Created by: zhengya01

tagspace 模型 直接跑样例数据单机多卡训练一个epoch后报错 paddle1.3 CUDA9 cudnn7 gpu ：4卡

以下是报错日志： TRAIN --> pass: 0 batch_num: 950 avg_cost: 0.0400084555149, acc: 1.0 TRAIN --> pass: 0 batch_num: 1000 avg_cost: 0.112019307911, acc: 0.4 epoch:1 num_steps:199 time_cost(s):1.300218 Traceback (most recent call last): File "train.py", line 128, in train() File "train.py", line 123, in train train_exe) File "/home/paddle/anaconda2/lib/python2.7/site-packages/paddle/fluid/io.py", line 1008, in save_inference_model save_persistables(executor, dirname, main_program, params_filename) File "/home/paddle/anaconda2/lib/python2.7/site-packages/paddle/fluid/io.py", line 487, in save_persistables filename=filename) File "/home/paddle/anaconda2/lib/python2.7/site-packages/paddle/fluid/io.py", line 174, in save_vars filename=filename) File "/home/paddle/anaconda2/lib/python2.7/site-packages/paddle/fluid/io.py", line 210, in save_vars executor.run(save_program) File "/home/paddle/anaconda2/lib/python2.7/site-packages/paddle/fluid/parallel_executor.py", line 301, in run self.executor.run(fetch_list, fetch_var_name) TypeError: run(): incompatible function arguments. The following argument types are supported: 1. (self: paddle.fluid.core.ParallelExecutor, arg0: List[unicode], arg1: unicode) -> None

Invoked with: <paddle.fluid.core.ParallelExecutor object at 0x7f78001ca688>, blocks { idx: 0 parent_idx: -1 vars { name: "cnn" type { type: LOD_TENSOR lod_tensor { tensor { data_type: FP32 dims: 50 dims: 1000