Upload folder using huggingface_hub

README.md

@@ -1,12 +1,6 @@
 ---
-title: Mtr Sd
-emoji: 🚀
-colorFrom: red
-colorTo: green
+title: mtr-sd
+app_file: server.py
 sdk: gradio
-sdk_version: 3.39.0
-app_file: app.py
-pinned: false
+sdk_version: 3.35.2
 ---
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

sd.py

@@ -0,0 +1,79 @@
+import gradio as gr
+import os, sys
+
+import argparse
+import copy
+
+from IPython.display import display
+from PIL import Image, ImageDraw, ImageFont
+from torchvision.ops import box_convert
+
+import supervision as sv
+
+# segment anything
+from segment_anything import build_sam, SamPredictor 
+import cv2
+import numpy as np
+import matplotlib.pyplot as plt
+
+# diffusers
+import PIL
+import requests
+import torch
+from io import BytesIO
+from diffusers import StableDiffusionInpaintPipeline
+
+from huggingface_hub import hf_hub_download
+
+# load models
+device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+
+# stable diffusion (inpainting)
+sd_pipe = StableDiffusionInpaintPipeline.from_pretrained(
+    "stabilityai/stable-diffusion-2-inpainting",
+    torch_dtype=torch.float16,
+).to(device)
+
+def generate_image(image, mask, prompt, negative_prompt, pipe, seed):
+    # resize for inpainting
+    w, h = image.size
+    in_image = image.resize((512, 512))
+    in_mask = mask.resize((512, 512))
+
+    generator = torch.Generator(device).manual_seed(seed) 
+
+    result = pipe(image=in_image, mask_image=in_mask, prompt=prompt, negative_prompt=negative_prompt, generator=generator)
+    result = result.images[0]
+
+    return result.resize((w, h))
+
+prompt="perfect skin"
+negative_prompt=""
+seed = 7 # for reproducibility
+def predict(inputs):
+    # load image 
+    image, mask = inputs["image"], inputs["mask"]
+
+    # convert
+    image_source_pil = Image.fromarray(image)
+    image_mask_pil = Image.fromarray(mask)
+    
+    # inference
+    generated_image = generate_image(image=image_source_pil, mask=image_mask_pil, prompt=prompt, negative_prompt=negative_prompt, pipe=sd_pipe, seed=seed)
+    return generated_image
+
+# gradio interface
+demo = gr.Interface(fn=predict, 
+                    inputs=gr.Image(source="upload",
+                                    # interactive=True,
+                                    height=512,
+                                    tool="sketch",
+                                    type="numpy"),
+                    outputs=gr.Image(),
+                    title="Perfect Skin",
+                    article="<p style='text-align: center'>Perfect Skin | Demo</p>",
+                    allow_flagging="never",
+)
+
+if __name__ == "__main__":
+    demo.launch(server_name="0.0.0.0") #share=True

server.py

@@ -0,0 +1,63 @@
+import gradio as gr
+import os, sys
+
+import argparse
+import copy
+
+from IPython.display import display
+from PIL import Image, ImageDraw, ImageFont
+from torchvision.ops import box_convert
+
+import supervision as sv
+
+# segment anything
+from segment_anything import build_sam, SamPredictor 
+import cv2
+import numpy as np
+import matplotlib.pyplot as plt
+
+# diffusers
+import PIL
+import requests
+import torch
+from io import BytesIO
+from diffusers import StableDiffusionInpaintPipeline
+
+from huggingface_hub import hf_hub_download
+
+# load models
+device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+
+# stable diffusion (inpainting)
+sd_pipe = StableDiffusionInpaintPipeline.from_pretrained(
+    "stabilityai/stable-diffusion-2-inpainting",
+    torch_dtype=torch.float16,
+).to(device)
+
+def generate_image(image, mask, prompt, negative_prompt, pipe, seed):
+    # resize for inpainting
+    w, h = image.size
+    in_image = image.resize((512, 512))
+    in_mask = mask.resize((512, 512))
+
+    generator = torch.Generator(device).manual_seed(seed) 
+
+    result = pipe(image=in_image, mask_image=in_mask, prompt=prompt, negative_prompt=negative_prompt, generator=generator)
+    result = result.images[0]
+
+    return result.resize((w, h))
+
+prompt="perfect skin"
+negative_prompt=""
+seed = 7 # for reproducibility
+def predict(image, mask):
+    # convert
+    image_source_pil = Image.fromarray(image)
+    image_mask_pil = Image.fromarray(mask)
+    
+    # inference
+    generated_image = generate_image(image=image_source_pil, mask=image_mask_pil, prompt=prompt, negative_prompt=negative_prompt, pipe=sd_pipe, seed=seed)
+    return generated_image
+
+if __name__ == "__main__":
+    io = gr.Interface(predict, ["image", "image"], "image").launch(server_name="0.0.0.0", share=True)