Text Generation
Transformers
PyTorch
English
llama
conversational
text-generation-inference
Inference Endpoints
tulu-2-dpo-7b / README.md
hamishivi's picture
Update README.md
f05fb21
|
raw
history blame
6.76 kB
metadata
model-index:
  - name: tulu-2-dpo-7b
    results: []
datasets:
  - HuggingFaceH4/ultrafeedback_binarized
  - allenai/tulu-v2-sft-mixture
language:
  - en
base_model: meta-llama/Llama-2-7b-hf
TuluV2 banner

Model Card for Tulu V2 DPO 7B

Tulu is a series of language models that are trained to act as helpful assistants. Tulu V2 DPO 7B is a fine-tuned version of Llama 2 that was trained on on a mix of publicly available, synthetic and human datasets using Direct Preference Optimization (DPO). This model is a strong alternative to Llama 2 7b Chat.

Model description

  • Model type: A model belonging to a suite of instruction and RLHF tuned chat models on a mix of publicly available, synthetic and human-created datasets.
  • Language(s) (NLP): Primarily English
  • License: AI2 ImpACT Low-risk license.
  • Finetuned from model: meta-llama/Llama-2-7b-hf

Model Sources

Performance

Model Size Alignment MT-Bench (score) AlpacaEval (win rate %)
Tulu-v2-7b 🐪 7B dDPO 6.30 73.9
Tulu-v2-dpo-7b 🐪 7B dDPO 6.29 85.1
StableLM-Tuned-α 7B dSFT 2.75 -
MPT-Chat 7B dSFT 5.42 -
Xwin-LMv0.1 7B dPPO 6.19 87.83
Mistral-Instructv0.1 7B - 6.84 -
Zephyr-7b-α 7B dDPO 6.88 -
Zephyr-7b-β 🪁 7B dDPO 7.34 90.60
Tulu-v2-13b 🐪 13B dDPO 6.70 78.9
Tulu-v2-dpo-13b 🐪 13B dDPO 7.00 89.5
Falcon-Instruct 40B dSFT 5.17 45.71
Guanaco 65B SFT 6.41 71.80
Llama2-Chat 70B RLHF 6.86 92.66
Vicuna v1.3 33B dSFT 7.12 88.99
WizardLM v1.0 70B dSFT 7.71 -
Xwin-LM v0.1 70B dPPO - 95.57
Tulu-v2-70b 🐪 70B dDPO 7.49 86.6
Tulu-v2-dpo-70b 🐪 70B dDPO 7.89 95.1
GPT-3.5-turbo - RLHF 7.94 89.37
Claude 2 - RLHF 8.06 91.36
GPT-4 - RLHF 8.99 95.28

Input Format

The model is trained to use the following format (note the newlines):

<|user|>
Your message here!
<|assistant|>

For best results, format all inputs in this manner. Make sure to include a newline after <|assistant|>, this can affect generation quality quite a bit.

Intended uses & limitations

The model was initially fine-tuned on a filtered and preprocessed of the Tulu V2 mix dataset, which contains a diverse range of human created instructions and synthetic dialogues generated primarily by other LLMs. We then further aligned the model with a Jax DPO trainer built on EasyLM on the openbmb/UltraFeedback dataset, which contains 64k prompts and model completions that are ranked by GPT-4.

Bias, Risks, and Limitations

The Tulu models have not been aligned to generate safe completions within the RLHF phase or deployed with in-the-loop filtering of responses like ChatGPT, so the model can produce problematic outputs (especially when prompted to do so). It is also unknown what the size and composition of the corpus was used to train the base Llama 2 models, however it is likely to have included a mix of Web data and technical sources like books and code. See the Falcon 180B model card for an example of this.

Training hyperparameters

The following hyperparameters were used during DPO training:

  • learning_rate: 5e-07
  • total_train_batch_size: 32
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 3.0

Citation

If you find Tulu 2 is useful in your work, please cite it with:

@misc{ivison2023changing,
   title={Camels in a Changing Climate: Enhancing LM Adaptation with Tulu 2}, 
   author={Hamish Ivison and Yizhong Wang and Valentina Pyatkin and Nathan Lambert and Matthew Peters and Pradeep Dasigi and Joel Jang and David Wadden and Noah A. Smith and Iz Beltagy and Hannaneh Hajishirzi},
   year={2023},
   archivePrefix={arXiv},
   primaryClass={cs.CL}
}

Model card adapted from Zephyr Beta