/LG-AI-Challenge-for-Plant-Classification

πŸ₯‡ LG-AI-Challenge 2022 1μœ„ μ†”λ£¨μ…˜ μž…λ‹ˆλ‹€.

Primary LanguageJupyter Notebook

LG-AI-Challenge-for-Plant-Classification

Daconμ—μ„œ μ§„ν–‰λœ 농업 ν™˜κ²½ 변화에 λ”°λ₯Έ μž‘λ¬Ό 병해 진단 AI κ²½μ§„λŒ€νšŒ 에 λŒ€ν•œ μ½”λ“œμž…λ‹ˆλ‹€. (colab directory에 μ½”λ“œκ°€ 잘 정리 λ˜μ–΄μžˆμŠ΅λ‹ˆλ‹€.)

Requirements

  • python==3.7.12
  • albumentations==1.1.0
  • numpy==1.19.5
  • pandas==1.3.5
  • cv2==4.1.2
  • sklearn==1.0.2
  • json==2.0.9
  • torch==1.10.0+cu111
  • timm==0.5.4
  • transformers==4.16.2

Preprocessing(image)

  • Augmentations: Transpose, Flip, Rotate, RandomBrightnessContrast, Cutmix ...

Preprocessing(sequence)

  • Augmentations : λ§Œμ•½ 길이가 500이상이라면 random으둜 sampling(2μ”©)
  • MinMax Scaling

Model

  • resnext50_32x4d

Training

  • K-fold Cross Validation(k=5)
  • Use Cutmix augmentation until epoch 15
  • Save model's weight when score is highest

inference

  • K-fold Model Ensemble(Soft Voting)