Start bert_as_service fail.
Created by: ShenYuhan
Env: PaddlePaddle: 1.6.2 Paddle_gpu_serving: 0.8.2 platform:Linux version 3.10.0_3-0-0-20 (gcc version 4.8.2 20140120 (Red Hat 4.8.2-16) (GCC) ) #1 (closed) SMP Thu May 24 14:41:54 CST 2018
When execute:
from paddle_gpu_serving.run import BertServer
bs = BertServer(with_gpu=False)
bs.with_model("ernie")
bs.run(port=8010)