Merge pull request #11 from haseeb-heaven/bug/local-model-env-check

Update offline model configuration and skip client initialization for…

README.md

@@ -108,8 +108,8 @@ This Interpreter supports offline models via **LM Studio** so to download it fro
 - Download any model from **LM Studio** like _Phi-2,Code-Llama,Mistral_.
 - Then in the app go to **Local Server** option and select the model.
 - Start the server and copy the **URL**.
-- Open config file `configs/offline-model.config` and paste the **URL** in the `api_base` field.
-- Now you can use the model with the interpreter set the model name to `offline-model` and run the interpreter.</br>
+- Open config file `configs/local-model.config` and paste the **URL** in the `api_base` field.
+- Now you can use the model with the interpreter set the model name to `local-model` and run the interpreter.</br>
 
 4. Run the interpreter with Python:</br>
 ### Running with Python.

libs/interpreter_lib.py

@@ -103,6 +103,11 @@ def initialize_client(self):
         self.INTERPRETER_MODEL = str(self.config_values.get('HF_MODEL', self.INTERPRETER_MODEL))       
         hf_model_name = self.INTERPRETER_MODEL.strip().split("/")[-1]
 
+        # skip init client for local models.(Bug#10 https://github.com/haseeb-heaven/code-interpreter/issues/10)
+        if 'local' in self.INTERPRETER_MODEL:
+            self.logger.info(f"Skipping client initialization for local model.")
+            return
+
         self.logger.info(f"Using model {hf_model_name}")
 
         model_api_keys = {