#InsTag: Instruction Tagging for Analyzing Supervised Fine-tuning of Large Language Models
Paper • 2308.07074 • Published
InsTagger
InsTagger is an tool for automatically providing instruction tags by distilling tagging results from InsTag.
InsTag aims analyzing supervised fine-tuning (SFT) data in LLM aligning with human preference. For local tagging deployment, we release InsTagger, fine-tuned on InsTag results, to tag the queries in SFT data. Through the scope of tags, we sample a 6K subset of open-resourced SFT data to fine-tune LLaMA and LLaMA-2 and the fine-tuned models TagLM-13B-v1.0 and TagLM-13B-v2.0 outperform many open-resourced LLMs on MT-Bench.
Model Description
- Model type: Auto-regressive Models
- Language(s) (NLP): English
- License: apache-2.0
- Finetuned from model: LLaMa-2
Model Sources [optional]
- Repository: https://github.com/OFA-Sys/InsTag
- Paper: Arxiv
- Demo: ModelScope Demo
Uses
This model is directly developed with FastChat. So it can be easily infer or serve with FastChat selecting the vicuna template.
- Downloads last month
- 905
Install from pip and serve model
# Install vLLM from pip: pip install vllm# Start the vLLM server: vllm serve "OFA-Sys/InsTagger"# Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "OFA-Sys/InsTagger", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'