Failed to allocate memory [Op:Mul]

Autonomous Machines Jetson & Embedded Systems Jetson Nano
1

I want to load a model in .h5 format, I do it this way:

model_new = keras.models.load_model("./Descargas/my_model(1)

After some time I get the following error:

File “/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/ops.py”, line 6941, in raise_from_not_ok_status
six.raise_from(core._status_to_exception(e.code, message), None)
File “”, line 3, in raise_from
tensorflow.python.framework.errors_impl.ResourceExhaustedError: failed to allocate memory [Op:Mul]

Apparently it’s a memory error and I can’t find a way to fix it

3

Hi,

The ResourceExhaustedError indicates that you are running out of memory.
Please double-check this by monitoring the system status with the following command at the same time:

$ sudo tegrastats

Since inference requires physical GPU memory, you will need to reduce the model complexity if OOM occurs.

Thanks.

4

Gracias por la colaboración.
En este momento me surge una duda, Si aumento la capacidad de memoria de la tarjeta micro SD, ¿De esta manera puedo solucionar el error?

5

Tambien me gustaría saber cual es o puede ser la causa de la siguiente alerta " System throttled due to Over-current"

6

Hi,

Unfortunately not.

GPU can only access the real physical memory.
The storage-based swap memory can only be used by the CPU.

For the “System throttled due to Over-current” issue, please check the following topic for details:

Thanks.

8

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.