adding support for Qwen models and fixed issue with passing PATH vari… (#583)

* adding support for Qwen models and fixed issue with passing PATH variable

* don't need to have qwen in the model alias routing example

* fixed base_url for qwen

---------

Co-authored-by: Salman Paracha <salmanparacha@MacBook-Pro-288.local>

arch/tools/cli/config_generator.py

@@ -22,6 +22,7 @@ SUPPORTED_PROVIDERS = [
     "ollama",
     "moonshotai",
     "zhipu",
+    "qwen",
 ]
 
 
@@ -187,7 +188,7 @@ def validate_and_render_schema():
 
             # Validate azure_openai and ollama provider requires base_url
             if (
-                provider == "azure_openai" or provider == "ollama"
+                provider == "azure_openai" or provider == "ollama" or provider == "qwen"
             ) and model_provider.get("base_url") is None:
                 raise Exception(
                     f"Provider '{provider}' requires 'base_url' to be set for model {model_name}"

arch/tools/cli/main.py

@@ -205,6 +205,8 @@ def up(file, path, service, foreground):
         "MODEL_SERVER_PORT": os.getenv("MODEL_SERVER_PORT", "51000"),
     }
     env = os.environ.copy()
+    # Remove PATH variable if present
+    env.pop("PATH", None)
     # check if access_keys are preesnt in the config file
     access_keys = get_llm_provider_access_keys(arch_config_file=arch_config_file)