#
places2
Here are 2 public repositories matching this topic...
Config-driven fine-tuning and benchmarking of Vision Transformers (ViT, DINOv2, Swin, ConvNeXt) against a ResNet-50 baseline for 40-class scene recognition on Places2, with attention-map explainability.
computer-vision deep-learning pytorch image-classification vit scene-recognition huggingface-transformers vision-transformer swin-transformer convnext dinov2 places2
-
Updated
Jun 18, 2026 - Python
Add this topic to your repo
To associate your repository with the places2 topic, visit your repo's landing page and select "manage topics."