Showing 1 - 20 results of 2,821 for search '(( trained cnn decrease ) OR ( a ((larger decrease) OR (marked decrease)) ))', query time: 0.46s Refine Results
  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6

    Training Data/Validation/Test. by Mudhafar Jalil Jassim Ghrabat (22177655)

    Published 2025
    “…Every model was subjected to individual testing. The SMO_CNN model we developed demonstrated exceptional testing and training accuracies of 98.95% and 99.20% respectively, surpassing CNN, VGG19, and ResNet50 models. …”
  7. 7
  8. 8

    CNN Model Architecture. by Mudhafar Jalil Jassim Ghrabat (22177655)

    Published 2025
    “…Every model was subjected to individual testing. The SMO_CNN model we developed demonstrated exceptional testing and training accuracies of 98.95% and 99.20% respectively, surpassing CNN, VGG19, and ResNet50 models. …”
  9. 9

    Biases in larger populations. by Sander W. Keemink (21253563)

    Published 2025
    “…<p>(<b>A</b>) Maximum absolute bias vs the number of neurons in the population for the Bayesian decoder. …”
  10. 10
  11. 11
  12. 12

    Flowchart of Proposed SMO_CNN. by Mudhafar Jalil Jassim Ghrabat (22177655)

    Published 2025
    “…Every model was subjected to individual testing. The SMO_CNN model we developed demonstrated exceptional testing and training accuracies of 98.95% and 99.20% respectively, surpassing CNN, VGG19, and ResNet50 models. …”
  13. 13

    CNN Model Layers Summary. by Mudhafar Jalil Jassim Ghrabat (22177655)

    Published 2025
    “…Every model was subjected to individual testing. The SMO_CNN model we developed demonstrated exceptional testing and training accuracies of 98.95% and 99.20% respectively, surpassing CNN, VGG19, and ResNet50 models. …”
  14. 14
  15. 15
  16. 16
  17. 17

    Performance Results of Proposed SMO_CNN Model. by Mudhafar Jalil Jassim Ghrabat (22177655)

    Published 2025
    “…Every model was subjected to individual testing. The SMO_CNN model we developed demonstrated exceptional testing and training accuracies of 98.95% and 99.20% respectively, surpassing CNN, VGG19, and ResNet50 models. …”
  18. 18
  19. 19
  20. 20