add initial code

.gitignore

@@ -0,0 +1,5 @@
+/access_token.txt
+/docs/
+/docs/*
+/.idea/
+/.idea/*

README.md

@@ -11,3 +11,15 @@ license: mit
 ---
 
 Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+
+
+## Room Decoration
+Tools for Room Decoration with AI
+ 
+## Status
+
+Currently under development
+
+## Cite this Project
+
+foxelas, Tools for Room Decoration with AI, 2023

app.py

@@ -0,0 +1,58 @@
+#from huggingface_hub import login
+#login()
+access_token = ""
+with open("access_token.txt", "r") as f:
+    access_token = f.read()
+
+import torch
+from diffusers import StableDiffusionImg2ImgPipeline
+model_path = "CompVis/stable-diffusion-v1-4"
+import requests
+from io import BytesIO
+from PIL import Image
+
+run_on_cpu = not(torch.cuda.is_available())
+if run_on_cpu:
+    device = torch.device("cpu")
+    pipe = StableDiffusionImg2ImgPipeline.from_pretrained(
+        model_path,
+        use_auth_token=access_token
+    )
+    pipe = pipe.to(device)
+    pipe.enable_attention_slicing()
+
+else:
+    device = torch.device("cuda")
+    pipe = StableDiffusionImg2ImgPipeline.from_pretrained(
+        model_path,
+        revision="fp16",
+        torch_dtype=torch.float16,
+        use_auth_token=access_token
+    )
+    pipe = pipe.to(device)
+    #pipe.enable_attention_slicing()
+
+#######################################################
+## Input and Parameters
+target_url = "https://t4.ftcdn.net/jpg/03/46/82/87/360_F_346828778_LM1vQWt1RLsrlHZMeNYeHnBBP90APQab.jpg"
+prompt = "A living room decorated in boho style with lots of furniture."
+strength = 0.75
+guidance_scale = 7.5
+
+#######################################################
+response = requests.get(target_url)
+init_img = Image.open(BytesIO(response.content)).convert("RGB")
+init_img = init_img.resize((768, 512))
+init_img.save("input.jpg")
+print("Saved input image as [input.jpg] in local directory.")
+
+generator = torch.Generator(device=device).manual_seed(1024)
+if run_on_cpu:
+    image = pipe(prompt=prompt, init_image=init_img, strength=strength, guidance_scale=guidance_scale, generator=generator).images[0]
+else:
+    with torch.autocast("cuda"):
+        image = pipe(prompt=prompt, init_image=init_img, strength=strength, guidance_scale=guidance_scale, generator=generator).images[0]
+
+image.save("output.jpg")
+
+print("Saved output image as [output.jpg] in local directory.")

requirements.txt

@@ -0,0 +1,5 @@
+gradio
+torch~=1.13.1
+requests~=2.28.1
+diffusers~=0.8.0
+Pillow~=9.4.0