Adjust some expression in User Guide

docs/hugectr_user_guide.md

@@ -11,13 +11,13 @@ Highlighted features of HugeCTR
 ## New Features in Version 2.1
 HugeCTR version 2.1 is a major update which aims to provide a flexible, fast, scalable and reliable solution for CTR Training. Framework designers can consider it as a high-performance reference design. 
 
-* Supporting three important networks: Wide and Deep Learning (WDL)[1], Deep Cross Network (DCN)[2] and DeepFM [3] 
-* A new embedding implementation `LocalizedSlotSparseEmbedding` which reduces the memory transactions across GPUs and nodes resiliently to the number of GPUs.
-* Supporting multiple Embeddings in one network
-* Supporting dense feature input, which doesn't need any embdding layer
-* Supporting new layers like: Dropout / Split / Reshape / Multiply / FmOrder2 / MultCross / Add
-* Check bits in data reader to enable data check and error skip.
-* L1 / L2 Regularization
+* Supporting three important networks: Wide and Deep Learning (WDL)[1], Deep Cross Network (DCN)[2] and DeepFM [3];
+* A new embedding implementation `LocalizedSlotSparseEmbedding` which reduces the memory transactions across GPUs and nodes resiliently to the number of GPUs;
+* Supporting multiple Embeddings in one network;
+* Supporting dense feature input;
+* Supporting new layers like: Dropout / Split / Reshape / Multiply / FmOrder2 / MultCross / Add;
+* Check bits in data reader to enable data check and error skip;
+* L1 / L2 Regularization.
 
 ## Architecture and Supported Networks
 To enable large embedding training, the embedding table in HugeCTR is model parallel and distributed across all the GPUs in a homogeneous cluster, which consists of multiple nodes and multiple GPUs. Meanwhile, the dense model such as DNN is data parallel, which has one copy in each GPU (see Fig.1).