Added examples

.gitattributes

@@ -32,3 +32,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+*.png filter=lfs diff=lfs merge=lfs -text
+*.jpg filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1 @@
+.history

.history/app_20230528143707.py

@@ -0,0 +1,61 @@
+import torch
+import clip
+from PIL import Image
+import gradio as gr
+
+device = "cuda" if torch.cuda.is_available() else "cpu"
+model, preprocess = clip.load("ViT-B/32", device=device)
+
+def allure(image, gender):
+    image = Image.fromarray(image.astype("uint8"), "RGB")
+    gender = gender.lower()
+    image = preprocess(image).unsqueeze(0).to(device)
+    positive_terms = [f'a hot {gender}', f'a beautiful {gender}', f'an alluring {gender}']
+    negative_terms = [f'a gross {gender}', f'an ugly {gender}', f'a hideous {gender}']
+
+    pairs = list(zip(positive_terms, negative_terms))
+
+    def evaluate(terms):
+        text = clip.tokenize(terms).to(device)
+
+        with torch.no_grad():
+            logits_per_image, logits_per_text = model(image, text)
+            probs = logits_per_image.softmax(dim=-1).cpu().numpy()
+            return probs[0]
+
+    probs = [evaluate(pair) for pair in pairs]
+    
+    positive_probs = [prob[0] for prob in probs]
+    negative_probs = [prob[1] for prob in probs]
+
+    hotness_score = round((probs[0][0] - probs[0][1] + 1) * 50, 2)
+    beauty_score = round((probs[1][0] - probs[1][1] + 1) * 50, 2)
+    attractiveness_score = round((probs[2][0] - probs[2][1] + 1) * 50, 2)
+
+    hot_score = sum(positive_probs)/len(positive_probs)
+    ugly_score = sum(negative_probs)/len(negative_probs)
+    composite = ((hot_score - ugly_score)+1) * 50
+    composite = round(composite, 2)
+    return composite, hotness_score, beauty_score, attractiveness_score
+
+iface = gr.Interface(
+    fn=allure,
+    inputs=[
+        gr.inputs.Image(label="Image"),
+        gr.inputs.Dropdown(
+            [
+                'Person', 'Man', 'Woman'
+            ],
+            default='Person',
+        )
+    ],
+    outputs=[
+        gr.Textbox(label="Composite (%)"),
+        gr.Textbox(label="Hotness (%)"),
+        gr.Textbox(label="Beauty (%)"),
+        gr.Textbox(label="Allure (%)"),
+    ],
+    title="Attractiveness Evaluator (using OpenAI CLIP)",
+    description="A simple attractiveness evaluation app using OpenAI's CLIP model. \nHow it works: \nThe input image is passed to OpenAI's CLIP image captioning model and evaluated for how much it conforms to the model's idea of hotness, beauty, and attractiveness. These values are then combined to produce a composite score on a scale of 0 to 100.\nNote: This is meant solely for educational use.",
+)
+iface.launch()

.history/app_20230528145539.py

@@ -0,0 +1,66 @@
+import torch
+import clip
+from PIL import Image
+import gradio as gr
+
+device = "cuda" if torch.cuda.is_available() else "cpu"
+model, preprocess = clip.load("ViT-B/32", device=device)
+
+def allure(image, gender):
+    image = Image.fromarray(image.astype("uint8"), "RGB")
+    gender = gender.lower()
+    image = preprocess(image).unsqueeze(0).to(device)
+    positive_terms = [f'a hot {gender}', f'a beautiful {gender}', f'an alluring {gender}']
+    negative_terms = [f'a gross {gender}', f'an ugly {gender}', f'a hideous {gender}']
+
+    pairs = list(zip(positive_terms, negative_terms))
+
+    def evaluate(terms):
+        text = clip.tokenize(terms).to(device)
+
+        with torch.no_grad():
+            logits_per_image, logits_per_text = model(image, text)
+            probs = logits_per_image.softmax(dim=-1).cpu().numpy()
+            return probs[0]
+
+    probs = [evaluate(pair) for pair in pairs]
+    
+    positive_probs = [prob[0] for prob in probs]
+    negative_probs = [prob[1] for prob in probs]
+
+    hotness_score = round((probs[0][0] - probs[0][1] + 1) * 50, 2)
+    beauty_score = round((probs[1][0] - probs[1][1] + 1) * 50, 2)
+    attractiveness_score = round((probs[2][0] - probs[2][1] + 1) * 50, 2)
+
+    hot_score = sum(positive_probs)/len(positive_probs)
+    ugly_score = sum(negative_probs)/len(negative_probs)
+    composite = ((hot_score - ugly_score)+1) * 50
+    composite = round(composite, 2)
+    return composite, hotness_score, beauty_score, attractiveness_score
+
+iface = gr.Interface(
+    fn=allure,
+    inputs=[
+        gr.inputs.Image(label="Image"),
+        gr.inputs.Dropdown(
+            [
+                'Person', 'Man', 'Woman'
+            ],
+            default='Person',
+            label="Gender"
+        )
+    ],
+    outputs=[
+        gr.Textbox(label="Composite (%)"),
+        gr.Textbox(label="Hotness (%)"),
+        gr.Textbox(label="Beauty (%)"),
+        gr.Textbox(label="Allure (%)"),
+    ],
+    examples = [
+        ['Mansib_01_x2048.png', 'Man'],
+        ['Mansib_02_x2048.png', 'Man']
+    ],
+    title="Attractiveness Evaluator (using OpenAI CLIP)",
+    description="A simple attractiveness evaluation app using OpenAI's CLIP model. \nHow it works: \nThe input image is passed to OpenAI's CLIP image captioning model and evaluated for how much it conforms to the model's idea of hotness, beauty, and attractiveness. These values are then combined to produce a composite score on a scale of 0 to 100.\nNote: This is meant solely for educational use.",
+)
+iface.launch()

.history/app_20230528145615.py

@@ -0,0 +1,66 @@
+import torch
+import clip
+from PIL import Image
+import gradio as gr
+
+device = "cuda" if torch.cuda.is_available() else "cpu"
+model, preprocess = clip.load("ViT-B/32", device=device)
+
+def allure(image, gender):
+    image = Image.fromarray(image.astype("uint8"), "RGB")
+    gender = gender.lower()
+    image = preprocess(image).unsqueeze(0).to(device)
+    positive_terms = [f'a hot {gender}', f'a beautiful {gender}', f'an alluring {gender}']
+    negative_terms = [f'a gross {gender}', f'an ugly {gender}', f'a hideous {gender}']
+
+    pairs = list(zip(positive_terms, negative_terms))
+
+    def evaluate(terms):
+        text = clip.tokenize(terms).to(device)
+
+        with torch.no_grad():
+            logits_per_image, logits_per_text = model(image, text)
+            probs = logits_per_image.softmax(dim=-1).cpu().numpy()
+            return probs[0]
+
+    probs = [evaluate(pair) for pair in pairs]
+    
+    positive_probs = [prob[0] for prob in probs]
+    negative_probs = [prob[1] for prob in probs]
+
+    hotness_score = round((probs[0][0] - probs[0][1] + 1) * 50, 2)
+    beauty_score = round((probs[1][0] - probs[1][1] + 1) * 50, 2)
+    attractiveness_score = round((probs[2][0] - probs[2][1] + 1) * 50, 2)
+
+    hot_score = sum(positive_probs)/len(positive_probs)
+    ugly_score = sum(negative_probs)/len(negative_probs)
+    composite = ((hot_score - ugly_score)+1) * 50
+    composite = round(composite, 2)
+    return composite, hotness_score, beauty_score, attractiveness_score
+
+iface = gr.Interface(
+    fn=allure,
+    inputs=[
+        gr.inputs.Image(label="Image"),
+        gr.inputs.Dropdown(
+            [
+                'Person', 'Man', 'Woman'
+            ],
+            default='Person',
+            label="Gender"
+        )
+    ],
+    outputs=[
+        gr.Textbox(label="Composite (%)"),
+        gr.Textbox(label="Hotness (%)"),
+        gr.Textbox(label="Beauty (%)"),
+        gr.Textbox(label="Allure (%)"),
+    ],
+    examples = [
+        ['Mansib_01_x2048.png', 'Man'],
+        ['Mansib_02_x2048.png', 'Man']
+    ],
+    title="Attractiveness Evaluator (powered by OpenAI CLIP)",
+    description="A simple attractiveness evaluation app using OpenAI's CLIP model. \nHow it works: \nThe input image is passed to OpenAI's CLIP image captioning model and evaluated for how much it conforms to the model's idea of hotness, beauty, and attractiveness. These values are then combined to produce a composite score on a scale of 0 to 100.\nNote: This is meant solely for educational use.",
+)
+iface.launch()

app.py

@@ -47,6 +47,7 @@ iface = gr.Interface(
                 'Person', 'Man', 'Woman'
             ],
             default='Person',
+            label="Gender"
         )
     ],
     outputs=[
@@ -55,7 +56,11 @@ iface = gr.Interface(
         gr.Textbox(label="Beauty (%)"),
         gr.Textbox(label="Allure (%)"),
     ],
-    title="Attractiveness Evaluator (using OpenAI CLIP)",
+    examples = [
+        ['Mansib_01_x2048.png', 'Man'],
+        ['Mansib_02_x2048.png', 'Man']
+    ],
+    title="Attractiveness Evaluator (powered by OpenAI CLIP)",
     description="A simple attractiveness evaluation app using OpenAI's CLIP model. \nHow it works: \nThe input image is passed to OpenAI's CLIP image captioning model and evaluated for how much it conforms to the model's idea of hotness, beauty, and attractiveness. These values are then combined to produce a composite score on a scale of 0 to 100.\nNote: This is meant solely for educational use.",
 )
 iface.launch()