search

SrijanShovit / CrySense

An ML project aimed at helping parents detect the reason of their cute babies' cry!!
6 stars 8 forks source link

Models and results #19

Open adi271001 opened 6 months ago

adi271001 commented 6 months ago

@SrijanShovit I have made 2 models one using CNN and other using efficientNet CNN gave me around 66% accuracy and EfficientNet gave me 82.5% accuracy now thing is as in the dataset 300+ audios are there for hungry rest all range from 8-32 so dtaatset is imbalanced because of which predictions aren't that accurate

SrijanShovit commented 6 months ago

Hey @adi271001 You can once try the augmented dataset. Still I admire your efforts.

adi271001 commented 6 months ago

Ok will definitely try

adi271001 commented 6 months ago

@SrijanShovit i have made 6 models 1. using CNN 2. using EfficientNet 3. VGG 16 4. ResNet50 5. DenseNet121 6. VGG19 with the augmented dataset all gave accuracy from range 82-86% can you please assign this to me under codepeak i'll add all models

SrijanShovit commented 6 months ago

Great @adi271001

SrijanShovit commented 6 months ago

@adi271001 All 3 of your PRs resolve the same issue pointed by you. Hence, in total you will be getting 3 points as the issue was pointed out by you. However, I do appreciate your efforts and they deserve more.