update README

README.md

@@ -22,7 +22,6 @@ This example code fine-tunes BERT on the SQuAD1.0 dataset.
 python /nx/transformers/examples/pytorch/question-answering/run_qa.py \
   --model_name_or_path /models/google-bert/bert-large-uncased-whole-word-masking-finetuned-squad \
   --dataset_name squad \
-  --do_train \
   --do_eval \
   --per_device_train_batch_size 12 \
   --learning_rate 3e-5 \
@@ -32,17 +31,9 @@ python /nx/transformers/examples/pytorch/question-answering/run_qa.py \
   --output_dir /nx/transformers/debug_squad/
 ```
 
-### 多卡**微调（Fine-tuning）**
+### 多卡inference
 
 ```
-bash fine-tuning_SQuAD1.0.sh
-```
-
-### 推理
-
-使用[evaluate-v1.1.py](https://github.com/allenai/bi-att-flow/blob/master/squad/evaluate-v1.1.py)对BERT-Large微调模型对SQuAD1.0 validation （[dev-v1.1.json](https://rajpurkar.github.io/SQuAD-explorer/dataset/dev-v1.1.json)）进行推理
-
-```
-bash infer.sh
+bash inference_SQuAD1.0.sh
 ```