Added online model.

py/api.py

@@ -1,10 +1,20 @@
 import requests
 import json
-
+from transformers import AutoTokenizer, LlamaForCausalLM
 
 class API:
     @staticmethod
-    def process_text(prompt, model):
+    def process_text_transformers(prompt, model):
+        model = LlamaForCausalLM.from_pretrained(model)
+        tokenizer = AutoTokenizer.from_pretrained(model)
+
+        inputs = tokenizer(prompt, return_tensors="pt")
+
+        generate_ids = model.generate(inputs.input_ids, max_length=30)
+        return tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
+
+    @staticmethod
+    def process_text_local(prompt, model):
         ollama_url = "http://localhost:11434"
 
         response = requests.post(
@@ -26,13 +36,15 @@ class API:
 
     def send_message(self, message, model):
         if model == 1:
-            answer = self.process_text(message, "phi3.5")
+            answer = self.process_text_local(message, "phi3.5")
         elif model == 2:
-            answer = self.process_text(message, "gemma2:2b")
+            answer = self.process_text_local(message, "gemma2:2b")
         elif model == 3:
-            answer = self.process_text(message, "qwen2:0.5b")
+            answer = self.process_text_local(message, "qwen2:0.5b")
         elif model == 4:
-            answer = self.process_text(message, "codegemma:2b")
+            answer = self.process_text_local(message, "codegemma:2b")
+        elif model == 5:
+            answer = self.process_text_transformers(message, "meta-llama/Meta-Llama-3.1-8B")
         else:
             return "Invalid choice"
         return answer