I am trying to conduct inference doing real-time object tracking on a jetson nano using tensorflow, but I am getting this error:
F tensorflow/core/kernels/resize_bilinear_op_gpu.cu.cc:493] Non-OK-status: GpuLaunchKernel(kernel, config.block_count, config.thread_per_block, 0, d.stream(), config.virtual_thread_count, images.data(), height_scale, width_scale, batch, in_height, in_width, channels, out_height, out_width, output.data()) status: Internal: too many resources requested for launch Fatal Python error: Aborted
My settings are as follows :
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3' os.environ['CUDA_VISIBLE_DEVICES'] = '0' physical_devices = tf.config.experimental.list_physical_devices('GPU') tf.config.experimental.set_memory_growth(physical_devices, True)
In another post, someone suggested turning the CUDA_VISIBLE_DEVICES to 1, but this disables the GPU, which is the whole reason I am using the Nano. I know this has to do with the number of threads, but I don’t know what to do about it. Does anyone have any advice? Thanks!