Tensor flow allocator ran out of memory when ever trying run a tensorflow object tracking program in tx1

My TX1 has a RAM of 4GB and internal memory space remaining 1.5 GB and have used part of sd card for swap memory of 10 GB and allocated swappiness of 100. Still when i run a tensorflow code for object detection using webcam it still shows below

“tensorflow/core/common_runtime/bfc_allocator.cc:239] Allocator (GPU_0_bfc) ran out of memory trying to allocate 2.27GiB with freed_by_count=0. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available”

Even when i have allocated 100 swappines when tensorflow is running only swap memory of 1.4 to 1.6 gb is used and Ram memory gets almost full. Any one please help me how to make swap memory utilization higher than RAM and finally to get results for my tensorflow program.


SWAP space cannot be accessed via GPU so it won’t increase the memory amount if your TensorFlow is on GPU mode.