Add editable LLM penalties and preserve sampling through backup restore

chat_test.py

@@ -6,6 +6,7 @@ import threading
 import time
 import uuid
 from inference import InferenceError
+from profiles import generation_parameters
 
 class ChatTests:
  def __init__(self,profiles,worker,scheduler):
@@ -52,7 +53,7 @@ class ChatTests:
    with self.scheduler.lease(key,profile['parameters']['slots'],prepare,allowed=allowed):
     if self.cancel.is_set():raise InterruptedError()
     self.phase('Antwort wird erzeugt')
-    body=dict(model=profile['name'],messages=messages,max_tokens=limit,stream=True,stream_options={"include_usage":True},**{k:profile['parameters'][k] for k in ('temperature','top_p','top_k')})
+    body=dict(model=profile['name'],messages=messages,max_tokens=limit,stream=True,stream_options={"include_usage":True},**generation_parameters(profile['parameters']))
     conn,token=self.worker.connect()
     conn.request('POST','/v1/chat/completions',json.dumps(body).encode(),headers={'Content-Type':'application/json','Authorization':'Bearer '+token})
     with self.lock:self.socket=conn.sock