app.py

import spaces
import gradio as gr
import gradio_client.utils as _gc_utils

_original_get_type = _gc_utils.get_type
_original_json_schema = _gc_utils._json_schema_to_python_type

def _safe_get_type(schema):
    if not isinstance(schema, dict):
        return "Any"
    return _original_get_type(schema)

def _safe_json_schema_to_python_type(schema, defs=None):
    if isinstance(schema, bool):
        return "Any"
    return _original_json_schema(schema, defs)

_gc_utils.get_type = _safe_get_type
_gc_utils._json_schema_to_python_type = _safe_json_schema_to_python_type

from PIL import Image
from transparent_background import Remover
import numpy as np

# Lazy-init on the GPU worker
remover = None


def _get_mask(img: Image.Image) -> Image.Image:
    """Run the model and return a clean grayscale mask, regardless of
    whether the library hands us a PIL image, float numpy array, or uint8 array."""
    out = remover.process(img, type='map')
    if isinstance(out, Image.Image):
        return out.convert('L')
    arr = np.asarray(out)
    if arr.dtype != np.uint8:
        arr = (arr * 255).clip(0, 255).astype(np.uint8)
    if arr.ndim == 3:
        arr = arr[..., 0]
    return Image.fromarray(arr, mode='L')


@spaces.GPU
def process_image(input_image, output_type):
    global remover
    if remover is None:
        remover = Remover(jit=False)

    input_image = input_image.convert('RGB')
    mask = _get_mask(input_image)

    if output_type == "Mask only":
        return mask

    # Compose RGBA ourselves: original pixels + our mask as alpha.
    rgba = input_image.convert('RGBA')
    rgba.putalpha(mask)
    return rgba


description = """<h1 align="center">InSPyReNet Background Remover</h1>
<p><center>
<a href="https://github.com/plemeri/InSPyReNet" target="_blank">[Github]</a>
<a href="https://dualview.ai" target="_blank">[Compare Results]</a>
</center></p>
"""

iface = gr.Interface(
    fn=process_image,
    inputs=[
        gr.Image(type="pil", label="Input Image", height=512),
        gr.Radio(["Default", "Mask only"], label="Output Type", value="Default")
    ],
    outputs=gr.Image(
        type="pil",
        label="Output Image",
        height=512,
        image_mode="RGBA",
        format="png",
    ),
    description=description,
    theme='bethecloud/storj_theme',
    examples=[
        ["1.png", "Default"],
        ["2.png", "Default"],
        ["3.jfif", "Default"],
        ["4.webp", "Default"]
    ],
    cache_examples=True
)

if __name__ == "__main__":
    iface.launch()