modify sh

demo/text-classification/run_classifier.sh

 export FLAGS_eager_delete_tensor_gb=0.0
-export CUDA_VISIBLE_DEVICES=0
 
 # User can select chnsenticorp, nlpcc_dbqa, lcqmc and so on for different task
 DATASET="chnsenticorp"
@@ -15,8 +14,7 @@ python -u text_classifier.py \
                    --max_seq_len=128 \
                    --num_epoch=3 \
                    --use_pyreader=True \
-                   --use_data_parallel=True \
-                   --use_taskid=False
+                   --use_data_parallel=True
 
 # Recommending hyper parameters for difference task
 # for ChineseGLUE: