• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 98 subscribers
  • Views 197 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

SK-AM62B-P1: Training face recognition with custom dataset

Namitha K Nair
Prodigy 20 points
Part Number: SK-AM62B-P1

Tool/software:

Hi,

I am currently working on a solution for driver authentication using face recognition and drowsiness detection .For this ,I am using face_recognition  and Opencv models .However,training the model with custom dataset in AM62B-P1 platform takes a considerable amount of time.Are there any AI tools or alternative methods which could help this training duration ?

  • 0
    TI__Genius 15556 points

    Hello,

    Are you training on the AM62B itself or training a model on PC, with the intent of deploying the model to the SOC? I would not recommend training on the device - I would expect it to be very slow, if it even runs (training usually requires more RAM than the EVM's have onboard.

    There are some open source models out there for face recognition but I have found that datasets and information for using those models is sparse (legal restrictions on models notwithstanding. TI does not currently have an officially supported face recognition / drowsiness detection model.

  • 0 in reply to Reese Grimsley
    Prodigy 20 points

    Hi Reese,


    I am trying to train the model  on AM62B itself and it takes time.Thankyou for the response.

  • 0 in reply to Namitha K Nair
    TI__Genius 15556 points

    Hi Namitha,

    Training on the device itself is not officially supported. Training typically happens on a much more powerful computer, like a PC or a server with GPUs. The GPU on AM62B does not support acceleration of AI model training