image-medium-classifier-efficientnet-b0-v1

Anime/Real/Rendered Image Classifier (EfficientNet-B0)

Fast, lightweight classifier for distinguishing photographs from anime and 3D rendered images.

Model Details

• Architecture: EfficientNet-B0 (timm)
• Input Size: 224×224 RGB
• Classes: anime, real, rendered
• Parameters: 5.3M
• Validation Accuracy: 97.44%
• Training Speed: ~1 min/epoch (GPU)
• Inference Speed: ~20ms per image (RTX 3060)

Performance

Usage

from PIL import Image
import torch
from torchvision import transforms
import timm
from safetensors.torch import load_file

# Load model
model = timm.create_model('efficientnet_b0', num_classes=3, pretrained=False)
state_dict = load_file('model.safetensors')
model.load_state_dict(state_dict)
model.eval()

# Prepare image
transform = transforms.Compose([
    transforms.Resize((224, 224)),
    transforms.ToTensor(),
    transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225]),
])

image = Image.open('image.jpg').convert('RGB')
x = transform(image).unsqueeze(0)

# Predict
with torch.no_grad():
    logits = model(x)
    probs = torch.softmax(logits, dim=1)
    pred_class = probs.argmax(dim=1).item()

labels = ['anime', 'real', 'rendered']
print(f"{labels[pred_class]}: {probs[0, pred_class]:.2%}")

Dataset

• Real: 5,000 COCO 2017 validation images (diverse real-world scenarios)
• Anime: 2,357 curated anime/animation frames
• Rendered: 1,610 AAA game screenshots + 61 Pixar movie stills
• Total: 8,967 images (8,070 train / 897 val)

Training Details

• Augmentation: None (raw resize to 224×224)
• Optimizer: AdamW (lr=0.001)
• Loss: CrossEntropyLoss with class weighting
• Epochs: 20
• Batch Size: 80
• Hardware: NVIDIA RTX 3060 (12GB)

Known Limitations

• Real vs Rendered: Some confusion (photorealistic games misclassified as real)
• Stylized Games: Cel-shaded games (e.g., Fate/Extella) may score as anime
• Pixar: Stylized rendered images may show mixed confidence

Recommendations

• Use ensemble with tf_efficientnetv2_s for critical applications
• Apply confidence threshold: only trust predictions >85% confidence
• For edge cases, use the full confusion matrix to understand failure modes

License

OpenRAIL - Free for research and commercial use with proper attribution