AIbase
EN
Home
Instagger
I

Instagger

Developed by OFA-Sys
InsTagger is a tool for automatically providing instruction tags. It achieves its function by extracting tag results from InsTag and is mainly used to analyze large language model supervised fine-tuning data consistent with human preferences.
Large Language Model
Transformers
 EnglishOpen Source License:Apache-2.0#Automatic generation of instruction tags#SFT data annotation#LLM preference analysis
Downloads 2,303
Release Time : 8/15/2023

Model Overview

InsTagger is a tool fine-tuned based on InsTag results, used to label queries in SFT data and supports local tag deployment.

Model Features

Automatic instruction tags
It can automatically extract tag results from InsTag and provide instruction tags for queries in SFT data.
Support for local deployment
Supports local tag deployment, facilitating users to perform localized processing on SFT data.
High-performance fine-tuned model
The fine-tuned model TagLM-13B performs better than many open-source large language models on MT-Bench.

Model Capabilities

Automatic generation of instruction tags
Analysis of supervised fine-tuning data
Local tag deployment

Use Cases

Natural language processing
SFT data annotation
Used to automatically label queries in supervised fine-tuning data.
Improve annotation efficiency and consistency.
Model fine-tuning
Fine-tune LLaMA and LLaMA-2 models based on tag results.
The fine-tuned models perform excellently on MT-Bench.
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
English简体中文繁體中文にほんご
© 2025AIbase