This is the GPU utilization of my model while training on A100 GPU on cloud. What could be the possible bottleneck? I'm using tfrecord datasets of each 550MB. GPU memory usage is almost 100% all the time.
0 Answers
Related
This is the GPU utilization of my model while training on A100 GPU on cloud. What could be the possible bottleneck? I'm using tfrecord datasets of each 550MB. GPU memory usage is almost 100% all the time.