Quellcode durchsuchen

Add link to LiteLLM to make-setup (#614)

* Update Makefile

* fix tab

* add note to readme
Robert Brennan vor 1 Jahr
Ursprung
Commit
310cd7017d
2 geänderte Dateien mit 14 neuen und 3 gelöschten Zeilen
  1. 6 3
      Makefile
  2. 8 0
      README.md

+ 6 - 3
Makefile

@@ -42,14 +42,17 @@ run:
 # Setup config.toml
 setup-config:
 	@echo "Setting up config.toml..."
-	@read -p "Enter your LLM API key: " llm_api_key; \
-	 echo "LLM_API_KEY=\"$$llm_api_key\"" >> $(CONFIG_FILE).tmp
-	@read -p "Enter your LLM Model name [default: $(DEFAULT_MODEL)]: " llm_model; \
+	@read -p "Enter your LLM Model name (see docs.litellm.ai/docs/providers for full list) [default: $(DEFAULT_MODEL)]: " llm_model; \
 	 llm_model=$${llm_model:-$(DEFAULT_MODEL)}; \
 	 echo "LLM_MODEL=\"$$llm_model\"" >> $(CONFIG_FILE).tmp
+
+	@read -p "Enter your LLM API key: " llm_api_key; \
+	 echo "LLM_API_KEY=\"$$llm_api_key\"" >> $(CONFIG_FILE).tmp
+
 	@read -p "Enter your workspace directory [default: $(DEFAULT_WORKSPACE_DIR)]: " workspace_dir; \
 	 workspace_dir=$${workspace_dir:-$(DEFAULT_WORKSPACE_DIR)}; \
 	 echo "WORKSPACE_DIR=\"$$workspace_dir\"" >> $(CONFIG_FILE).tmp
+
 	@mv $(CONFIG_FILE).tmp $(CONFIG_FILE)
 
 # Help

+ 8 - 0
README.md

@@ -130,6 +130,14 @@ Getting started with the OpenDevin project is incredibly easy. Follow these simp
     make setup-config
     ```
 
+You'll need to choose your LLM model as part of this step. By default, we use OpenAI's gpt-4, but you can
+use Anthropic's Claude, ollama, or any other LLM provider supported by LiteLLM. See the full
+list of supported models at [docs.litellm.ai/docs/providers](https://docs.litellm.ai/docs/providers).
+
+(Note: alternative models can be hard to work with. We will make LLM-specific available documentation available soon.
+If you've gotten OpenDevin working with a model other than OpenAI's GPT models,
+please [add your setup instructions here](https://github.com/OpenDevin/OpenDevin/issues/417).)
+
 ### 3. Run the Application
 
 - **Run the Application:** Once the setup is complete, launching OpenDevin is as simple as running a single command. This command starts both the backend and frontend servers seamlessly, allowing you to interact with OpenDevin without any hassle.