bump defaults, updated lite, fixed glm4.7 autoguess template

koboldcpp.py

@@ -57,6 +57,7 @@ savestate_limit_default = 5
 savestate_limit = 0 #savestate slots start at 0, only set when load model
 default_vae_tile_threshold = 768
 default_native_ctx = 16384
+default_genlen = 1024
 overridekv_max = 4
 default_autofit_padding = 1024
 lora_filenames_max = 4
@@ -69,7 +70,7 @@ dry_seq_break_max = 128
 extra_images_max = 4 # for kontext/qwen img
 
 # global vars
-KcppVersion = "1.108.2"
+KcppVersion = "1.109"
 showdebug = True
 kcpp_instance = None #global running instance
 global_memory = {"tunnel_url": "", "restart_target":"", "input_to_exit":False, "load_complete":False, "restart_override_config_target":""}
@@ -5587,7 +5588,7 @@ def show_gui():
     jinja_tools_var = ctk.IntVar(value=0)
     moeexperts_var = ctk.StringVar(value=str(-1))
     moecpu_var = ctk.StringVar(value=str(0))
-    defaultgenamt_var = ctk.StringVar(value=str(896))
+    defaultgenamt_var = ctk.StringVar(value=str(default_genlen))
     genlimit_var = ctk.StringVar(value=str(0))
     nobostoken_var = ctk.IntVar(value=0)
     override_kv_var = ctk.StringVar(value="")
@@ -6647,7 +6648,7 @@ def show_gui():
             args.overridenativecontext = 0
         args.moeexperts = int(moeexperts_var.get()) if moeexperts_var.get()!="" else -1
         args.moecpu = int(moecpu_var.get()) if moecpu_var.get()!="" else 0
-        args.defaultgenamt = int(defaultgenamt_var.get()) if defaultgenamt_var.get()!="" else 896
+        args.defaultgenamt = int(defaultgenamt_var.get()) if defaultgenamt_var.get()!="" else default_genlen
         args.genlimit = int(genlimit_var.get()) if genlimit_var.get()!="" else 0
         args.nobostoken = (nobostoken_var.get()==1)
         args.jinja = (jinja_var.get()==1)
@@ -9002,7 +9003,7 @@ if __name__ == '__main__':
     advparser.add_argument("--nomodel", help="Allows you to launch the GUI alone, without selecting any model.", action='store_true')
     advparser.add_argument("--moeexperts", metavar=('[num of experts]'), help="How many experts to use for MoE models (default=follow gguf)", type=int, default=-1)
     advparser.add_argument("--moecpu","--n-cpu-moe", "-ncmoe", metavar=('[layers affected]'), help="Keep the Mixture of Experts (MoE) weights of the first N layers in the CPU. If no value is provided, applies to all layers.", nargs='?', const=999, type=int, default=0)
-    advparser.add_argument("--defaultgenamt", help="How many tokens to generate by default, if not specified. Must be smaller than context size. Usually, your frontend GUI will override this.", type=check_range(int,64,8192), default=896)
+    advparser.add_argument("--defaultgenamt", help="How many tokens to generate by default, if not specified. Must be smaller than context size. Usually, your frontend GUI will override this.", type=check_range(int,64,8192), default=default_genlen)
     advparser.add_argument("--nobostoken", help="Prevents BOS token from being added at the start of any prompt. Usually NOT recommended for most models.", action='store_true')
     advparser.add_argument("--enableguidance", help="Enables the use of Classifier-Free-Guidance, which allows the use of negative prompts. Has performance and memory impact.", action='store_true')
     advparser.add_argument("--maxrequestsize", metavar=('[size in MB]'), help="Specify a max request payload size. Any requests to the server larger than this size will be dropped. Do not change if unsure.", type=int, default=32)