vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

ChatON: initial go at OrionStar Ai chat model template

Browse source 
Got from its tokenizer config json. Also same found in existing
hardcoded template in default chat apply template logic of llamacpp
This commit is contained in:
HanishKVC 2024-05-05 15:05:14 +05:30
parent 989c6c4125
commit 93115a9733
1 changed files with 30 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

32
examples/chaton_meta.json
View file

@ -320,6 +320,36 @@
"systemuser-system-has-end": true,
"systemuser-1st-user-has-begin": true,
"systemuser-1st-user-has-prefix": true
}
},
"orion": {
"global": {
"begin": "",
"end": ""
},
"system": {
"begin": "",
"prefix": "Human: ",
"suffix": "\n\n",
"end": ""
},
"user": {
"begin": "",
"prefix": "Human: ",
"suffix": "\n\nAssistant: </s>",
"end": ""
},
"assistant": {
"begin": "",
"prefix": "",
"suffix": "</s>",
"end": ""
},
"reverse-prompt": "</s>",
"systemuser-system-has-suffix": true,
"systemuser-system-has-end": true,
"systemuser-1st-user-has-begin": false,
"systemuser-1st-user-has-prefix": false
}
}