Train-Cityscape-data-using-Pretrained-Resnet-18-model-using-Discriminative-Learning

Train Cityscape data using Pretrained Resnet 18 model using Discriminative Learning using Pytorch

Dataset:

Cityscapes Image Pairs

Implementation:

• The Resnet18 pretrained model was used to do the semantic segmentation task using Pytorch.
• Due to lack of time I couldn't extend the implementation in FastAPI
• The detailed implementation can be found from the colab notebook along with the documented step by step implementation guide.

References:

1A. Download Datasets from Kaggle on Colab

1. TORCHVISION.DATASETS

2. TRANSFER LEARNING FOR COMPUTER VISION TUTORIAL

3. Discriminative Fine-Tuning

4. ColabCode: Deploying Machine Learning Models From Google Colab

5. Image-Segmentation-with-UNet-PyTorch

6. Discriminative-learning-rates-PyTorch

7. semantic_segmentation_cityscape

8. Guide to Pytorch Learning Rate Scheduling

9. Discriminative Fine-Tuning

10. Differential Learning Rates

11. PyTorch learning rate finder