inference: false
longchat-13b-16k Model Card
Usage
Please use load_model from FastChat or LongChat repo to load the model (or chatting API from FastChat). There is a monkey patch needed to use the model. Usage referece:
(LongChat) python3 eval.py --model-name-or-path lmsys/longchat-13b-16k --task topics
(FastChat) python3 -m fastchat.serve.cli --model-path lmsys/longchat-13b-16k
Under the hood, the monkey patch is added in:
Model details
Model type: longchat-13b-16k is an open-source chatbot trained by fine-tuning llama-13b on user-shared conversations collected from ShareGPT, using the condensing rotary embedding technique reported in the blog.
Model date: longchat-13b-16k was trained on June 2023.
Organizations developing the model: The LongChat developers: Dacheng Li*, Rulin Shao*, Anze Xie, Ying Sheng, Lianmin Zheng, Ion Stoica, Xuezhe Ma, and Hao Zhang
Paper or resources for more information: https://github.com/DachengLi1/LongChat
Where to send questions or comments about the model: https://github.com/DachengLi1/LongChat
Intended use
Primary intended uses: The primary use of longchat-13b-16k is for research purposes.
Primary intended users: The primary intended users of the model are researchers in natural language processing, machine learning, and artificial intelligence.
Training dataset
18K conversations collected from ShareGPT.com.
Evaluation dataset
A preliminary evaluation of the model quality is conducted by our released LongEval.
Open LLM Leaderboard Evaluation Results
Detailed results can be found here
| Metric | Value |
|---|---|
| Avg. | 46.32 |
| ARC (25-shot) | 53.58 |
| HellaSwag (10-shot) | 77.67 |
| MMLU (5-shot) | 45.24 |
| TruthfulQA (0-shot) | 47.07 |
| Winogrande (5-shot) | 70.09 |
| GSM8K (5-shot) | 4.17 |
| DROP (3-shot) | 26.42 |