README.md

---
license: apache-2.0
language:
- en
pipeline_tag: text-generation
base_model:
- allenai/OLMo-2-1124-13B-SFT
library_name: transformers
datasets:
- allenai/olmo-2-1124-13b-preference-mix
---

<img src="https://allenai.org/olmo/olmo-7b-animation.gif" alt="OLMo Logo" width="800" style="margin-left:'auto' margin-right:'auto' display:'block'"/>

# OLMo-2-1124-13B-DPO

OLMo-2 13B DPO November 2024 is finetuned variant of the [OLMo-2 13B November 2024](https://huggingface.co/allenai/OLMo2-13B-1124) model, which has undergone supervised finetuning on the [Tülu 3 dataset](https://huggingface.co/datasets/allenai/tulu-3-sft-mixture) and further DPO training on [this dataset](https://huggingface.co/datasets/allenai/olmo-2-1124-13b-preference-mix).
Tülu 3 is designed for state-of-the-art performance on a diversity of tasks in addition to chat, such as MATH, GSM8K, and IFEval.
Check out [the OLMo-2 paper](https://TODO) or [Tülu 3 paper](https://arxiv.org/abs/2411.15124) for more details!

OLMo is a series of **O**pen **L**anguage **Mo**dels designed to enable the science of language models. 
These models are trained on the Dolma dataset. We are releasing all code, checkpoints, logs (coming soon), and associated training details. 
The core models released in this batch include the following:


| **Stage**           | **OLMo-2 7B**                                                                                          | **OLMo-2 7B**                                                                                         |
|----------------------|----------------------------------------------------------------------------------------------------------|----------------------------------------------------------------------------------------------------------|
| **Base Model**       | [allenai/OLMo2-7B-1124](https://huggingface.co/allenai/OLMo2-7B-1124)                                | [allenai/OLMo-2-13B-1124](https://huggingface.co/allenai/OLMo-2-13B-1124)                             |
| **SFT**              | [allenai/OLMo-2-1124-7B-SFT](https://huggingface.co/allenai/OLMo-2-1124-7B-SFT)                | [allenai/OLMo-2-1124-13B-SFT](https://huggingface.co/allenai/OLMo-2-1124-13B-SFT)              |
| **DPO**              | [allenai/OLMo-2-1124-7B-DPO](https://huggingface.co/allenai/OLMo-2-1124-7B-DPO)                | [allenai/OLMo-2-1124-13B-DPO](https://huggingface.co/allenai/OLMo-2-1124-13B-DPO)              |
| **Final Models (RLVR)** | [allenai/OLMo-2-1124-7B-Instruct](https://huggingface.co/allenai/OLMo-2-1124-7B-Instruct)                        | [allenai/OLMo-2-1124-13B-Instruct](https://huggingface.co/allenai/OLMo-2-1124-13B-Instruct)                      |
| **Reward Model (RM)**| [allenai/OLMo-2-1124-7B-RM](https://huggingface.co/allenai/OLMo-2-1124-7B-RM)                                                     | (Same as 8B)                                                     |



## Model description

- **Model type:** A model trained on a mix of publicly available, synthetic and human-created datasets.
- **Language(s) (NLP):** Primarily English
- **License:** Apache 2.0
- **Finetuned from model:** allenai/OLMo-2-13B-1124-SFT

### Model Sources

- **Project Page:** https://allenai.org/olmo
- **Repositories:** 
    - Core repo (training, inference, fine-tuning etc.): https://github.com/allenai/OLMo
    - Evaluation code: https://github.com/allenai/olmes
    - Further fine-tuning code: https://github.com/allenai/open-instruct
- **Paper:** Coming soon! TODO
- **Demo:** https://playground.allenai.org/

## Using the model

### Loading with HuggingFace

To load the model with HuggingFace, use the following snippet:
```
from transformers import AutoModelForCausalLM

olmo_model = AutoModelForCausalLM.from_pretrained("allenai/OLMo-2-1124-13B-DPO")
```

### Chat template

The chat template for our models is formatted as:
```
<|endoftext|><|user|>\nHow are you doing?\n<|assistant|>\nI'm just a computer program, so I don't have feelings, but I'm functioning as expected. How can I assist you today?<|endoftext|>
```
Or with new lines expanded:
```
<|endoftext|><|user|>
How are you doing?
<|assistant|>
I'm just a computer program, so I don't have feelings, but I'm functioning as expected. How can I assist you today?<|endoftext|>
```
It is embedded within the tokenizer as well, for `tokenizer.apply_chat_template`.

### System prompt

In Ai2 demos, we use this system prompt by default:
```
You are OLMo 2, a helpful and harmless AI Assistant built by the Allen Institute for AI.
```
The model has not been trained with a specific system prompt in mind.

### Bias, Risks, and Limitations

The OLMo-2 models have limited safety training, but are not deployed automatically with in-the-loop filtering of responses like ChatGPT, so the model can produce problematic outputs (especially when prompted to do so). 
See the Falcon 180B model card for an example of this.


## Performance

TODO

## Hyperparameters

Note we use a length-normalized variant of DPO for training.

DPO:
- **Learning Rate**: 8E-7 (7B, 13B)
- **Beta**: 5
- **Effective Batch Size:** 128 (7B, 13B)
- **Max. Sequence Length:** 2048
- **Learning Rate Schedule:** Linear
- **LR Warmup Ratio:** 0.1
- **Num. Epochs:** 1

## License and use

OLMo-2 is licensed under the Apache 2.0 license.
OLMo-2 is intended for research and educational use.
For more information, please see our [Responsible Use Guidelines](https://allenai.org/responsible-use).

## Citation

If OLMo-2 or any of the related materials were helpful to your work, please cite:
```
TODO
```