+        switches.append(f'-s {params["steps"]}')
+    if "seed" in params:
+        switches.append(f'-S {params["seed"]}')
+    if "width" in params:
+        switches.append(f'-W {params["width"]}')
+    if "height" in params:
+        switches.append(f'-H {params["height"]}')
+    if "cfg_scale" in params:
+        switches.append(f'-C {params["cfg_scale"]}')
+    if "sampler_name" in params:
+        switches.append(f'-A {params["sampler_name"]}')
+    if "seamless" in params and params["seamless"] == True:
+        switches.append(f"--seamless")
+    if "hires_fix" in params and params["hires_fix"] == True:
+        switches.append(f"--hires")
+    if "init_img" in params and len(params["init_img"]) > 0:
+        switches.append(f'-I {params["init_img"]}')
+    if "init_mask" in params and len(params["init_mask"]) > 0:
+        switches.append(f'-M {params["init_mask"]}')
+    if "init_color" in params and len(params["init_color"]) > 0:
+        switches.append(f'--init_color {params["init_color"]}')
+    if "strength" in params and "init_img" in params:
+        switches.append(f'-f {params["strength"]}')
+        if "fit" in params and params["fit"] == True:
+            switches.append(f"--fit")
+    if "facetool" in params:
+        switches.append(f'-ft {params["facetool"]}')
+    if "facetool_strength" in params and params["facetool_strength"]:
+        switches.append(f'-G {params["facetool_strength"]}')
+    elif "gfpgan_strength" in params and params["gfpgan_strength"]:
+        switches.append(f'-G {params["gfpgan_strength"]}')
+    if "codeformer_fidelity" in params:
+        switches.append(f'-cf {params["codeformer_fidelity"]}')
+    if "upscale" in params and params["upscale"]:
+        switches.append(f'-U {params["upscale"][0]} {params["upscale"][1]}')
+    if "variation_amount" in params and params["variation_amount"] > 0:
+        switches.append(f'-v {params["variation_amount"]}')
+        if "with_variations" in params:
+            seed_weight_pairs = ",".join(
+                f"{seed}:{weight}" for seed, weight in params["with_variations"]
+            )
+            switches.append(f"-V {seed_weight_pairs}")
+
+    return " ".join(switches)

backend/modules/parse_seed_weights.py

@@ -0,0 +1,47 @@
+def parse_seed_weights(seed_weights):
+    """
+    Accepts seed weights as string in "12345:0.1,23456:0.2,3456:0.3" format
+    Validates them
+    If valid: returns as [[12345, 0.1], [23456, 0.2], [3456, 0.3]]
+    If invalid: returns False
+    """
+
+    # Must be a string
+    if not isinstance(seed_weights, str):
+        return False
+    # String must not be empty
+    if len(seed_weights) == 0:
+        return False
+
+    pairs = []
+
+    for pair in seed_weights.split(","):
+        split_values = pair.split(":")
+
+        # Seed and weight are required
+        if len(split_values) != 2:
+            return False
+
+        if len(split_values[0]) == 0 or len(split_values[1]) == 1:
+            return False
+
+        # Try casting the seed to int and weight to float
+        try:
+            seed = int(split_values[0])
+            weight = float(split_values[1])
+        except ValueError:
+            return False
+
+        # Seed must be 0 or above
+        if not seed >= 0:
+            return False
+
+        # Weight must be between 0 and 1
+        if not (weight >= 0 and weight <= 1):
+            return False
+
+        # This pair is valid
+        pairs.append([seed, weight])
+
+    # All pairs are valid
+    return pairs

configs/INITIAL_MODELS.yaml

@@ -0,0 +1,80 @@
+stable-diffusion-1.5:
+   description: The newest Stable Diffusion version 1.5 weight file (4.27 GB)
+   repo_id: runwayml/stable-diffusion-v1-5
+   config: v1-inference.yaml
+   file: v1-5-pruned-emaonly.ckpt
+   recommended: true
+   width: 512
+   height: 512
+inpainting-1.5:
+   description: RunwayML SD 1.5 model optimized for inpainting (4.27 GB)
+   repo_id: runwayml/stable-diffusion-inpainting
+   config: v1-inpainting-inference.yaml
+   file: sd-v1-5-inpainting.ckpt
+   recommended: True
+   width: 512
+   height: 512
+ft-mse-improved-autoencoder-840000:
+   description: StabilityAI improved autoencoder fine-tuned for human faces (recommended; 335 MB)
+   repo_id: stabilityai/sd-vae-ft-mse-original
+   config: VAE/default
+   file: vae-ft-mse-840000-ema-pruned.ckpt
+   recommended: True
+   width: 512
+   height: 512
+stable-diffusion-1.4:
+   description: The original Stable Diffusion version 1.4 weight file (4.27 GB)
+   repo_id: CompVis/stable-diffusion-v-1-4-original
+   config: v1-inference.yaml
+   file: sd-v1-4.ckpt
+   recommended: False
+   width: 512
+   height: 512
+waifu-diffusion-1.3:
+   description: Stable Diffusion 1.4 fine tuned on anime-styled images (4.27 GB)
+   repo_id: hakurei/waifu-diffusion-v1-3
+   config: v1-inference.yaml
+   file: model-epoch09-float32.ckpt
+   recommended: False
+   width: 512
+   height: 512
+trinart-2.0:
+   description: An SD model finetuned with ~40,000 assorted high resolution manga/anime-style pictures (2.13 GB)
+   repo_id: naclbit/trinart_stable_diffusion_v2
+   config: v1-inference.yaml
+   file: trinart2_step95000.ckpt
+   recommended: False
+   width: 512
+   height: 512
+trinart_characters-1.0:
+   description: An SD model finetuned with 19.2M anime/manga style images (2.13 GB)
+   repo_id: naclbit/trinart_characters_19.2m_stable_diffusion_v1
+   config: v1-inference.yaml
+   file: trinart_characters_it4_v1.ckpt
+   recommended: False
+   width: 512
+   height: 512
+trinart_vae:
+   description: Custom autoencoder for trinart_characters
+   repo_id: naclbit/trinart_characters_19.2m_stable_diffusion_v1
+   config: VAE/trinart
+   file: autoencoder_fix_kl-f8-trinart_characters.ckpt
+   recommended: False
+   width: 512
+   height: 512
+papercut-1.0:
+   description: SD 1.5 fine-tuned for papercut art (use "PaperCut" in your prompts) (2.13 GB)
+   repo_id: Fictiverse/Stable_Diffusion_PaperCut_Model
+   config: v1-inference.yaml
+   file: PaperCut_v1.ckpt
+   recommended: False
+   width: 512
+   height: 512
+voxel_art-1.0:
+   description: Stable Diffusion trained on voxel art (use "VoxelArt" in your prompts) (4.27 GB)
+   repo_id: Fictiverse/Stable_Diffusion_VoxelArt_Model
+   config: v1-inference.yaml
+   file: VoxelArt_v1.ckpt
+   recommended: False
+   width: 512
+   height: 512

configs/models.yaml.example

@@ -0,0 +1,29 @@
+# This file describes the alternative machine learning models
+# available to InvokeAI script.
+#
+# To add a new model, follow the examples below. Each
+# model requires a model config file, a weights file,
+# and the width and height of the images it
+# was trained on.
+stable-diffusion-1.5:
+  description: The newest Stable Diffusion version 1.5 weight file (4.27 GB)
+  weights: models/ldm/stable-diffusion-v1/v1-5-pruned-emaonly.ckpt
+  config: configs/stable-diffusion/v1-inference.yaml
+  width: 512
+  height: 512
+  vae: ./models/ldm/stable-diffusion-v1/vae-ft-mse-840000-ema-pruned.ckpt
+  default: true
+stable-diffusion-1.4:
+  description: Stable Diffusion inference model version 1.4
+  config: configs/stable-diffusion/v1-inference.yaml
+  weights: models/ldm/stable-diffusion-v1/sd-v1-4.ckpt
+  vae: models/ldm/stable-diffusion-v1/vae-ft-mse-840000-ema-pruned.ckpt
+  width: 512
+  height: 512
+inpainting-1.5:
+  weights: models/ldm/stable-diffusion-v1/sd-v1-5-inpainting.ckpt
+  config: configs/stable-diffusion/v1-inpainting-inference.yaml
+  vae: models/ldm/stable-diffusion-v1/vae-ft-mse-840000-ema-pruned.ckpt
+  description: RunwayML SD 1.5 model optimized for inpainting
+  width: 512
+  height: 512