---
language:
- ja
tags:
- heron
- vision
- image-captioning
- VQA
pipeline_tag: image-to-text
license:
- cc-by-nc-4.0
inference: false
---
# Heron GIT Japanese ELYZA Llama 2 Fast 7B

![heron](./heron_image.png)

## Model Details
Heron GIT Japanese ELYZA Llama 2 Fast 7B is a vision-language model that can converse about input images.<br>
This model was trained using [the heron library](https://github.com/turingmotors/heron). Please refer to the code for details.


## Usage

Follow [the installation guide](https://github.com/turingmotors/heron/#1-clone-this-repository). 

```python
import requests
from PIL import Image

import torch
from transformers import AutoProcessor
from heron.models.git_llm.git_gpt_neox import GitGPTNeoXForCausalLM

device_id = 0

# prepare a pretrained model
model = GitGPTNeoXForCausalLM.from_pretrained(
    'turing-motors/heron-chat-git-ELYZA-fast-7b-v0', torch_dtype=torch.float16
)
model.eval()
model.to(f"cuda:{device_id}")

# prepare a processor
processor = AutoProcessor.from_pretrained('turing-motors/heron-chat-git-ELYZA-fast-7b-v0')

# prepare inputs
url = "https://www.barnorama.com/wp-content/uploads/2016/12/03-Confusing-Pictures.jpg"
image = Image.open(requests.get(url, stream=True).raw)

text = f"##human: これは何の写真ですか？\n##gpt: "

# do preprocessing
inputs = processor(
    text,
    image,
    return_tensors="pt",
    truncation=True,
)
inputs = {k: v.to(f"cuda:{device_id}") for k, v in inputs.items()}

# set eos token
eos_token_id_list = [
    processor.tokenizer.pad_token_id,
    processor.tokenizer.eos_token_id,
]

# do inference
with torch.no_grad():
    out = model.generate(**inputs, max_length=256, do_sample=False, temperature=0., eos_token_id=eos_token_id_list)

# print result
print(processor.tokenizer.batch_decode(out)[0])
```


## Model Details
* **Developed by**: [Turing Inc.](https://www.turing-motors.com/)
* **Adaptor type**: [GIT](https://arxiv.org/abs/2205.14100)
* **Lamguage Model**: [ELYZA Japanese Llama-2 7B fast instruct](https://huggingface.co/elyza/ELYZA-japanese-Llama-2-7b-fast-instruct)
* **Language(s)**: Japanese

### Training
This model was initially trained with the Adaptor using STAIR Captions. In the second phase, it was fine-tuned with LLaVA-Instruct-150K-JA and Japanese Visual Genome using LoRA.

### Training Dataset

- [LLaVA-Instruct-150K-JA](https://huggingface.co/datasets/turing-motors/LLaVA-Instruct-150K-JA)
- [Japanese STAIR Captions](http://captions.stair.center/)
- [Japanese Visual Genome VQA dataset](https://github.com/yahoojapan/ja-vg-vqa)

## Use and Limitations

### Intended Use

This model is intended for use in chat-like applications and for research purposes.

### Limitations

The model may produce inaccurate or false information, and its accuracy is not guaranteed. It is still in the research and development stage.

## How to cite
```bibtex
@misc{GitElyzaFast, 
    url    = {[https://huggingface.co/turing-motors/heron-chat-git-ELYZA-fast-7b-v0](https://huggingface.co/turing-motors/heron-chat-git-ELYZA-fast-7b-v0)}, 
    title  = {Heron GIT Japanese ELYZA Llama 2 Fast 7B}, 
    author = {Yuichi Inoue, Kotaro Tanahashi, and Yu Yamaguchi}
}
```

## Citations

```bibtex
@misc{touvron2023llama,
      title={Llama 2: Open Foundation and Fine-Tuned Chat Models}, 
      author={Hugo Touvron and Louis Martin and Kevin Stone and Peter Albert and Amjad Almahairi and Yasmine Babaei and Nikolay Bashlykov and Soumya Batra and Prajjwal Bhargava and Shruti Bhosale and Dan Bikel and Lukas Blecher and Cristian Canton Ferrer and Moya Chen and Guillem Cucurull and David Esiobu and Jude Fernandes and Jeremy Fu and Wenyin Fu and Brian Fuller and Cynthia Gao and Vedanuj Goswami and Naman Goyal and Anthony Hartshorn and Saghar Hosseini and Rui Hou and Hakan Inan and Marcin Kardas and Viktor Kerkez and Madian Khabsa and Isabel Kloumann and Artem Korenev and Punit Singh Koura and Marie-Anne Lachaux and Thibaut Lavril and Jenya Lee and Diana Liskovich and Yinghai Lu and Yuning Mao and Xavier Martinet and Todor Mihaylov and Pushkar Mishra and Igor Molybog and Yixin Nie and Andrew Poulton and Jeremy Reizenstein and Rashi Rungta and Kalyan Saladi and Alan Schelten and Ruan Silva and Eric Michael Smith and Ranjan Subramanian and Xiaoqing Ellen Tan and Binh Tang and Ross Taylor and Adina Williams and Jian Xiang Kuan and Puxin Xu and Zheng Yan and Iliyan Zarov and Yuchen Zhang and Angela Fan and Melanie Kambadur and Sharan Narang and Aurelien Rodriguez and Robert Stojnic and Sergey Edunov and Thomas Scialom},
      year={2023},
      eprint={2307.09288},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}
```

---
license: cc-by-nc-4.0
---