Coin-CLIP 🪙 : Enhancing Coin Image Retrieval with CLIP
Model Details / 模型细节
This model (Coin-CLIP) is built upon
OpenAI's CLIP (ViT-B/32) model and fine-tuned on
a dataset of more than 340,000 coin images using contrastive learning techniques. This specialized model is designed to significantly improve feature extraction for coin images, leading to more accurate image-based search capabilities. Coin-CLIP combines the power of Visual Transformer (ViT) with CLIP's multimodal learning capabilities, specifically tailored for the numismatic domain.
Key Features:
- State-of-the-art coin image retrieval;
- Enhanced feature extraction for numismatic images;
- Seamless integration with CLIP's multimodal learning.
本模型（Coin-CLIP）
在 OpenAI 的 CLIP (ViT-B/32) 模型基础上，利用对比学习技术在超过 340,000 张硬币图片数据上微调得到的。
Coin-CLIP 旨在提高模型针对硬币图片的特征提取能力，从而实现更准确的以图搜图功能。该模型结合了视觉变换器（ViT）的强大功能和 CLIP 的多模态学习能力，并专门针对硬币图片进行了优化。
Comparison: Coin-CLIP vs. CLIP / 效果对比
Example 1 (Left: Coin-CLIP; Right: CLIP)
Example 2 (Left: Coin-CLIP; Right: CLIP)
More examples can be found: breezedeus/Coin-CLIP: Coin CLIP .
Usage and Limitations / 使用和限制
-
Usage: This model is primarily used for extracting representation vectors from coin images, enabling efficient and precise image-based searches in a coin image database.
-
Limitations: As the model is trained specifically on coin images, it may not perform well on non-coin images.
-
用途：此模型主要用于提取硬币图片的表示向量，以实现在硬币图像库中进行高效、精确的以图搜图。
-
限制：由于模型是针对硬币图像进行训练的，因此在处理非硬币图像时可能效果不佳。
Documents / 文档
- Base Model: openai/clip-vit-base-patch32
Model Use / 模型使用
Transformers
from PIL import Image
import requests
import torch.nn.functional as F
from transformers import CLIPProcessor, CLIPModel
model = CLIPModel.from_pretrained("breezedeus/coin-clip-vit-base-patch32")
processor = CLIPProcessor.from_pretrained("breezedeus/coin-clip-vit-base-patch32")
image_fp = "path/to/coin_image.jpg"
image = Image.open(image_fp).convert("RGB")
inputs = processor(images=image, return_tensors="pt")
img_features = model.get_image_features(**inputs)
img_features = F.normalize(img_features, dim=1)
Tool / 工具
To further simplify the use of the Coin-CLIP model, we provide a simple Python library breezedeus/Coin-CLIP: Coin CLIP for quickly building a coin image retrieval engine.
为了进一步简化 Coin-CLIP 模型的使用，我们提供了一个简单的 Python 库 breezedeus/Coin-CLIP: Coin CLIP，以便快速构建硬币图像检索引擎。
Install
pip install coin_clip
Extract Feature Vectors
from coin_clip import CoinClip
# Automatically download the model from Huggingface
model = CoinClip(model_name='breezedeus/coin-clip-vit-base-patch32')
images = ['examples/10_back.jpg', 'examples/16_back.jpg']
img_feats, success_ids = model.get_image_features(images)
print(img_feats.shape) # --> (2, 512)
More Tools can be found: breezedeus/Coin-CLIP: Coin CLIP .
Training Data / 训练数据
The model was trained on a specialized coin image dataset. This dataset includes images of various currencies' coins.
本模型使用的是专门的硬币图像数据集进行训练。这个数据集包含了多种货币的硬币图片。
Training Process / 训练过程
The model was fine-tuned on the OpenAI CLIP (ViT-B/32) pretrained model using a coin image dataset. The training process involved Contrastive Learning fine-tuning techniques and parameter settings.
模型是在 OpenAI 的 CLIP (ViT-B/32) 预训练模型的基础上，使用硬币图像数据集进行微调。训练过程采用了对比学习的微调技巧和参数设置。
Performance / 性能
This model demonstrates excellent performance in coin image retrieval tasks.
该模型在硬币图像检索任务上展现了优异的性能。
Feedback / 反馈
Where to send questions or comments about the model.
Welcome to contact the author Breezedeus.
欢迎联系作者 Breezedeus 。