Update README.md

README.md

@@ -4,6 +4,10 @@ datasets:
 - TNILab/DomainNet_FL
 ---
 
+# **Multisource-121-DomainNet**
+
+> **Multisource-121-DomainNet** is an image classification vision-language encoder model fine-tuned from **google/siglip2-base-patch16-224** for a single-label classification task. It is designed to classify images into 121 domain categories using the **SiglipForImageClassification** architecture.
+
 ![- visual selection(2).png](https://cdn-uploads.huggingface.co/production/uploads/65bb837dbfb878f46c77de4c/yfp_IYXqDyZfgZsJQ-7Bo.png)
 
 ```py
@@ -135,4 +139,221 @@ The_Great_Wall_of_China     0.8893    0.8333    0.8604       270
                accuracy                         0.7995     32670
               macro avg     0.8052    0.7995    0.8006     32670
            weighted avg     0.8052    0.7995    0.8006     32670
-```
+```
+
+
+
+
+The model categorizes images into the following 121 classes:
+- **Class 0:** "barn"  
+- **Class 1:** "baseball_bat"  
+- **Class 2:** "basket"  
+- **Class 3:** "beach"  
+- **Class 4:** "bear"  
+- **Class 5:** "beard"  
+- **Class 6:** "bee"  
+- **Class 7:** "bird"  
+- **Class 8:** "blueberry"  
+- **Class 9:** "bowtie"  
+- **Class 10:** "bracelet"  
+- **Class 11:** "brain"  
+- **Class 12:** "bread"  
+- **Class 13:** "broccoli"  
+- **Class 14:** "bus"  
+- **Class 15:** "butterfly"  
+- **Class 16:** "circle"  
+- **Class 17:** "cloud"  
+- **Class 18:** "cruise_ship"  
+- **Class 19:** "dolphin"  
+- **Class 20:** "dumbbell"  
+- **Class 21:** "elephant"  
+- **Class 22:** "eye"  
+- **Class 23:** "eyeglasses"  
+- **Class 24:** "feather"  
+- **Class 25:** "fish"  
+- **Class 26:** "flower"  
+- **Class 27:** "foot"  
+- **Class 28:** "frog"  
+- **Class 29:** "giraffe"  
+- **Class 30:** "goatee"  
+- **Class 31:** "golf_club"  
+- **Class 32:** "grapes"  
+- **Class 33:** "grass"  
+- **Class 34:** "guitar"  
+- **Class 35:** "hamburger"  
+- **Class 36:** "hand"  
+- **Class 37:** "hat"  
+- **Class 38:** "headphones"  
+- **Class 39:** "helicopter"  
+- **Class 40:** "hexagon"  
+- **Class 41:** "hockey_stick"  
+- **Class 42:** "horse"  
+- **Class 43:** "hourglass"  
+- **Class 44:** "house"  
+- **Class 45:** "ice_cream"  
+- **Class 46:** "jacket"  
+- **Class 47:** "ladder"  
+- **Class 48:** "leg"  
+- **Class 49:** "lipstick"  
+- **Class 50:** "megaphone"  
+- **Class 51:** "monkey"  
+- **Class 52:** "moon"  
+- **Class 53:** "mushroom"  
+- **Class 54:** "necklace"  
+- **Class 55:** "owl"  
+- **Class 56:** "panda"  
+- **Class 57:** "pear"  
+- **Class 58:** "peas"  
+- **Class 59:** "penguin"  
+- **Class 60:** "pig"  
+- **Class 61:** "pillow"  
+- **Class 62:** "pineapple"  
+- **Class 63:** "pizza"  
+- **Class 64:** "pool"  
+- **Class 65:** "popsicle"  
+- **Class 66:** "rabbit"  
+- **Class 67:** "rhinoceros"  
+- **Class 68:** "rifle"  
+- **Class 69:** "river"  
+- **Class 70:** "sailboat"  
+- **Class 71:** "sandwich"  
+- **Class 72:** "sea_turtle"  
+- **Class 73:** "shark"  
+- **Class 74:** "shoe"  
+- **Class 75:** "skyscraper"  
+- **Class 76:** "snorkel"  
+- **Class 77:** "snowman"  
+- **Class 78:** "soccer_ball"  
+- **Class 79:** "speedboat"  
+- **Class 80:** "spider"  
+- **Class 81:** "spoon"  
+- **Class 82:** "square"  
+- **Class 83:** "squirrel"  
+- **Class 84:** "stethoscope"  
+- **Class 85:** "strawberry"  
+- **Class 86:** "streetlight"  
+- **Class 87:** "submarine"  
+- **Class 88:** "suitcase"  
+- **Class 89:** "sun"  
+- **Class 90:** "sweater"  
+- **Class 91:** "sword"  
+- **Class 92:** "table"  
+- **Class 93:** "teapot"  
+- **Class 94:** "teddy-bear"  
+- **Class 95:** "telephone"  
+- **Class 96:** "tent"  
+- **Class 97:** "The_Eiffel_Tower"  
+- **Class 98:** "The_Great_Wall_of_China"  
+- **Class 99:** "The_Mona_Lisa"  
+- **Class 100:** "tiger"  
+- **Class 101:** "toaster"  
+- **Class 102:** "tooth"  
+- **Class 103:** "tornado"  
+- **Class 104:** "tractor"  
+- **Class 105:** "train"  
+- **Class 106:** "tree"  
+- **Class 107:** "triangle"  
+- **Class 108:** "trombone"  
+- **Class 109:** "truck"  
+- **Class 110:** "trumpet"  
+- **Class 111:** "umbrella"  
+- **Class 112:** "vase"  
+- **Class 113:** "violin"  
+- **Class 114:** "watermelon"  
+- **Class 115:** "whale"  
+- **Class 116:** "windmill"  
+- **Class 117:** "wine_glass"  
+- **Class 118:** "yoga"  
+- **Class 119:** "zebra"  
+- **Class 120:** "zigzag"
+
+# **Run with Transformers🤗**
+
+```python
+!pip install -q transformers torch pillow gradio
+```
+
+```python
+import gradio as gr
+from transformers import AutoImageProcessor, SiglipForImageClassification
+from transformers.image_utils import load_image
+from PIL import Image
+import torch
+
+# Load model and processor
+model_name = "prithivMLmods/Multisource-121-DomainNet"
+model = SiglipForImageClassification.from_pretrained(model_name)
+processor = AutoImageProcessor.from_pretrained(model_name)
+
+def multisource_classification(image):
+    """Predicts the domain category for an input image."""
+    # Convert the input numpy array to a PIL Image and ensure it is in RGB format
+    image = Image.fromarray(image).convert("RGB")
+    
+    # Process the image and convert it to model inputs
+    inputs = processor(images=image, return_tensors="pt")
+    
+    # Get model predictions without gradient calculations
+    with torch.no_grad():
+        outputs = model(**inputs)
+        logits = outputs.logits
+        # Convert logits to probabilities using softmax
+        probs = torch.nn.functional.softmax(logits, dim=1).squeeze().tolist()
+    
+    # Mapping from class indices to domain labels
+    labels = {
+        "0": "barn", "1": "baseball_bat", "2": "basket", "3": "beach", "4": "bear",
+        "5": "beard", "6": "bee", "7": "bird", "8": "blueberry", "9": "bowtie",
+        "10": "bracelet", "11": "brain", "12": "bread", "13": "broccoli", "14": "bus",
+        "15": "butterfly", "16": "circle", "17": "cloud", "18": "cruise_ship", "19": "dolphin",
+        "20": "dumbbell", "21": "elephant", "22": "eye", "23": "eyeglasses", "24": "feather",
+        "25": "fish", "26": "flower", "27": "foot", "28": "frog", "29": "giraffe",
+        "30": "goatee", "31": "golf_club", "32": "grapes", "33": "grass", "34": "guitar",
+        "35": "hamburger", "36": "hand", "37": "hat", "38": "headphones", "39": "helicopter",
+        "40": "hexagon", "41": "hockey_stick", "42": "horse", "43": "hourglass", "44": "house",
+        "45": "ice_cream", "46": "jacket", "47": "ladder", "48": "leg", "49": "lipstick",
+        "50": "megaphone", "51": "monkey", "52": "moon", "53": "mushroom", "54": "necklace",
+        "55": "owl", "56": "panda", "57": "pear", "58": "peas", "59": "penguin",
+        "60": "pig", "61": "pillow", "62": "pineapple", "63": "pizza", "64": "pool",
+        "65": "popsicle", "66": "rabbit", "67": "rhinoceros", "68": "rifle", "69": "river",
+        "70": "sailboat", "71": "sandwich", "72": "sea_turtle", "73": "shark", "74": "shoe",
+        "75": "skyscraper", "76": "snorkel", "77": "snowman", "78": "soccer_ball", "79": "speedboat",
+        "80": "spider", "81": "spoon", "82": "square", "83": "squirrel", "84": "stethoscope",
+        "85": "strawberry", "86": "streetlight", "87": "submarine", "88": "suitcase", "89": "sun",
+        "90": "sweater", "91": "sword", "92": "table", "93": "teapot", "94": "teddy-bear",
+        "95": "telephone", "96": "tent", "97": "The_Eiffel_Tower", "98": "The_Great_Wall_of_China",
+        "99": "The_Mona_Lisa", "100": "tiger", "101": "toaster", "102": "tooth", "103": "tornado",
+        "104": "tractor", "105": "train", "106": "tree", "107": "triangle", "108": "trombone",
+        "109": "truck", "110": "trumpet", "111": "umbrella", "112": "vase", "113": "violin",
+        "114": "watermelon", "115": "whale", "116": "windmill", "117": "wine_glass", "118": "yoga",
+        "119": "zebra", "120": "zigzag"
+    }
+    
+    # Create a dictionary mapping each label to its corresponding probability (rounded)
+    predictions = {labels[str(i)]: round(probs[i], 3) for i in range(len(probs))}
+    return predictions
+
+# Create Gradio interface
+iface = gr.Interface(
+    fn=multisource_classification,
+    inputs=gr.Image(type="numpy"),
+    outputs=gr.Label(label="Prediction Scores"),
+    title="Multisource-121-DomainNet Classification",
+    description="Upload an image to classify it into one of 121 domain categories."
+)
+
+# Launch the app
+if __name__ == "__main__":
+    iface.launch()
+```
+
+---
+
+# **Intended Use:**
+
+The **Multisource-121-DomainNet** model is designed for multi-source image classification. It can categorize images into a diverse set of 121 domains, covering various objects, scenes, and landmarks. Potential use cases include:
+
+- **Cross-Domain Image Analysis:** Enabling robust classification across a wide range of visual domains.
+- **Multimedia Retrieval:** Assisting in content organization and retrieval in multimedia databases.
+- **Computer Vision Research:** Serving as a benchmark for evaluating domain adaptation and transfer learning techniques.
+- **Interactive Applications:** Enhancing user interfaces with diverse, real-time image recognition capabilities.