usage: koboldcpp-nocuda.exe [-h] [--usecuda [[main GPU ID] [[main GPU ID]
                            ...]]] [--usevulkan [[Device IDs] [[Device IDs]
                            ...]]] [--usecpu] [--model [filenames]
                            [[filenames] ...]] [--config [filename]]
                            [--contextsize [256 to 524288]]
                            [--gpulayers [[GPU layers]]] [--host [ipaddr]]
                            [--launch] [--tensor_split [Ratios] [[Ratios]
                            ...]] [--threads [threads]] [--port [portnumber]]
                            [--version] [--agent] [--agentload [filename]]
                            [--analyze [filename]] [--autofit]
                            [--autofitpadding [padding in MB]]
                            [--batchsize {-1,16,32,64,128,256,512,1024,2048,4096}]
                            [--ubatchsize {-1,16,32,64,128,256,512,1024,2048,4096}]
                            [--benchmark [[filename]]]
                            [--blasthreads [threads]]
                            [--chatcompletionsadapter [filename]] [--cli]
                            [--debugmode [DEBUGMODE]]
                            [--defaultgenamt DEFAULTGENAMT]
                            [--device <dev1,dev2,..>]
                            [--downloaddir [directory]]
                            [--draftamount [tokens]]
                            [--draftgpulayers [layers]]
                            [--draftgpusplit [Ratios] [[Ratios] ...]]
                            [--draftmodel [filename]] [--enableguidance]
                            [--exportconfig [filename]]
                            [--exporttemplate [filename]] [--failsafe]
                            [--foreground]
                            [--gendefaults {"parameter":"value",...}]
                            [--gendefaultsoverwrite]
                            [--genlimit [token limit]] [--highpriority]
                            [--ignoremissing] [--jinja]
                            [--jinja_kwargs {"parameter":"value",...}]
                            [--jinjatemplate [filename]]
                            [--jinjathink {default,true,false}]
                            [--lora [lora_filename] [[lora_filename] ...]]
                            [--loramult [amount]] [--lowvram]
                            [--maingpu [Device ID]]
                            [--maxrequestsize [size in MB]]
                            [--mcpfile [mcp json file]] [--mmproj [filename]]
                            [--mmprojcpu] [--lazy-mode {auto,on,off}]
                            [--moe-cache-mib MiB]
                            [--moecpu [[layers affected]]]
                            [--ffncpu [[layers affected]]]
                            [--moeexperts [num of experts]] [--multiplayer]
                            [--nocertify] [--nofastforward]
                            [--noflashattention] [--noavx2] [--nobostoken]
                            [--nommq] [--nomodel] [--noshift] [--noswa]
                            [--onready [shell command]]
                            [--allow-config-onready]
                            [--overridekv [name=type:value]]
                            [--overridenativecontext [trained context]]
                            [--overridetensors [tensor name pattern=buffer type]]
                            [--parallelrequests [slots]]
                            [--paralleltotalcontext [tokens]]
                            [--password [API key]] [--preloadstory [savefile]]
                            [--prompt [prompt]]
                            [--quantkv [quantization level f16/bf16/q8_0/q5_1/q4_0]]
                            [--queuelimit limit] [--quiet]
                            [--ratelimit [seconds]]
                            [--reasoningeffort {default,none,low,medium,high,xhigh}]
                            [--remotetunnel] [--ropeconfig [rope-freq-scale]
                            [[rope-freq-base] ...]]
                            [--savedatafile [savefile]] [--singleinstance]
                            [--smartcache [limit]] [--smartcontext]
                            [--splitmode [split mode]] [--ssl [cert_pem]
                            [[key_pem] ...]] [--swapadding SWAPADDING]
                            [--unpack destination] [--usemtp] [--usemlock]
                            [--usemmap] [--usedirectio]
                            [--visionmaxres [max px]]
                            [--visionmaxtokens [tokens]]
                            [--visionmintokens [tokens]] [--websearch]
                            [--showgui | --skiplauncher]
                            [--hordemodelname [name]]
                            [--hordeworkername [name]] [--hordekey [apikey]]
                            [--hordemaxctx [amount]] [--hordegenlen [amount]]
                            [--sdaudiovae [filename]] [--sdclamped [[maxres]]]
                            [--sdclampedsoft [maxres]] [--sdclip1 [filename]]
                            [--sdclip2 [filename]]
                            [--sdclipdevice [Device ID]]
                            [--sdconvdirect {off,vaeonly,full}]
                            [--sdflashattention] [--sdlora [filename]
                            [[filename] ...]] [--sdloramult [amounts]
                            [[amounts] ...]] [--sdmaingpu [Device ID]]
                            [--sdmodel [filename]] [--sdoffloadcpu]
                            [--sdphotomaker [filename]]
                            [--sdquant [[quantization level 0/1/2]]]
                            [--sdllm [filename]] [--sdthreads [threads]]
                            [--sdtiledvae [maxres]] [--sdupscaler [filename]]
                            [--sdvae [filename] | --sdvaeauto]
                            [--sdvaedevice [Device ID]]
                            [--sdvramlimit [limit MB]]
                            [--whispermodel [filename]]
                            [--ttsmodel [filename]]
                            [--ttswavtokenizer [filename]] [--ttsgpu]
                            [--ttsmaxlen TTSMAXLEN] [--ttsthreads [threads]]
                            [--ttsdir [directory]] [--musicllm [filename]]
                            [--musicembeddings [filename]]
                            [--musicdiffusion [filename]]
                            [--musicvae [filename]] [--musiclowvram]
                            [--embeddingsmodel [filename]]
                            [--embeddingsmaxctx [amount]] [--embeddingsgpu]
                            [--rpcmode [disabled/connect/host]]
                            [--rpcport [portnumber]] [--rpchost [IP address]]
                            [--rpcdevice <dev1,dev2,..>]
                            [--rpccache <directory>]
                            [--rpctargets [remotehost1:port1,remotehost2:port2]]
                            [--admin] [--adminpassword [password]]
                            [--admindir [directory]]
                            [--adminunloadtimeout ADMINUNLOADTIMEOUT]
                            [--routermode] [--reqtimeout [seconds]]
                            [--autoswapmode]
                            [--autoswapthreshold AUTOSWAPTHRESHOLD]
                            [--baseconfig BASECONFIG]
                            [model_param] [port_param]
koboldcpp-nocuda.exe: error: argument --contextsize/--ctx-size/-c: must be within [256, 524288]