Datasets
Models
Downloads: 1
[14ef49]:
/
Model_Tracker_Deeptek_classification.csv
| # | model_id | architecture | batch_size | img_size | learning_rate | optimizer | loss function | weight_path | logs_path | Colab_URL | comments | lossfunction | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 0 | DN_1 | DenseNet121 | 1 | 512 | 0.001 | Adam | Binary crossentropy | /content/drive/MyDrive/classification/saved Models/Pretrained DenseNet121/DenseNet121-model.h5 | /content/drive/MyDrive/logs/Binary_Classification_DenseNet121_1.csv | https://colab.research.google.com/drive/1U3dCtpGbpCuS09N4D_WhQ_S2Y3ybsR75#scrollTo=33PPUeC6S44e | The image size was 512 x 512 and had a good AU|ROC score of 0.85 | Binary Crossentropy |
| 2 | 1 | IV3_2 | InceptionV3 | 1 | 512 | 0.001 | Adam | /content/drive/MyDrive/classification/saved Models/Pretrained InceptionV3_1/InceptionV3_1-model.h5 | /content/drive/MyDrive/logs/Binary_Classification_InceptionV3_2.csv | https://colab.research.google.com/drive/1ZDqxDcG3IAaftRLxIEMKBSjCXxNzy4Nl#scrollTo=yWrAA3OrGBcP | The image size for this model training was changed to 224 x 224 instead of 512 x 512 | Binary Crossentropy | |
| 3 | 2 | RN_1 | ResNet50 | 1 | 512 | 0.001 | Adam | /content/drive/MyDrive/classification/saved Models/Pretrained ResNet50/ResNet50-model.h5 | /content/drive/MyDrive/logs/Binary_Classification_ResNet_1.csv | https://colab.research.google.com/drive/17J-0KJ8WQiHLx1zVFjjgr7_wNB9xRmA1#scrollTo=sXKhxRpwMOxn | The image size for this model training was 512 x 512 and an average Au|ROC score of 0.69 | Binary Crossentropy | |
| 4 | 3 | VG_1 | VGG16 | 1 | 512 | 0.001 | Adam | /content/drive/MyDrive/classification/saved Models/Pretrained VGG16/VGG16-model.h5 | /content/drive/MyDrive/logs/Binary_Classification_VGG16_1.csv | https://colab.research.google.com/drive/1DxIczGkuI7LQS-kSyqixAq4DzxLspp0H#scrollTo=0l9o2hwUgwDG | The image size for this model training was 512 x 512 and a good AU|ROC score of 0.82 | Binary Crossentropy | |
| 5 | 4 | IV3_1 | InceptionV3 | 1 | 512 | 0.001 | Adam | /content/drive/MyDrive/classification/saved Models/Pretrained InceptionV3/InceptionV3-model.h5 | /content/drive/MyDrive/logs/Binary_Classification_InceptionV3_1.csv | https://colab.research.google.com/drive/1ZM3_SiDqyMx9aLezFZkEYX6fCXTL2hUP#scrollTo=cuvaItqflfiQ | The image size for this model training was 512 x 512 with a good AU|ROC score of 0.86 | Binary Crossentropy |