not detecting

[sahil-dhuri]

i have trained my model in same way as u did for max 50 epochs

but then to its not detecting

i have seen in train folder the images are labeled very well

but it was not able to predict while training itself i think so..

[prp-e]

Use what you see in last instead of best or increase the number of epochs. I had same problem(s).

[wacky7]

i have trained my model in same way as u did for max 50 epochs

but then to its not detecting

i have seen in train folder the images are labeled very well

but it was not able to predict while training itself i think so..

same issues here, let me know if you've found the solution.

[gaurav00700]

Hi,
Actually there is no issue in your model. Your model is not detecting due to low prediction threshold limit. You can increase the number of echos, this will improve accuracy and reduce loss. This way you can get prediction on test data.

[prashantmore277]

Hi,
Actually there is no issue in your model. Your model is not detecting due to low prediction threshold limit. You can increase the number of epochs, this will improve accuracy and reduce loss. This way you can get prediction on test data.

for 50 epochs accuracy(F1 Score) is 0.49
and It is only detecting drowsy as I have trained this model with only 20 image (10 Awake and 10 Drowsy)
and sometime aways with close eyes also

after increasing training image and epochs the result is improved so there is not any problem with model its just that you have to fine tune model until it gives you an perfect result