added backend which maintains ai instances;

added frontend as control panel

backend/ai/ais/AiBase.py

@@ -0,0 +1,51 @@
+import datetime
+import json
+import os
+import torch
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+
+class AiBase:
+    model_id_or_path: str = None
+    model = None
+    tokenizer = None
+
+    def __init__(self, model_id_or_path: str = None):
+        self.model_id_or_path = model_id_or_path
+
+        print('Loading ' + model_id_or_path)
+        self.model_id_or_path = model_id_or_path
+        self.model = AutoModelForCausalLM.from_pretrained(model_id_or_path, torch_dtype='auto', device_map='auto')
+        self.tokenizer = AutoTokenizer.from_pretrained(model_id_or_path)
+        # print(self.tokenizer.default_chat_template)
+        # print(type(self.model))
+        # print(type(self.tokenizer))
+        print('Loaded')
+
+    def generate(self, messages):
+        return []
+
+    def record_conversation(self, messages, response):
+        messages = messages + [response]
+
+        this_dir = os.path.dirname(os.path.abspath(__file__))
+        conversations_folder = this_dir + '/../../../conversations'
+        folder = conversations_folder + '/' + self.model_id_or_path.replace('/', '_')
+        self._mkdir(conversations_folder)
+        self._mkdir(folder)
+        timestamp = datetime.datetime.utcnow().strftime('%Y%m%d%H%M%S')
+        pickle_filename = folder + '/' + timestamp + '.json'
+        with open(pickle_filename, 'w') as file:
+            json.dump(messages, file)
+
+    def _mkdir(self, path):
+        if not os.path.isdir(path):
+            os.mkdir(path)
+
+    def __del__(self):
+        del self.model
+        del self.tokenizer
+
+        import gc
+        gc.collect()
+        torch.cuda.empty_cache()