Update app.py

app.py

@@ -11,7 +11,7 @@ from peft import PeftModel
 
 # Constants
 MAX_MAX_NEW_TOKENS = 2048
-DEFAULT_MAX_NEW_TOKENS = 1024
+DEFAULT_MAX_NEW_TOKENS = 900
 MAX_INPUT_TOKEN_LENGTH = int(os.getenv("MAX_INPUT_TOKEN_LENGTH", "4096"))
 
 # Description and License Texts
@@ -54,7 +54,7 @@ class Story(Document):
 
 # Utility function for prompts
 def make_prompt(entry):
-    return  f"### Human: Don't repeat the assesments, limit to 1024 tokens ONLY {entry} ### Assistant:"
+    return  f"### Human:{entry} ### Assistant:"
     # f"TELL A STORY, RELATE TO COMPUTER SCIENCE, INCLUDE ASSESMENTS. MAKE IT REALISTIC AND AROUND 800 WORDS, END THE STORY WITH "THE END.": {entry}"
 
 def process_text(text):
@@ -68,9 +68,9 @@ def generate(
     message: str,
     chat_history: list[tuple[str, str]],
     max_new_tokens: int = DEFAULT_MAX_NEW_TOKENS,
-    temperature: float = 0.6,
+    temperature: float = 0.8,
     top_p: float = 0.7,
-    top_k: int = 20,
+    top_k: int = 30,
     repetition_penalty: float = 1.0,
 ) -> Iterator[str]:
     conversation = []