Update app.py

app.py

@@ -1,71 +1,71 @@
-from fastai.vision.all import *
-import gradio as gr
-from colorthief import ColorThief
-from PIL import Image
-import matplotlib.colors as mcolors
-import io
-import colorsys
-
-# Load the model
-learn = load_learner('outfit_recommender_resnet18.pkl')
-
-# Define class names
-class_names = ['jeans', 'top']
-
-# Helper function to get the name of the closest color
-def closest_color(requested_color):
-    min_colors = {}
-    for key, name in mcolors.CSS4_COLORS.items():
-        r_c, g_c, b_c = mcolors.hex2color(name)
-        rd = (r_c - requested_color[0]) ** 2
-        gd = (g_c - requested_color[1]) ** 2
-        bd = (b_c - requested_color[2]) ** 2
-        min_colors[(rd + gd + bd)] = key
-    return min_colors[min(min_colors.keys())]
-
-def get_dominant_color(image):
-    img_byte_arr = io.BytesIO()
-    image.save(img_byte_arr, format='PNG')
-    img_byte_arr = img_byte_arr.getvalue()
-    
-    color_thief = ColorThief(io.BytesIO(img_byte_arr))
-    dominant_color = color_thief.get_color(quality=1)
-    dominant_color = tuple(c / 255 for c in dominant_color)  # Normalize RGB values to [0, 1]
-    return dominant_color
-
-def get_complementary_color(rgb_color):
-    h, s, v = colorsys.rgb_to_hsv(*rgb_color)
-    complementary_h = (h + 0.5) % 1.0
-    r, g, b = colorsys.hsv_to_rgb(complementary_h, s, v)
-    return closest_color((r, g, b))
-
-def get_outfit_recommendation(pred_class):
-    if pred_class == 'top':
-        return 'Jeans'
-    elif pred_class == 'jeans':
-        return 'Top'
-    else:
-        return 'Item'
-
-def predict(image):
-    pred_class, pred_idx, outputs = learn.predict(image)
-    pred_class = class_names[pred_idx]  # Convert index to class name
-    dominant_color = get_dominant_color(image)
-    complementary_color = get_complementary_color(dominant_color)
-    garment_recommendation = get_outfit_recommendation(pred_class)
-    return f"Complementary item: {garment_recommendation} in {complementary_color}"
-
-def gradio_predict(image):
-    if isinstance(image, np.ndarray):
-        image = Image.fromarray(image.astype('uint8'), 'RGB')
-    return predict(image)
-
-interface = gr.Interface(
-    fn=gradio_predict,
-    inputs=gr.Image(),
-    outputs="text",
-    title="Outfit Recommender",
-    description="Upload an image of a jeans or a top to get a recommendation for a complementary outfit based on fashion theory."
-)
-
-interface.launch()
+from fastai.vision.all import *
+import gradio as gr
+from colorthief import ColorThief
+from PIL import Image
+import matplotlib.colors as mcolors
+import io
+import colorsys
+
+# Load the model
+learn = load_learner('outfit_recommender_resnet18.pkl')
+
+# Define class names
+class_names = ['jeans', 'top']
+
+# Helper function to get the name of the closest color
+def closest_color(requested_color):
+    min_colors = {}
+    for key, name in mcolors.CSS4_COLORS.items():
+        r_c, g_c, b_c = mcolors.hex2color(name)
+        rd = (r_c - requested_color[0]) ** 2
+        gd = (g_c - requested_color[1]) ** 2
+        bd = (b_c - requested_color[2]) ** 2
+        min_colors[(rd + gd + bd)] = key
+    return min_colors[min(min_colors.keys())]
+
+def get_dominant_color(image):
+    img_byte_arr = io.BytesIO()
+    image.save(img_byte_arr, format='PNG')
+    img_byte_arr = img_byte_arr.getvalue()
+    
+    color_thief = ColorThief(io.BytesIO(img_byte_arr))
+    dominant_color = color_thief.get_color(quality=1)
+    dominant_color = tuple(c / 255 for c in dominant_color)  # Normalize RGB values to [0, 1]
+    return dominant_color
+
+def get_complementary_color(rgb_color):
+    h, s, v = colorsys.rgb_to_hsv(*rgb_color)
+    complementary_h = (h + 0.5) % 1.0
+    r, g, b = colorsys.hsv_to_rgb(complementary_h, s, v)
+    return closest_color((r, g, b))
+
+def get_outfit_recommendation(pred_class):
+    if pred_class == 'top':
+        return 'Jeans'
+    elif pred_class == 'jeans':
+        return 'Top'
+    else:
+        return 'Item'
+
+def predict(image):
+    pred_class, _, _ = learn.predict(image)
+    pred_class = class_names[pred_idx]  # Convert index to class name
+    dominant_color = get_dominant_color(image)
+    complementary_color = get_complementary_color(dominant_color)
+    garment_recommendation = get_outfit_recommendation(pred_class)
+    return f"Complementary item: {garment_recommendation} in {complementary_color}"
+
+def gradio_predict(image):
+    if isinstance(image, np.ndarray):
+        image = Image.fromarray(image.astype('uint8'), 'RGB')
+    return predict(image)
+
+interface = gr.Interface(
+    fn=gradio_predict,
+    inputs=gr.Image(),
+    outputs="text",
+    title="Outfit Recommender",
+    description="Upload an image of a jeans or a top to get a recommendation for a complementary outfit based on fashion theory."
+)
+
+interface.launch()