Update generate.py

generate.py

@@ -157,6 +157,16 @@ def generate(
     self.use_complex_talk_head = use_complex_talk_head
     self.use_weighted_talk_head = use_weighted_talk_head
 
+    # Set model properties
+    self.use_end_thought_token = True
+    self.use_start_thought_token = True
+    self.n_ahead = n_ahead
+    self.n_passes = 1
+    self.eval_mode = True
+    self.first_run = False
+    self.rm_initialized = True
+    self.original_mode = False
+
     # Generate using the custom generate function
     generated_token_ids = custom_generate(
         self,