Embedded Files
Model 1 : Simple CNN
Model 1 : Simple CNN
The performance with simple CNN depicts the training and validation accuracy are performing well with val loss flattening down
We performed hyperparameter tuning and saw the combinations that could give us the best output, " Adam with 0.001 and 32 epochs" gave us an accuracy of 88%.
Model 2 : Resnet ( Residual Network)
Model 2 : Resnet ( Residual Network)
The performance with Resnet depicts the training and validation accuracy are struggling to fit well with the data along with val loss fluctuating down
Model 3 : Pretrained Resnet (Imagenet) Model
Model 3 : Pretrained Resnet (Imagenet) Model
The performance with Pretrained Resnet depicts the training and validation accuracy are struggling a little while the performance looks to be doing great. Moreover, the data along with val loss remains down with constant
Model 4: Final Convolutional Neural Network
Model 4: Final Convolutional Neural Network
The performance with Pretrained Resnet depicts the training and validation accuracy are struggling a little while the performance looks to be doing great. Moreover, the data along with val loss remains down with constant
Model 4 : Final CNN Model Predictions
Model 4 : Final CNN Model Predictions
Page updated
Google Sites
Report abuse