Polish README.md (#2138)

* Fix grammar errors.

* Update the train command sample in README.md

* Update argument values.

Author: brightcoder01

README.md

@@ -29,8 +29,12 @@ with Keras API, train the model distributedly with a command line.
 
 ```bash
 elasticdl train \
-  --model_def=mnist_functional_api.custom_model \
-  --training_data=/mnist/train --output=output
+  --image_name=elasticdl:mnist \
+  --model_zoo=model_zoo \
+  --model_def=mnist_functional_api.mnist_functional_api.custom_model \
+  --training_data=/data/mnist/train \
+  --job_name=test-mnist \
+  --volume="host_path=/data,mount_path=/data"
 ```
 
 ### Integration with SQLFlow
@@ -56,9 +60,9 @@ computing job would fail; however, we can restart the job and recover its status
 from the most recent checkpoint files.
 
 ElasticDL, as an enhancement of TensorFlow's distributed training feature,
-supports fault-tolerance. In the case that some processes fail, the job would go
-on running. Therefore, ElasticDL doesn't need to checkpoint nor recover from
-checkpoints.
+supports fault-tolerance. In the case that some processes fail, the job would
+go on running. Therefore, ElasticDL doesn't need to save checkpoint nor recover
+from checkpoints.
 
 The feature of fault-tolerance makes ElasticDL works with the priority-based
 preemption of Kubernetes to achieve elastic scheduling.  When Kubernetes kills