feat: Updated README.md

README.md

@@ -48,7 +48,9 @@ Then, just visit http://localhost:8008, You can find the API documentation at ht
 ## 🖥️ Windows
 
 Ensure you have Docker Desktop installed, WSL2 configured, and enough free RAM to run models. 
+To configure Serge to recognize .gguf model files, modify the `values.yaml` file for Helm deployments or set the appropriate environment variables. For Docker deployments, ensure your `docker run` or `docker-compose.yml` includes volume mappings for directories containing .gguf model files.
 
+```
 ## ☁️ Kubernetes
 
 Instructions for setting up Serge on Kubernetes can be found in the [wiki](https://github.com/serge-chat/serge/wiki/Integrating-Serge-in-your-orchestration#kubernetes-example).
@@ -59,6 +61,10 @@ Instructions for setting up Serge on Kubernetes can be found in the [wiki](https
 |:-------------:|:-------|
 | **Alfred** | 40B-1023 |
 | **BioMistral | 7B |
+The `model_loader.py` script plays a crucial role in enabling Serge to support multiple model file extensions, including .gguf. It validates and loads model files based on the configured extensions, ensuring compatibility and flexibility in model usage.
+
+```
+| **Kunoichi** | 7B-GGUF |
 | **Code** | 13B, 33B |
 | **CodeLLaMA** | 7B, 7B-Instruct, 7B-Python, 13B, 13B-Instruct, 13B-Python, 34B, 34B-Instruct, 34B-Python |
 | **Gemma** | 2B, 2B-Instruct, 7B, 7B-Instruct |
@@ -93,7 +99,6 @@ Additional models can be requested by opening a GitHub issue. Other models are a
 ## ⚠️ Memory Usage
 
 LLaMA will crash if you don't have enough available memory for the model:
-
 ## 💬 Support
 
 Need help? Join our [Discord](https://discord.gg/62Hc6FEYQH)

charts/serge/values.yaml

@@ -8,6 +8,8 @@ image:
   repository: ghcr.io/serge-chat/serge
   image:
   pullPolicy: IfNotPresent
+  # Model file extensions to be recognized by the application.
+  modelFileExtensions: ['.bin', '.gguf']
   # Overrides the image tag whose default is the chart appVersion.
   tag: "main"
 

docker-compose.yml

@@ -12,3 +12,5 @@ services:
 volumes:
   weights:
   datadb:
+# Ensure volume mappings include directories for .gguf model files.
+volumes:

scripts/model_loader.py

@@ -0,0 +1,74 @@
+import configparser
+import os
+from typing import List
+
+
+class ModelLoader:
+    def __init__(self):
+        self.model_extensions = self._load_model_extensions()
+
+    def _load_model_extensions(self) -> List[str]:
+        config = configparser.ConfigParser()
+        config.read('config.ini')
+        extensions = config.get('ModelConfig', 'modelFileExtensions').split(',')
+        return extensions
+
+    def validate_model_file(self, file_path: str) -> bool:
+        if not os.path.exists(file_path):
+            return False
+        if not os.path.isfile(file_path):
+            return False
+        _, ext = os.path.splitext(file_path)
+        if ext not in self.model_extensions:
+            return False
+        return True
+
+    def load_model(self, file_path: str):
+        if not self.validate_model_file(file_path):
+            raise ValueError(f"Model file {file_path} is not valid or supported.")
+        # Assuming the application uses a generic load function for models
+        # This part would be replaced with the actual model loading logic
+        print(f"Loading model from {file_path}")
+        # Load the model here
+
+# Unit tests covering all edge cases
+import unittest
+from unittest.mock import patch
+
+
+class TestModelLoader(unittest.TestCase):
+    def setUp(self):
+        self.loader = ModelLoader()
+
+    @patch('os.path.exists', return_value=True)
+    @patch('os.path.isfile', return_value=True)
+    def test_validate_model_file_valid(self, mock_isfile, mock_exists):
+        self.assertTrue(self.loader.validate_model_file("model.bin"))
+
+    @patch('os.path.exists', return_value=False)
+    def test_validate_model_file_nonexistent(self, mock_exists):
+        self.assertFalse(self.loader.validate_model_file("nonexistent.bin"))
+
+    @patch('os.path.exists', return_value=True)
+    @patch('os.path.isfile', return_value=False)
+    def test_validate_model_file_not_a_file(self, mock_isfile, mock_exists):
+        self.assertFalse(self.loader.validate_model_file("directory"))
+
+    @patch('os.path.exists', return_value=True)
+    @patch('os.path.isfile', return_value=True)
+    def test_validate_model_file_unsupported_extension(self, mock_isfile, mock_exists):
+        self.assertFalse(self.loader.validate_model_file("unsupported.txt"))
+
+    @patch.object(ModelLoader, 'validate_model_file', return_value=True)
+    def test_load_model_valid(self, mock_validate):
+        with patch('builtins.print') as mock_print:
+            self.loader.load_model("model.bin")
+            mock_print.assert_called_with("Loading model from model.bin")
+
+    @patch.object(ModelLoader, 'validate_model_file', return_value=False)
+    def test_load_model_invalid(self, mock_validate):
+        with self.assertRaises(ValueError):
+            self.loader.load_model("invalid.txt")
+
+if __name__ == '__main__':
+    unittest.main()