Skip to content

v2.0.0 : High-Performance Optimization (mAP 0.922) & Model Pruning

Latest
Latest

Choose a tag to compare

View all tags
@songCode404 songCode404 released this 18 Jan 04:58
· 5 commits to main since this release
Immutable release. Only release title and notes can be modified.
v2.0.0
ca61a0d
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

💎 Highlights

  • Significant mAP Boost: 0.700 → 0.922 (+31%)
  • Data-Centric AI: Resolved dataset imbalance and added hard negative samples.
  • Model Compression: Applied 30% L1 Pruning and successfully recovered performance via fine-tuning.
  • Optimized for Speed: Added ONNX export & FP16 quantization for real-time inference (1.9ms).

📌 Attach Files Info

  • v2_baseline_yolov5m.pt : 하이퍼파라미터 조정, 비율 재조정, background data 추가된 set으로 학습된 yolov5m baseline 가중치
  • pruned_v2_30.pt : 30% pruning 된 모델 가중치 (fine-tuning 전)
  • v2_30pruning_finetuning_best.pt : 30% pruning 후 50에폭 추가 학습으로 fine-tuning
Assets 6
Loading