explicit buffer allocation from python

expose.cpp

@@ -33,7 +33,7 @@ extern "C" {
     };
     struct generation_outputs
     {
-        int status;
+        int status = -1;
         char text[16384]; //16kb should be enough for any response
     };
 
@@ -63,7 +63,7 @@ extern "C" {
         return true;
     }
 
-    generation_outputs generate(const generation_inputs inputs, generation_outputs output)
+    generation_outputs generate(const generation_inputs inputs, generation_outputs & output)
     {
         api_params.prompt = inputs.prompt;
         api_params.seed = inputs.seed;

llama_for_kobold.py

@@ -48,7 +48,7 @@ def load_model(model_filename,batch_size=8,max_context_length=512,threads=4,n_pa
 
 def generate(prompt,max_length=20,temperature=0.8,top_k=100,top_p=0.85,rep_pen=1.1,rep_pen_range=128,seed=-1,reset_state=True):
     inputs = generation_inputs()
-    outputs = generation_outputs()
+    outputs = ctypes.create_unicode_buffer(ctypes.sizeof(generation_outputs))
     inputs.prompt = prompt.encode("UTF-8")
     inputs.max_length = max_length
     inputs.temperature = temperature