Modelpublic

princeton-nlp/Llama-3-8B-ProLong-64k-Instruct

sourceHugging Facellama3updated 2y agoView on Hugging Face

13likes8.6kdownloads

Model Card

princeton_nlp/Llama-3-8B-ProLong-64k-Instruct

ProLong (Princeton long-context language models) is a family of long-context models that are continued trained and supervised fine-tuned from Llama-3-8B, with a maximum context window of 512K tokens. Our main ProLong model is one of the best-performing long-context models at the 10B scale (evaluated by HELMET).

To train this strong long-context model, we conduct thorough ablations on the long-context pre-training data, SFT data, and numerous other design choices. We demonstrate our findings in our paper, How to Train Long-Context Language Models (Effectively).

Authors: Tianyu Gao\, [Alexander Wettig](https://www.cs.princeton.edu/~awettig/)\, Howard Yen, Danqi Chen (* equal contribution)

Contact: {tianyug, awettig}@princeton.edu

The ProLong Models

—princeton_nlp/Llama-3-8B-ProLong-64k-Base
—princeton_nlp/Llama-3-8B-ProLong-64k-Instruct ← you are here!
—princeton_nlp/Llama-3-8B-ProLong-512k-Base
—⭐ princeton_nlp/Llama-3-8B-ProLong-512k-Instruct

Model card

Here are some quick facts about our main ProLong model: princeton-nlp/Llama-3-8B-ProLong-512k-Instruct.

—Base model: meta-llama/Meta-Llama-3-8B-Instruct
—Long-context continued training: 20B tokens on 64K training data (princeton-nlp/prolong-data-64K), and 20B tokens on 512K training data (princeton-nlp/prolong-data-512K)
—Supervised fine-tuning (SFT): UltraChat
—Maximum context window: 512K tokens

<img width="80%" alt="image" src="https://github.com/user-attachments/assets/c31c9671-49fe-4776-91d2-de70ffd9f9a1"> ProLong performance on <a href="https://github.com/princeton-nlp/helmet">HELMET</a> averaged over 32K, 64K, and 128K lengths. All models are instruct models.

<img width="80%" alt="image" src="https://github.com/user-attachments/assets/a36a7d0f-4480-4a29-80f3-208477707fb7"> ProLong training recipe.

Citation

bibtex

@article{gao2024prolong,
  title={How to Train Long-Context Language Models (Effectively)},
  author={Gao, Tianyu and Wettig, Alexander and Yen, Howard and Chen, Danqi},
  journal={arXiv preprint arXiv:2410.02660},
  year={2024}
}