add cli to list the supported model names (#639)

* update

* resolve comment

README.md

@@ -52,7 +52,7 @@ LMDeploy is a toolkit for compressing, deploying, and serving LLM, developed by
 
 ## Supported Models
 
-`LMDeploy` has two inference backends, `Pytorch` and `TurboMind`.
+`LMDeploy` has two inference backends, `Pytorch` and `TurboMind`. You can run `lmdeploy list` to check the supported model names.
 
 ### TurboMind
 

README_zh-CN.md

@@ -53,7 +53,7 @@ LMDeploy 由 [MMDeploy](https://github.com/open-mmlab/mmdeploy) 和 [MMRazor](ht
 
 ## 支持的模型
 
-`LMDeploy` 支持 `TurboMind` 和 `Pytorch` 两种推理后端
+`LMDeploy` 支持 `TurboMind` 和 `Pytorch` 两种推理后端。运行`lmdeploy list`可查看支持模型列表
 
 ### TurboMind
 

lmdeploy/cli/cli.py

@@ -49,6 +49,30 @@ class CLI(object):
                 quant_path=quant_path,
                 group_size=group_size)
 
+    def list(self, engine: str = 'turbomind'):
+        """List supported model names.
+
+        Examples 1:
+            lmdeploy list
+
+        Examples 2:
+            lmdeploy list --engine pytorch
+
+        Args:
+            engine (str): The backend for the model to run. Choice from
+                ['turbomind', 'pytorch'].
+        """
+        assert engine in ['turbomind', 'pytorch']
+        if engine == 'pytorch':
+            model_names = ['llama', 'llama2', 'internlm-7b']
+        elif engine == 'turbomind':
+            from lmdeploy.model import MODELS
+            model_names = list(MODELS.module_dict.keys())
+            model_names = [n for n in model_names if n.lower() not in ['base']]
+        model_names.sort()
+        print('Supported model names:')
+        print('\n'.join(model_names))
+
 
 def run():
     """The entry point of running LMDeploy CLI."""