File size: 6,547 Bytes
523e955 6c176ee 523e955 8f26cbd 24c17fd e597514 24c17fd 523e955 e597514 24c17fd e597514 24c17fd e597514 24c17fd e597514 24c17fd e597514 24c17fd e597514 24c17fd e597514 24c17fd e597514 24c17fd | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 | ---
license: apache-2.0
language:
- en
base_model:
- google/siglip2-base-patch16-224
pipeline_tag: image-classification
library_name: transformers
tags:
- color
- cloth
---
# **Fashion-Product-baseColour**
> **Fashion-Product-baseColour** is a visual classification model fine-tuned from **google/siglip2-base-patch16-224** using the **SiglipForImageClassification** architecture. It predicts the **base color** of fashion products from images — enabling accurate tagging, search, and recommendation in fashion-related applications.
```py
Classification Report:
precision recall f1-score support
Beige 0.4338 0.5409 0.4815 745
Black 0.8051 0.8656 0.8342 9699
Blue 0.7513 0.7858 0.7682 4906
Bronze 0.0000 0.0000 0.0000 89
Brown 0.6812 0.7596 0.7183 3440
Burgundy 0.0000 0.0000 0.0000 44
Charcoal 0.4941 0.1842 0.2684 228
Coffee Brown 0.0000 0.0000 0.0000 29
Copper 0.5000 0.0120 0.0235 83
Cream 0.3940 0.3446 0.3677 383
Fluorescent Green 0.0000 0.0000 0.0000 5
Gold 0.4935 0.6747 0.5701 621
Green 0.7286 0.7760 0.7516 2103
Grey 0.6313 0.5002 0.5581 2735
Grey Melange 0.5728 0.4041 0.4739 146
Khaki 0.3540 0.2878 0.3175 139
Lavender 0.5049 0.3250 0.3954 160
Lime Green 0.0000 0.0000 0.0000 5
Magenta 0.5909 0.1016 0.1733 128
Maroon 0.5121 0.2929 0.3727 577
Mauve 0.0000 0.0000 0.0000 28
Metallic 0.0000 0.0000 0.0000 41
Multi 0.4005 0.3832 0.3917 394
Mushroom Brown 0.0000 0.0000 0.0000 16
Mustard 0.4912 0.2887 0.3636 97
Navy Blue 0.6290 0.4905 0.5512 1784
Nude 0.0000 0.0000 0.0000 21
Off White 0.5789 0.2418 0.3411 182
Olive 0.5259 0.5208 0.5233 409
Orange 0.6838 0.6119 0.6458 523
Peach 0.4727 0.4216 0.4457 185
Pink 0.6912 0.7423 0.7158 1824
Purple 0.6846 0.7568 0.7189 1612
Red 0.6916 0.8273 0.7534 2432
Rose 0.0000 0.0000 0.0000 21
Rust 0.5000 0.1692 0.2529 65
Sea Green 0.0000 0.0000 0.0000 22
Silver 0.6088 0.4830 0.5387 1089
Skin 0.5479 0.6319 0.5869 163
Steel 0.2857 0.0381 0.0672 315
Tan 0.6667 0.0357 0.0678 112
Taupe 0.0000 0.0000 0.0000 11
Teal 0.4857 0.2857 0.3598 119
Turquoise Blue 0.0000 0.0000 0.0000 69
White 0.7518 0.7950 0.7728 5497
Yellow 0.7714 0.8003 0.7856 776
accuracy 0.7072 44072
macro avg 0.4112 0.3343 0.3469 44072
weighted avg 0.6919 0.7072 0.6935 44072
```
The model categorizes fashion product images into the following **46 base color classes**:
- Beige, Black, Blue, Bronze, Brown, Burgundy, Charcoal, Coffee Brown, Copper, Cream
- Fluorescent Green, Gold, Green, Grey, Grey Melange, Khaki, Lavender, Lime Green
- Magenta, Maroon, Mauve, Metallic, Multi, Mushroom Brown, Mustard, Navy Blue
- Nude, Off White, Olive, Orange, Peach, Pink, Purple, Red, Rose, Rust
- Sea Green, Silver, Skin, Steel, Tan, Taupe, Teal, Turquoise Blue, White, Yellow
---
# **Run with Transformers 🤗**
```python
!pip install -q transformers torch pillow gradio
```
```python
import gradio as gr
from transformers import AutoImageProcessor, SiglipForImageClassification
from PIL import Image
import torch
# Load model and processor
model_name = "prithivMLmods/Fashion-Product-baseColour" # Replace with actual model path
model = SiglipForImageClassification.from_pretrained(model_name)
processor = AutoImageProcessor.from_pretrained(model_name)
# Label mapping
id2label = {
0: "Beige", 1: "Black", 2: "Blue", 3: "Bronze", 4: "Brown", 5: "Burgundy",
6: "Charcoal", 7: "Coffee Brown", 8: "Copper", 9: "Cream", 10: "Fluorescent Green",
11: "Gold", 12: "Green", 13: "Grey", 14: "Grey Melange", 15: "Khaki", 16: "Lavender",
17: "Lime Green", 18: "Magenta", 19: "Maroon", 20: "Mauve", 21: "Metallic",
22: "Multi", 23: "Mushroom Brown", 24: "Mustard", 25: "Navy Blue", 26: "Nude",
27: "Off White", 28: "Olive", 29: "Orange", 30: "Peach", 31: "Pink", 32: "Purple",
33: "Red", 34: "Rose", 35: "Rust", 36: "Sea Green", 37: "Silver", 38: "Skin",
39: "Steel", 40: "Tan", 41: "Taupe", 42: "Teal", 43: "Turquoise Blue", 44: "White", 45: "Yellow"
}
def classify_base_color(image):
"""Predicts the base color of a fashion product from an image."""
image = Image.fromarray(image).convert("RGB")
inputs = processor(images=image, return_tensors="pt")
with torch.no_grad():
outputs = model(**inputs)
logits = outputs.logits
probs = torch.nn.functional.softmax(logits, dim=1).squeeze().tolist()
predictions = {id2label[i]: round(probs[i], 3) for i in range(len(probs))}
return predictions
# Gradio interface
iface = gr.Interface(
fn=classify_base_color,
inputs=gr.Image(type="numpy"),
outputs=gr.Label(label="Base Colour Prediction Scores"),
title="Fashion-Product-baseColour",
description="Upload a fashion product image to detect its primary color (e.g., Red, Black, Cream, Navy Blue, etc.)."
)
# Launch the app
if __name__ == "__main__":
iface.launch()
```
---
# **Intended Use**
This model is ideal for:
- **E-commerce platforms** for accurate product color labeling
- **Fashion search engines** and recommendation systems
- **Inventory and catalog automation**
- **Fashion analytics and trends tracking**
- **Design tools** for color-based sorting and filters |