Update app.py

app.py

@@ -63,8 +63,13 @@ tokenizer = AutoTokenizer.from_pretrained("TinyLlama/TinyLlama-1.1B-Chat-v1.0")
 #    temperature=0.4,
 #    stream = True
   #  eos_token_id=terminators
-)
+#)
 # send additional parameters to model for generation
+terminators = [
+    tokenizer.eos_token_id, # End-of-Sequence Token that indicates where the model should consider the text sequence to be complete
+    tokenizer.convert_tokens_to_ids("<|eot_id|>") # Converts a token strings in a single/ sequence of integer id using the vocabulary
+]
+# indicates the end of a sequence
 
 #model = llama_cpp.Llama(model_path = tinyllama-1.1b-chat-v1.0.Q4_K_M.gguf AutoModelForCausalLM.from_pretrained(llm_model, model_file = "tinyllama-1.1b-chat-v1.0.Q4_K_M.gguf", model_type="llama", gpu_layers=0, config = generation_config)
 model = Llama.from_pretrained(
@@ -74,6 +79,7 @@ model = Llama.from_pretrained(
     temperature=0.75,
     max_tokens=500,
     top_p=0.95,
+    eos_token_id=terminators
     # callback_manager=callback_manager,
     # verbose=True,  # Verbose is required to pass to the callback manager
     )