Create working_app.py

working_app.py

@@ -0,0 +1,139 @@
+import gradio as gr
+import numpy as np
+# from edict_functions import EDICT_editing
+from PIL import Image
+from utils import Endpoint, get_token
+from io import BytesIO
+import requests
+
+
+endpoint = Endpoint()
+
+def local_edict(x, source_text, edit_text,
+         edit_strength, guidance_scale,
+          steps=50, mix_weight=0.93, ):
+    x = Image.fromarray(x)
+    return_im =  EDICT_editing(x,
+                         source_text,
+                         edit_text,
+                  steps=steps,
+                  mix_weight=mix_weight,
+                  init_image_strength=edit_strength,
+                  guidance_scale=guidance_scale
+                              )[0]
+    return np.array(return_im)
+
+def encode_image(image):
+    buffered = BytesIO()
+    image.save(buffered, format="JPEG", quality=95)
+    buffered.seek(0)
+
+    return buffered
+
+
+
+def decode_image(img_obj):
+    img = Image.open(img_obj).convert("RGB")
+    return img
+
+def edict(x, source_text, edit_text,
+         edit_strength, guidance_scale,
+          steps=50, mix_weight=0.93, ):
+
+    url = endpoint.url
+    url = url + "/api/edit"
+    headers = {### Misc.
+
+        "User-Agent": "EDICT HuggingFace Space",
+        "Auth-Token": get_token(),
+    }
+
+    data = {
+        "source_text": source_text,
+        "edit_text": edit_text,
+        "edit_strength": edit_strength,
+        "guidance_scale": guidance_scale,
+    }
+
+    image = encode_image(Image.fromarray(x))
+    files = {"image": image}  
+
+    response = requests.post(url, data=data, files=files, headers=headers)
+    
+    if response.status_code == 200:
+        return np.array(decode_image(BytesIO(response.content)))
+    else:
+        return "Error: " + response.text   
+    # x = decode_image(response)
+    # return np.array(x)
+
+examples = [
+        ['square_ims/american_gothic.jpg', 'A painting of two people frowning', 'A painting of two people smiling', 0.5, 3],
+        ['square_ims/colloseum.jpg', 'An old ruined building', 'A new modern office building', 0.8, 3],
+    ]
+
+
+examples.append(['square_ims/scream.jpg', 'A painting of someone screaming', 'A painting of an alien', 0.5, 3])
+examples.append(['square_ims/yosemite.jpg', 'Granite forest valley', 'Granite desert valley', 0.8, 3])
+examples.append(['square_ims/einstein.jpg', 'Mouth open', 'Mouth closed', 0.8, 3])
+examples.append(['square_ims/einstein.jpg', 'A man', 'A man in K.I.S.S. facepaint', 0.8, 3])
+"""
+examples.extend([
+        ['square_ims/imagenet_cake_2.jpg', 'A cupcake', 'A Chinese New Year cupcake', 0.8, 3],
+        ['square_ims/imagenet_cake_2.jpg', 'A cupcake', 'A Union Jack cupcake', 0.8, 3],
+        ['square_ims/imagenet_cake_2.jpg', 'A cupcake', 'A Nigerian flag cupcake', 0.8, 3],
+        ['square_ims/imagenet_cake_2.jpg', 'A cupcake', 'A Santa Claus cupcake', 0.8, 3],
+        ['square_ims/imagenet_cake_2.jpg', 'A cupcake', 'An Easter cupcake', 0.8, 3],
+        ['square_ims/imagenet_cake_2.jpg', 'A cupcake', 'A hedgehog cupcake', 0.8, 3],
+        ['square_ims/imagenet_cake_2.jpg', 'A cupcake', 'A rose cupcake', 0.8, 3],
+    ])
+"""
+
+for dog_i in [1, 2]:
+    for breed in ['Golden Retriever', 'Chihuahua', 'Dalmatian']:
+        examples.append([f'square_ims/imagenet_dog_{dog_i}.jpg', 'A dog', f'A {breed}', 0.8, 3])
+
+
+description = 'A gradio demo for [EDICT](https://arxiv.org/abs/2211.12446) (CVPR23)'
+# description = gr.Markdown(description)
+
+article = """
+
+### Prompting Style
+
+As with many text-to-image methods, the prompting style of EDICT can make a big difference. When in doubt, experiment! Some guidance:
+* Parallel *Original Description* and *Edit Description* construction as much as possible. Inserting/editing single words often is enough to affect a change while maintaining a lot of the original structure
+* Words that will affect the entire setting (e.g. "A photo of " vs. "A painting of") can make a big difference. Playing around with them can help a lot
+
+### Parameters
+Both `edit_strength` and `guidance_scale` have similar properties qualitatively: the higher the value the more the image will change. We suggest
+* Increasing/decreasing `edit_strength` first, particularly to alter/preserve more of the original structure/content
+* Then changing `guidance_scale` to make the change in the edited region more or less pronounced.
+
+Usually we find changing `edit_strength` to be enough, but feel free to play around (and report any interesting results)!
+
+### Misc.
+
+Having difficulty coming up with a caption? Try [BLIP](https://huggingface.co/spaces/Salesforce/BLIP2) to automatically generate one!
+
+As with most StableDiffusion approaches, faces/text are often problematic to render, especially if they're small. Having these in the foreground will help keep them cleaner.
+
+A returned black image means that the [Safety Checker](https://huggingface.co/CompVis/stable-diffusion-safety-checker) triggered on the photo. This happens in odd cases sometimes (it often rejects
+the huggingface logo or variations), but we need to keep it in for obvious reasons.
+"""
+# article = gr.Markdown(description)
+
+iface = gr.Interface(fn=edict, inputs=["image",
+                                       gr.Textbox(label="Original Description"),
+                                       gr.Textbox(label="Edit Description"),
+                                       # 50, # gr.Slider(5, 50, value=20, step=1),
+                                       # 0.93, # gr.Slider(0.5, 1, value=0.7, step=0.05),
+                                       gr.Slider(0.0, 1, value=0.8, step=0.05),
+                                       gr.Slider(0, 10, value=3, step=0.5),
+                                      ],
+                     examples = examples,
+                     outputs="image",
+                     description=description,
+                     article=article,
+                     cache_examples=True)
+iface.launch()