Closed FurkanGozukara closed 1 week ago
I am getting cuda out of memory error very weirdly on Kaggle with 15 GB GPU
obviously there is a bug
meanwhile how can I force it to use CPU?
this is my usage
faces = RetinaFace.detect_faces(image_path)
found the solution :D
with tf.device('/CPU:0'): faces = RetinaFace.detect_faces(image_path)
I am getting cuda out of memory error very weirdly on Kaggle with 15 GB GPU
obviously there is a bug
meanwhile how can I force it to use CPU?
this is my usage
faces = RetinaFace.detect_faces(image_path)