Update README

examples/st_covost2/README.md

@@ -1,5 +1,15 @@
 # ST_covost2
 
+
+## Model Stracture
+<img src="image/framework.jpg" alt="示例图片" style="width:75%;">
+
+
+## Multitask 
+<img src="image/prompt.png" alt="示例图片" style="width:50%;">
+
+
+
 ## Download Model 
 We only train the q-former projector in this recipe.
 Encoder | Projector | LLM 
@@ -33,18 +43,16 @@ You can find the test jsonl in "test_st.jsonl"
 {"audio": "/userhome/speech/data/common/4/en/clips/common_voice_en_699711.mp3", "prompt": "\"She'll be all right.\"<|zh|>", "gt": "\"She'll be all right.\"<|zh|>她会没事的。", "source": "covost_enenzh"}
 ```
 ## Train Stage
-Here, we have designed a four-step training process, where each training session uses the checkpoint obtained from the previous training session.
+Here, we have designed a three-step training process, where each training session uses the checkpoint obtained from the previous training session.
 ```
 #In this step, we perform ASR pretraining to acquire speech recognition capabilities.
 bash asr_pretrain.sh
 
 #In this phase, we conduct multimodal machine translation training to enhance the final performance.
 bash mmt.sh
 
-#monolingual SRT training.
+#monolingual SRT training and multitask training.
 bash srt.sh
-
-#multilingual multitask training.
 bash zsrt.sh
 ```
 
@@ -53,7 +61,7 @@ bash zsrt.sh
 You can try our pre-trained model.
 
 ```
-bash infer.sh
+bash infer_enzh.sh
 ```
 
 ##  Citation