Miewid Msv3

🚀 MiewID-msv3 Model

MiewID-msv3 is a feature extractor designed for wildlife re-identification. It leverages contrastive learning on a large, high - quality dataset encompassing 64 terrestrial and aquatic wildlife species, including various body parts like fins, flukes, flanks, and faces.

🚀 Quick Start

The intended use of the MiewID - msv3 model is to re - identify individuals from different species by matching against a database of ground - truth samples. Additionally, the model features can be used for species classification through retrieval.

✨ Features

• Wildlife Re - identification: Capable of accurately re - identifying individuals across 64 different wildlife species.
• Feature Extraction: Extracts useful features for both re - identification and species classification tasks.

📦 Installation

No specific installation steps are provided in the original README. If you plan to use the model, you can follow the embedding extraction code which will download the model using the transformers library.

💻 Usage Examples

Basic Usage

import numpy as np
from PIL import Image
import torch
import torchvision.transforms as transforms
from transformers import AutoModel

model_tag = f"conservationxlabs/miewid-msv3"
model = AutoModel.from_pretrained(model_tag, trust_remote_code=True)

def generate_random_image(height=440, width=440, channels=3):
    random_image = np.random.randint(0, 256, (height, width, channels), dtype=np.uint8)
    return Image.fromarray(random_image)

random_image = generate_random_image()

preprocess = transforms.Compose([
    transforms.Resize((440, 440)),
    transforms.ToTensor(),
    transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),
])

input_tensor = preprocess(random_image)
input_batch = input_tensor.unsqueeze(0) 

with torch.no_grad():
    output = model(input_batch)

print(output)
print(output.shape)

Advanced Usage

View more usage examples at https://github.com/WildMeOrg/wbia-plugin-miew-id/tree/main/wbia_miew_id/examples

📚 Documentation

Model Details

Model Description

MiewID - msv3 is a wildlife re - identification feature backbone. Here are its key statistics:

Model Sources

• Repository: https://github.com/WildMeOrg/wbia-plugin-miew-id
• Backbone: https://huggingface.co/timm/efficientnetv2_rw_m.agc_in1k

Training Details

Training Data

The dataset used for these experiments was a combination of data from Wildbook platforms (multiple users), the Happywhale Kaggle competitions multi - species dataset, and multiple publicly available datasets. A small subset of data from Wildbook platforms is available at https://lila.science/datasets.

Example Images

Evaluation Results