README.md

---
library_name: peft
base_model: meta-llama/Llama-2-7b-chat-hf
---

# Model Card for Model ID

<!-- Provide a quick summary of what the model is/does. -->



## Model Details
We introduce the first text-based and multimodal LLMs capable of sign language processing called SignAlignLM, and propose new
prompting and fine-tuning strategies incorporating sign linguistic rules and conventions. We show that LLMs can be generalized interfaces
for both spoken and signed languages if trained with a multitasking paradigm.

### Tasks
As RWTH-PHOENIX-14T is a parallel corpus between spoken German and DGS, most previous research has focused on translation tasks between these languages. In this paper, we focus
on translating DGS to German (broadly considered as a sign understanding or recognition task) and German to DGS (broadly considered as sign
generation). In addition to these, we introduce additional tasks to test generalization. Specifically, we consider:
 - **(G2T) DGS Gloss to German Text:** a text-based
translation task from textual intermediary representations of DGS (glosses) to German text.
 - **(T2G) German Text to DGS Gloss:** the inverse
problem of the above and is text-based.
 - **(V2T) DGS Videos to German Text:** a multimodal task where the input is a video of a signer
signing in DGS, and the output is German text.
 - **(I-G2T) Intensified DGS Gloss to German Text:**
a text-based task with augmented DGS tokens.
Additional symbols <HIGH-INT> and <LOWINT> are wrapped around glosses to depict intensity in the video that is not depicted in traditional
gloss representations (Inan et al., 2022).
 - **(T2I-G) German Text to Intensified DGS Gloss:** the inverse problem of (I-G2T), still text-based.
 - **(G2E) DGS Gloss to English Text:** a novel task
of cross-modal translation, where DGS glosses from the German Sign Language family are translated to English text from the spoken IndoEuropean language family. Without any pretraining, this is a difficult test of generalization and composition of contextualized meanings
across spoken and signed languages.

To test generalizability and in-context learning, G2T is the only DGS task we use for any finetuning (see § 4.2). All the other tasks are used to
evaluate the models’ performance.

### Model Description

<!-- Provide a longer summary of what this model is. -->



- **Developed by:** Mert Inan
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** LLaMA adapter
- **Language(s) (NLP):** American Sign Language, German Sign Language, English, German
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** LLaMA-2 7B

### Model Sources [optional]

<!-- Provide the basic links for the model. -->

- **Repository:** https://github.com/Merterm/signAlignLM
- **Paper [optional]:** [SignAlignLM: Integrating Multimodal Sign Language Processing into Large Language Models](https://aclanthology.org/2025.findings-acl.190/)
- **Demo [optional]:** [More Information Needed]

## Uses

<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->

### Direct Use

<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->

[More Information Needed]

### Downstream Use [optional]

<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->

[More Information Needed]

### Out-of-Scope Use

<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->

[More Information Needed]

## Bias, Risks, and Limitations

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

[More Information Needed]

### Recommendations

<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->

Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.

## How to Get Started with the Model

Use the code below to get started with the model.

[More Information Needed]

## Training Details

### Training Data

<!-- This should link to a Dataset 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. -->

[More Information Needed]

### Training Procedure 

<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->

#### Preprocessing [optional]

[More Information Needed]


#### Training Hyperparameters

- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->

#### Speeds, Sizes, Times [optional]

<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->

[More Information Needed]

## Evaluation

<!-- This section describes the evaluation protocols and provides the results. -->

### Testing Data, Factors & Metrics

#### Testing Data

<!-- This should link to a Dataset Card if possible. -->

[More Information Needed]

#### Factors

<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->

[More Information Needed]

#### Metrics

<!-- These are the evaluation metrics being used, ideally with a description of why. -->

[More Information Needed]

### Results

[More Information Needed]

#### Summary



## Model Examination [optional]

<!-- Relevant interpretability work for the model goes here -->

[More Information Needed]

## Environmental Impact

<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->

Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).

- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]

## Technical Specifications [optional]

### Model Architecture and Objective

[More Information Needed]

### Compute Infrastructure

[More Information Needed]

#### Hardware

[More Information Needed]

#### Software

[More Information Needed]

## Citation [optional]

<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->

**BibTeX:**

~~~
@inproceedings{inan-etal-2025-signalignlm,
    title = "{S}ign{A}lign{LM}: Integrating Multimodal Sign Language Processing into Large Language Models",
    author = "Inan, Mert  and
      Sicilia, Anthony  and
      Alikhani, Malihe",
    editor = "Che, Wanxiang  and
      Nabende, Joyce  and
      Shutova, Ekaterina  and
      Pilehvar, Mohammad Taher",
    booktitle = "Findings of the Association for Computational Linguistics: ACL 2025",
    month = jul,
    year = "2025",
    address = "Vienna, Austria",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2025.findings-acl.190/",
    doi = "10.18653/v1/2025.findings-acl.190",
    pages = "3691--3706",
    ISBN = "979-8-89176-256-5",
    abstract = "Deaf and Hard-of-Hearing (DHH) users increasingly utilize Large Language Models (LLMs), yet face significant challenges due to these models' limited understanding of sign language grammar, multimodal sign inputs, and Deaf cultural contexts. Further, current approaches that try to address these limitations, frequently reduce sign language processing (SLP) to traditional translation tasks, neglecting the multimodal and linguistic complexity inherent in signed languages. In this paper, we present an empirical investigation informed by learning theory into natively integrating sign language support within LLMs, directly addressing the documented needs of DHH users. We introduce the first text-based and multimodal LLMs capable of sign language processing called SignAlignLM, and propose new prompting and fine-tuning strategies incorporating sign linguistic rules and conventions. We show that LLMs can be generalized interfaces for both spoken and signed languages if trained with a multitasking paradigm. Our code and model checkpoints are open-source."
}
~~~

**APA:**

[More Information Needed]

## Glossary [optional]

<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->

[More Information Needed]

## More Information [optional]

[More Information Needed]

## Model Card Authors [optional]

Mert Inan

## Model Card Contact

Mert Inan


### Framework versions

- PEFT 0.7.1