Upload 4 files

app (2).py

@@ -0,0 +1,53 @@
+#Script added by SPDraptor
+import spaces
+from typing import Optional
+import subprocess
+subprocess.run('pip install flash-attn --no-build-isolation', env={'FLASH_ATTENTION_SKIP_CUDA_BUILD': "TRUE"}, shell=True)
+import torch
+print("cuda present = ",torch.cuda.is_available())
+import os
+import sys
+import gradio as gr
+from PIL import Image
+
+from modules import masking_module
+from modules import controlnetSD
+# import modules.inpaint.utils_inpaint
+
+
+
+DESCRIPTION = "Welcome to Raptor APIs"
+
+css = """
+  #output {
+    height: 500px; 
+    overflow: auto; 
+    border: 1px solid #ccc; 
+  }
+"""
+
+with gr.Blocks(css=css) as demo:
+    gr.Markdown(DESCRIPTION)
+    with gr.Tab(label="OBJ_mask"):
+        with gr.Row():
+            with gr.Column():
+                image = gr.Image(label="Input main Picture")
+                image_object = gr.Textbox(label="object name")
+                mask_btn = gr.Button(value="createMask")
+                
+            with gr.Column():
+                output_mask = gr.Image(label="mask")
+                mask_btn.click(masking_module.masking_process,inputs=[image,image_object],outputs=output_mask,api_name="masking_step")
+    with gr.Tab(label="img_inpaint"):
+        with gr.Row():
+            with gr.Column():
+                org_image = gr.Image(label="Input main Picture")
+                mask = gr.Image(label="Input mask Picture")
+                prompt = gr.Textbox(label="prompt")
+                mask_btn = gr.Button(value="replace")
+                
+            with gr.Column():
+                output_img = gr.Image(label="edited_img")
+                mask_btn.click(controlnetSD.mask_based_updating2,inputs=[org_image,mask,prompt],outputs=output_img,api_name="/masked_based_img_editing")
+
+demo.launch()

modules_controlnetSD.py

@@ -0,0 +1,42 @@
+from utils_inpaint import resize_image_dimensions, make_inpaint_condition
+import torch
+from diffusers import ControlNetModel, StableDiffusionControlNetInpaintPipeline
+import spaces
+import time
+from PIL import Image
+import numpy as np
+
+device = torch.device('cuda')
+
+@spaces.GPU(duration=20)
+def mask_based_updating2(init_image_file,mask_image_file,prompt,strength=0.9, guidance_scale=9, num_inference_steps=100):
+    # load ControlNet
+    start_time = time.time()
+    controlnet = ControlNetModel.from_pretrained("lllyasviel/control_v11p_sd15_inpaint")
+
+    # pass ControlNet to the pipeline
+    pipeline = StableDiffusionControlNetInpaintPipeline.from_pretrained(
+        "fluently/Fluently-v4-inpainting", controlnet=controlnet
+    )
+    # pipeline.enable_model_cpu_offload()
+    pipeline.to(device)
+    
+    init_image = Image.fromarray(init_image_file)
+    mask_image = Image.fromarray(mask_image_file)
+    init_image = init_image.convert("RGB")
+    mask_image = mask_image.convert("1")
+    width, height = init_image.size
+    width_new, height_new = resize_image_dimensions(original_resolution_wh=init_image.size)
+    init_image = init_image.resize((width_new, height_new), Image.LANCZOS)
+    mask_image = mask_image.resize((width_new, height_new), Image.NEAREST)
+    #image and mask_image should be PIL images.
+    #The mask structure is white for inpainting and black for keeping as is
+    # image = pipe(prompt=prompt, image=init_image, mask_image=mask_image).images[0]
+    control_image = make_inpaint_condition(init_image, mask_image)
+    print("para: ",strength, guidance_scale,num_inference_steps)
+    negative_prompt = "ugly, deformed, nsfw, disfigured, worst quality, normal quality, low quality, low res, blurry, text, watermark, logo, banner, extra digits, cropped, jpeg artifacts, signature, username, error, sketch, duplicate, ugly, monochrome, horror, geometry, mutation, disgusting, bad anatomy, faint, unrealistic, Cartoon, drawing"
+    image = pipeline(prompt=prompt,negative_prompt=negative_prompt, image=init_image, mask_image=mask_image, control_image=control_image,strength = strength, guidance_scale=guidance_scale,num_inference_steps=num_inference_steps).images[0]
+    image = image.resize((width, height), Image.LANCZOS)
+    print(f'Time taken by inpainting model: {time.time() - start_time}')
+    torch.cuda.empty_cache()
+    return image

requirements (1).txt

@@ -0,0 +1,12 @@
+tqdm
+einops
+accelerate
+spaces
+timm
+transformers
+diffusers
+samv2
+gradio
+supervision
+opencv-python
+pytest

utils_inpaint.py

@@ -0,0 +1,35 @@
+from typing import Tuple
+
+def resize_image_dimensions(
+    original_resolution_wh: Tuple[int, int],
+    maximum_dimension: int = 2048
+) -> Tuple[int, int]:
+    width, height = original_resolution_wh
+
+    if width <= maximum_dimension and height <= maximum_dimension:
+        width = width - (width % 32)
+        height = height - (height % 32)
+        return width, height
+
+    if width > height:
+        scaling_factor = maximum_dimension / width
+    else:
+        scaling_factor = maximum_dimension / height
+
+    new_width = int(width * scaling_factor)
+    new_height = int(height * scaling_factor)
+
+    new_width = new_width - (new_width % 32)
+    new_height = new_height - (new_height % 32)
+
+    return new_width, new_height
+
+def make_inpaint_condition(init_image, mask_image):
+    init_image = np.array(init_image.convert("RGB")).astype(np.float32) / 255.0
+    mask_image = np.array(mask_image.convert("L")).astype(np.float32) / 255.0
+
+    assert init_image.shape[0:1] == mask_image.shape[0:1], "image and image_mask must have the same image size"
+    init_image[mask_image > 0.5] = -1.0  # set as masked pixel
+    init_image = np.expand_dims(init_image, 0).transpose(0, 3, 1, 2)
+    init_image = torch.from_numpy(init_image)
+    return init_image