text-generation-webui/models/config.yaml

.*:
  wbits: 'None'
  model_type: 'None'
  groupsize: 'None'
  pre_layer: 0
  mode: 'chat'
  skip_special_tokens: true
  custom_stopping_strings: ''
.*llama:
  model_type: 'llama'
.*gptq(?!u|arl|v2):
  wbits: 4
  groupsize: 128
.*(4bit|int4):
  wbits: 4
.*(3bit|int3):
  wbits: 3
.*(-2bit|_2bit|int2-):
  wbits: 2
.*(-1bit|_1bit|int1-):
  wbits: 1
.*(8bit|int8):
  wbits: 8
.*(-7bit|_7bit|int7-):
  wbits: 7
.*(-6bit|_6bit|int6-):
  wbits: 6
.*(-5bit|_5bit|int5-):
  wbits: 5
.*gptqv2:
  groupsize: 'None'
.*(-gr32-|-32g-|groupsize32):
  groupsize: 32
.*(-gr64-|-64g-|groupsize64):
  groupsize: 64
.*(gr128|128g|groupsize128):
  groupsize: 128
.*(gr1024|1024g|groupsize1024):
  groupsize: 1024
.*(oasst|stablelm-7b-sft-v7-epoch-3):
  mode: 'instruct'
  instruction_template: 'Open Assistant'
  skip_special_tokens: false
(?!.*v0)(?!.*1.1)(?!.*1_1)(?!.*stable).*vicuna:
  mode: 'instruct'
  instruction_template: 'Vicuna-v0'
.*vicuna.*v0:
  mode: 'instruct'
  instruction_template: 'Vicuna-v0'
.*vicuna.*(1.1|1_1):
  mode: 'instruct'
  instruction_template: 'Vicuna-v1.1'
.*stable.*vicuna:
  mode: 'instruct'
  instruction_template: 'StableVicuna'
.*alpaca:
  mode: 'instruct'
  instruction_template: 'Alpaca'
.*alpaca-native-4bit:
  mode: 'instruct'
  instruction_template: 'Alpaca'
  wbits: 4
  groupsize: 128
.*galactica:
  skip_special_tokens: false
.*dolly-v[0-9]-[0-9]*b:
  mode: 'instruct'
  instruction_template: 'Alpaca'
  skip_special_tokens: false
  custom_stopping_strings: '"### End"'
.*koala:
  mode: 'instruct'
  instruction_template: 'Koala'
.*chatglm:
  mode: 'instruct'
  instruction_template: 'ChatGLM'
.*metharme:
  mode: 'instruct'
  instruction_template: 'Metharme'
.*llava:
  mode: 'instruct'
  model_type: 'llama'
  instruction_template: 'LLaVA'
  custom_stopping_strings: '"\n###"'
.*raven:
  mode: 'instruct'
  instruction_template: 'RWKV-Raven'
.*moss-moon.*sft:
  mode: 'instruct'
  instruction_template: 'MOSS'
.*stablelm-tuned:
  mode: 'instruct'
  instruction_template: 'StableLM'
  truncation_length: 4096
  chat_prompt_size: 4096
  chat_prompt_size_max: 4096
.*stablelm-base:
  truncation_length: 4096
  chat_prompt_size: 4096
  chat_prompt_size_max: 4096
.*wizardlm:
  mode: 'instruct'
  model_type: 'llama'
  instruction_template: 'WizardLM'
.*galactica.*finetuned:
  mode: 'instruct'
  instruction_template: 'Galactica Finetuned'
.*galactica.*-v2:
  mode: 'instruct'
  instruction_template: 'Galactica v2'
(?!.*finetuned)(?!.*-v2).*galactica:
  mode: 'instruct'
  instruction_template: 'Galactica'
.*guanaco:
  mode: 'instruct'
  instruction_template: 'Guanaco non-chat'
.*baize:
  mode: 'instruct'
  instruction_template: 'Baize'
.*mpt-.*instruct:
  mode: 'instruct'
  instruction_template: 'Alpaca'
.*mpt-.*chat:
  mode: 'instruct'
  instruction_template: 'MPT-Chat'
(?!.*-flan-)(?!.*-t5-).*lamini-:
  mode: 'instruct'
  instruction_template: 'Alpaca'
.*incite.*chat:
  mode: 'instruct'
  instruction_template: 'INCITE-Chat'
.*incite.*instruct:
  mode: 'instruct'
  instruction_template: 'INCITE-Instruct'
.*pygmalion-7b:
  model_type: 'llama'
.*metharme-7b:
  model_type: 'llama'