text-generation-webui/models/config.yaml

231 lines
6.2 KiB
YAML
Raw Normal View History

.*(llama|alpac|vicuna|guanaco|koala|llava|wizardlm|metharme|pygmalion-7b|pygmalion-2|mythalion|wizard-mega|openbuddy|vigogne|h2ogpt-research|manticore):
model_type: 'llama'
2023-05-10 07:07:22 +02:00
.*(opt-|opt_|opt1|opt3|optfor|galactica|galpaca|pygmalion-350m):
model_type: 'opt'
.*(gpt-j|gptj|gpt4all-j|malion-6b|pygway|pygmalion-6b|dolly-v1):
2023-05-10 07:07:22 +02:00
model_type: 'gptj'
.*(gpt-neox|koalpaca-polyglot|polyglot.*koalpaca|polyglot-ko|polyglot_ko|pythia|stablelm|incite|dolly-v2|polycoder|h2ogpt-oig|h2ogpt-oasst1|h2ogpt-gm):
model_type: 'gptneox'
.*llama:
model_type: 'llama'
.*bloom:
model_type: 'bloom'
.*gpt2:
model_type: 'gpt2'
.*falcon:
model_type: 'falcon'
.*mpt:
model_type: 'mpt'
.*(starcoder|starchat):
model_type: 'starcoder'
.*dolly-v2:
model_type: 'dollyv2'
.*replit:
model_type: 'replit'
llama-65b-gptq-3bit:
groupsize: 'None'
.*(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
.*(-gr32-|-32g-|groupsize32|-32g$):
groupsize: 32
2023-05-12 11:11:58 +02:00
.*(-gr64-|-64g-|groupsize64|-64g$):
groupsize: 64
.*(gr128|128g|groupsize128):
groupsize: 128
.*(gr1024|1024g|groupsize1024):
groupsize: 1024
.*(oasst|openassistant-|stablelm-7b-sft-v7-epoch-3):
instruction_template: 'Open Assistant'
skip_special_tokens: false
(?!.*galactica)(?!.*reward).*openassistant:
instruction_template: 'Open Assistant'
skip_special_tokens: false
(?!.*v0)(?!.*1.1)(?!.*1_1)(?!.*stable)(?!.*chinese).*vicuna:
2023-04-26 21:20:27 +02:00
instruction_template: 'Vicuna-v0'
.*vicuna.*v0:
instruction_template: 'Vicuna-v0'
.*vicuna.*(1.1|1_1|1.3|1_3):
instruction_template: 'Vicuna-v1.1'
2023-08-10 18:42:24 +02:00
.*vicuna.*(1.5|1_5):
instruction_template: 'Vicuna-v1.1'
truncation_length: 4096
.*stable.*vicuna:
instruction_template: 'StableVicuna'
(?!.*chat).*chinese-vicuna:
instruction_template: 'Alpaca'
.*chinese-vicuna.*chat:
instruction_template: 'Chinese-Vicuna-Chat'
.*alpaca:
instruction_template: 'Alpaca'
2023-04-14 16:15:59 +02:00
.*alpaca-native-4bit:
instruction_template: 'Alpaca'
wbits: 4
groupsize: 128
.*galactica:
skip_special_tokens: false
.*dolly-v[0-9]-[0-9]*b:
instruction_template: 'Alpaca'
skip_special_tokens: false
custom_stopping_strings: '"### End"'
2023-04-16 19:40:45 +02:00
.*koala:
instruction_template: 'Koala'
.*chatglm:
instruction_template: 'ChatGLM'
.*(metharme|pygmalion|mythalion):
instruction_template: 'Metharme'
2023-04-24 01:32:22 +02:00
.*llava:
model_type: 'llama'
instruction_template: 'LLaVA'
2023-04-26 08:47:34 +02:00
custom_stopping_strings: '"\n###"'
2023-04-26 08:21:53 +02:00
.*raven:
instruction_template: 'RWKV-Raven'
.*ctx8192:
truncation_length: 8192
2023-05-05 04:19:23 +02:00
.*moss-moon.*sft:
instruction_template: 'MOSS'
.*stablelm-tuned:
instruction_template: 'StableLM'
truncation_length: 4096
.*stablelm-base:
truncation_length: 4096
.*galactica.*finetuned:
instruction_template: 'Galactica Finetuned'
.*galactica.*-v2:
instruction_template: 'Galactica v2'
(?!.*finetuned)(?!.*-v2).*galactica:
instruction_template: 'Galactica'
.*guanaco:
instruction_template: 'Guanaco non-chat'
.*baize:
instruction_template: 'Baize'
.*mpt-.*instruct:
instruction_template: 'Alpaca'
.*mpt-.*chat:
instruction_template: 'MPT-Chat'
(?!.*-flan-)(?!.*-t5-).*lamini-:
instruction_template: 'Alpaca'
.*incite.*chat:
instruction_template: 'INCITE-Chat'
.*incite.*instruct:
instruction_template: 'INCITE-Instruct'
.*wizard.*mega:
instruction_template: 'Wizard-Mega'
custom_stopping_strings: '"</s>"'
.*ziya-:
instruction_template: 'Ziya'
.*koalpaca:
instruction_template: 'KoAlpaca'
.*openbuddy:
instruction_template: 'OpenBuddy'
(?!.*chat).*vigogne:
instruction_template: 'Vigogne-Instruct'
.*vigogne.*chat:
instruction_template: 'Vigogne-Chat'
.*(llama-deus|supercot|llama-natural-instructions|open-llama-0.3t-7b-instruct-dolly-hhrlhf|open-llama-0.3t-7b-open-instruct):
instruction_template: 'Alpaca'
.*bactrian:
instruction_template: 'Bactrian'
.*(h2ogpt-oig-|h2ogpt-oasst1-|h2ogpt-research-oasst1-):
instruction_template: 'H2O-human_bot'
.*h2ogpt-gm-:
instruction_template: 'H2O-prompt_answer'
.*manticore:
instruction_template: 'Manticore Chat'
2023-06-01 19:42:12 +02:00
.*bluemoonrp-(30|13)b:
instruction_template: 'Bluemoon'
2023-06-05 17:07:52 +02:00
truncation_length: 4096
2023-06-05 16:52:13 +02:00
.*Nous-Hermes-13b:
instruction_template: 'Alpaca'
.*airoboros:
2023-06-05 16:52:13 +02:00
instruction_template: 'Vicuna-v1.1'
2023-07-17 16:27:42 +02:00
.*airoboros.*1.2:
instruction_template: 'Airoboros-v1.2'
.*alpa(cino|sta):
instruction_template: 'Alpaca'
.*hippogriff:
instruction_template: 'Hippogriff'
.*lazarus:
instruction_template: 'Alpaca'
.*guanaco-.*(7|13|33|65)b:
instruction_template: 'Guanaco'
.*hypermantis:
instruction_template: 'Alpaca'
.*open-llama-.*-open-instruct:
instruction_template: 'Alpaca'
.*starcoder-gpteacher-code-instruct:
instruction_template: 'Alpaca'
.*tulu:
instruction_template: 'Tulu'
.*chronos:
instruction_template: 'Alpaca'
.*samantha:
instruction_template: 'Samantha'
.*wizardcoder:
instruction_template: 'Alpaca'
.*starchat-beta:
2023-06-22 02:05:08 +02:00
instruction_template: 'Starchat-Beta'
custom_stopping_strings: '"<|end|>"'
2023-06-22 02:05:08 +02:00
.*minotaur:
instruction_template: 'Minotaur'
.*minotaur-15b:
2023-06-22 02:05:08 +02:00
truncation_length: 8192
2023-06-25 06:54:53 +02:00
.*orca_mini:
instruction_template: 'Orca Mini'
.*landmark:
truncation_length: 8192
.*superhot-8k:
truncation_length: 8192
2023-07-05 03:50:01 +02:00
.*xgen.*-inst:
2023-06-29 08:32:53 +02:00
truncation_length: 8192
2023-07-05 03:50:01 +02:00
instruction_template: 'Vicuna-v0'
.*(platypus|gplatty|superplatty):
instruction_template: 'Alpaca'
.*longchat:
instruction_template: 'Vicuna-v1.1'
.*vicuna-33b:
instruction_template: 'Vicuna-v1.1'
.*redmond-hermes-coder:
instruction_template: 'Alpaca'
truncation_length: 8192
.*wizardcoder-15b:
instruction_template: 'Alpaca'
truncation_length: 8192
.*wizardlm:
instruction_template: 'Vicuna-v1.1'
.*godzilla:
instruction_template: 'Alpaca'
.*llama-(2|v2):
truncation_length: 4096
.*llama(-?)(2|v2).*chat:
instruction_template: 'Llama-v2'
2023-08-01 22:27:29 +02:00
.*newhope:
instruction_template: 'NewHope'
2023-08-03 21:02:54 +02:00
.*stablebeluga2:
instruction_template: 'StableBeluga2'
truncation_length: 4096
2023-08-08 23:10:04 +02:00
.*openchat:
instruction_template: 'OpenChat'
.*falcon.*-instruct:
2023-08-18 00:03:08 +02:00
.*(openorca-platypus2):
instruction_template: 'OpenOrca-Platypus2'
custom_stopping_strings: '"### Instruction:", "### Response:"'
2023-08-25 16:06:57 +02:00
.*codellama:
rope_freq_base: 1000000
.*codellama.*instruct:
2023-09-29 00:16:28 +02:00
instruction_template: 'Llama-v2'
.*mistral.*instruct:
instruction_template: 'Mistral'