myneighborh's picture
Update README.md
c9e2c3b verified
|
Raw
History Blame Contribute Delete
1.46 kB
metadata
tags:
  - image-classification
  - computer-vision
  - deep-learning
  - vehicle
  - pytorch
  - convnext
  - k-fold
  - tta
license: cc-by-nc-sa-4.0
datasets:
  - custom
model-index:
  - name: Vehicle Model Classifier (ConvNeXt + MixUp/CutMix)
    results:
      - task:
          type: image-classification
          name: Image Classification
        dataset:
          name: Custom Used Car Dataset
          type: image
        metrics:
          - type: log_loss
            value: 0.1435
sdk_version: 5.38.0

Vehicle Model Classifier

This deep learning model classifies 396 real-world used car models from vehicle images.

Final Model Details

  • Backbone: ConvNeXt-Base
  • Framework: PyTorch
  • Augmentation: MixUp / CutMix / Albumentations
  • Scheduler: CosineAnnealingLR / ReduceLROnPlateau
  • Validation: Stratified K-Fold (3-fold)
  • Ensemble: Logit & Soft Voting Ensemble
  • TTA: 3-view & 5-view Test Time Augmentation

How to Use

from transformers import AutoImageProcessor, AutoModelForImageClassification
from PIL import Image
import torch

processor = AutoImageProcessor.from_pretrained("myneighborh/vehicle-model-classifier")
model = AutoModelForImageClassification.from_pretrained("myneighborh/vehicle-model-classifier")

image = Image.open("path_to_image.jpg")
inputs = processor(images=image, return_tensors="pt")
outputs = model(**inputs)
pred = outputs.logits.argmax(-1).item()
label = model.config.id2label[pred]
print("Predicted label:", label)