---
pipeline_tag: text-generation
tags:
- text-generation-inference
- backpack
- backpackmodel
library_name: transformers
license: apache-2.0
datasets:
- openwebtext
language:
- en
---


# Model Card for Backpack-GPT2

<!-- Provide a quick summary of what the model is/does. [Optional] -->
The Backpack-GPT2 language model is an instance of the [Backpack architecture](https://arxiv.org/abs/2305.16765), intended to combine strong modeling performance with an interface for interpretability and control.
Most details about this model and its training should be accessed in the paper, [Backpack Language Models](https://arxiv.org/abs/2305.16765).

See also [backpackmodels.science](backpackmodels.science).

![A depiction of the Backpack language modeling process, in which each word in the sequence is weighted and summed to predict each word in context.](http://backpackmodels.science/assets/backpack-process.gif)

##  Table of Contents

- [Model Card for Backpack-GPT2](#model-card-for--model_id-)
- [Table of Contents](#table-of-contents)
- [Model Details](#model-details)
  - [Model Description](#model-description)
- [Uses](#uses)
- [Bias, Risks, and Limitations](#bias-risks-and-limitations)
- [Training Details](#training-details)
  - [Training Data](#training-data)
  - [Training Procedure](#training-procedure)
- [Environmental Impact](#environmental-impact)
- [Technical Specifications [optional]](#technical-specifications-optional)
  - [Model Architecture and Objective](#model-architecture-and-objective)
  - [Compute Infrastructure](#compute-infrastructure)
    - [Hardware](#hardware)
    - [Software](#software)
- [Citation](#citation)
- [Model Card Authors [optional]](#model-card-authors-optional)
- [Model Card Contact](#model-card-contact)
- [How to Get Started with the Model](#how-to-get-started-with-the-model)


## Model Details

### Model Description

<!-- Provide a longer summary of what this model is/does. -->
The Backpack-GPT2 is a [Backpack-based language model](https://arxiv.org/abs/2305.16765), an architecture intended to combine strong modeling performance with an interface for interpretability and control.

- **Developed by:**  John Hewitt, John Thickstun, Christopher D. Manning, Percy Liang
- **Shared by [Optional]:** More information needed
- **Model type:** Language model
- **Language(s) (NLP):** en
- **License:** apache-2.0
- **Resources for more information:**
    - [GitHub Repo](https://github.com/john-hewitt/backpacks-flash-attn)
    - [Associated Paper](https://huggingface.co/datasets/openwebtext)

## Uses

This model is intended for use in the study and development of increasingly interpretable methods in natural language processing.
It is not directly fit for any production use.


## Bias, Risks, and Limitations

<!-- This section is meant to convey both technical and sociotechnical limitations. -->

Significant research has explored bias and fairness issues with language models (see, e.g., [Sheng et al. (2021)](https://aclanthology.org/2021.acl-long.330.pdf) and [Bender et al. (2021)](https://dl.acm.org/doi/pdf/10.1145/3442188.3445922)). Predictions generated by the model may include disturbing and harmful stereotypes across protected classes; identity characteristics; and sensitive, social, and occupational groups.
This model in particular is limited in its capabilities, and with a brand new architecture, less is known about its biases than, e.g., Transformer-based models.

## How to Get Started with the Model

```python
import torch
from transformers import AutoConfig, AutoModelForCausalLM

model_id = "stanfordnlp/backpack-gpt2"
config = AutoConfig.from_pretrained(model_id, trust_remote_code=True)
torch_model = AutoModelForCausalLM.from_pretrained(model_id, config=config, trust_remote_code=True)
torch_model.eval()

input = torch.randint(0, 50264, (1, 512), dtype=torch.long)
torch_out = torch_model(
    input,
    position_ids=None,
)
torch_out = torch.nn.functional.softmax(torch_out.logits, dim=-1)
print(torch_out)
```

## Training Details

### Training Data

<!-- This should link to a Data Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->

This model was trained on the [OpenWebText](https://huggingface.co/datasets/openwebtext) corpus.


### Training Procedure

This model was trained for 100k gradient steps with a batch size of 512k tokens and a linearly decaying learning rate from 6e-4 to zero, with a linear warmup of 5k steps.

### Environmental Impact

- **Hardware Type:** 4 A100 GPUs (40G)
- **Hours used:** Roughly 4 days.
- **Cloud Provider:** Stanford compute.
- **Compute Region:** Stanford energy grid.

### Model Architecture and Objective

This model was trained to minimize the cross-entropy loss, and is a [Backpack language model](https://arxiv.org/pdf/2305.16765.pdf).

### Compute Infrastructure

This model was trained on a slurm cluster.

### Hardware

This model was trained on 4 A100s.

### Software

This model was trained with [FlashAttention](https://github.com/HazyResearch/flash-attention) and [PyTorch](https://pytorch.org/)

## Citation

**BibTeX:**

```
@InProceedings{hewitt2023backpack,
  author =      "Hewitt, John and Thickstun, John and Manning, Christopher D. and Liang, Percy",
  title =       "Backpack Language Models",
  booktitle =   "Proceedings of the Association for Computational Linguistics",
  year =        "2023",
  publisher =   "Association for Computational Linguistics",
  location =    "Toronto, Canada",
}
```


## Model Card Authors [optional]

<!-- This section provides another layer of transparency and accountability. Whose views is this model card representing? How many voices were included in its construction? Etc. -->

John Hewitt

## Model Card Contact

johnhew@cs.stanford.edu