pacozaa / bonito

This is an unofficial version of [Bonito](https://github.com/BatsResearch/bonito). I convert and quantize the model myself and extract the template from the [paper](https://arxiv.org/abs/2402.18334)

7B

29 Pulls Updated 4 months ago

4 months ago

21006918e2d2 · 4.4GB

model
llama
·
7.24B
·
Q4_K_M
template
<|tasktype|> {{ if .System }}{{ .System }}{{ end }} <|context|> {{ .Prompt }} <|task|>
params
{"stop":["<|task|>","<|context|>","\"<|tasktype|>\""]}

Readme

This is an unofficial version of Bonito.
I convert and quantize the model myself and extract the template from the paper

Check the current task type by using >>>/show system
Set support task types by using >>>/set system [support task type]

The support task types are

SHORTFORM_TO_FULL_TASK_TYPES = {
    "exqa": "extractive question answering",
    "mcqa": "multiple-choice question answering",
    "qg": "question generation",
    "qa": "question answering without choices",
    "ynqa": "yes-no question answering",
    "coref": "coreference resolution",
    "paraphrase": "paraphrase generation",
    "paraphrase_id": "paraphrase identification",
    "sent_comp": "sentence completion",
    "sentiment": "sentiment",
    "summarization": "summarization",
    "text_gen": "text generation",
    "topic_class": "topic classification",
    "wsd": "word sense disambiguation",
    "te": "textual entailment",
    "nli": "natural language inference",
}

shown here https://github.com/BatsResearch/bonito/blob/main/bonito/abstract.py#L5 at SHORTFORM_TO_FULL_TASK_TYPES variable.