jcai / breeze-7b-32k-instruct-v1_0

实用性方面
- Breeze-7B-Base 通过增加额外的30000个繁体中文词汇扩展了原始词汇表。扩展词汇后，在所有其他条件都相同的情况下，Breeze-7B 在处理繁体中文方面相对于 Mistral-7B 和 Llama 7B 的推理速度提高了一倍。[参见推理性能。]
- Breeze-7B-Instruct 可以直接用于常见任务，如问答、RAG、多轮对话和总结。
- Breeze-7B-32k-Instruct 可以在文档级别执行任务（对于中文，20 ~ 40页）。

项目成员（按字母顺序排列）：許湛然、廖峰挺、許博竣、陳宜昌和指导老师許大山。

特点

Breeze-7B-32k-Base-v1_0
- 将词汇字典的大小从32k扩展到62k，以更好地支持繁体中文
- 32k个标记的上下文长度
Breeze-7B-32k-Instruct-v1_0
- 将词汇字典的大小从32k扩展到62k，以更好地支持繁体中文
- 32k个标记的上下文长度
- 多轮对话（无针对有害性的特殊处理）

模型详情

Breeze-7B-32k-Base-v1_0
- 预训练来自: Breeze-7B-Base
- 模型类型：因果解码器仅变换语言模型
- 语言：英语和繁体中文（zh-tw）
Breeze-7B-32k-Instruct-v1_0
- 微调来自: Breeze-7B-32k-Base
- 模型类型：因果解码器仅变换语言模型
- 语言：英语和繁体中文（zh-tw）

长上下文性能

针插草丛性能

我们使用密钥检索任务来测试模型在给定序列中关注不同深度的能力。
模型需要在一个分散的文档中检索。一个关键字已被放在一个上下文中。
关键字位置被分入16个桶中，每个桶有20个测试案例。
Breeze-7B-32k-Base 以90%以上的准确率完成了任务，如图所示。

长-DRCD 性能

模型/性能(EM)	DRCD	DRCD-16k	DRCD-32k
Breeze-7B-32k-Instruct-v1_0	76.9	54.82	44.26
Breeze-7B-32k-Base-v1_0	79.73	69.68	61.55
Breeze-7B-Base-v1_0	80.61	21.79	15.29

短期基准性能

模型/性能(EM)	TMMLU+	MMLU	TABLE	MT-Bench-tw	MT-Bench
Breeze-7B-32k-Instruct-v1_0	41.37	61.34	34	5.8	7.4
Breeze-7B-Instruct-v1_0	42.67	62.73	39.58	6.0	7.4

引用

@article{MediaTek-Research2024breeze7b,
      title={Breeze-7B Technical Report}, 
      author={Chan-Jan Hsu and Chang-Le Liu and Feng-Ting Liao and Po-Chun Hsu and Yi-Chang Chen and Da-Shan Shiu},
      year={2024},
      eprint={2403.02712},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}

## Model Card for MediaTek Research Breeze-7B-32k-Instruct-v1_0

## Source: [Link](https://hugging-face.cn/MediaTek-Research/Breeze-7B-32k-Instruct-v1_0)

MediaTek Research Breeze-7B (hereinafter referred to as Breeze-7B) is a language model family that builds on top of [Mistral-7B](https://hugging-face.cn/mistralai/Mistral-7B-v0.1), specifically intended for Traditional Chinese use.

[Breeze-7B-Base](https://hugging-face.cn/MediaTek-Research/Breeze-7B-Base-v1_0) is the base model for the Breeze-7B series. 
It is suitable for use if you have substantial fine-tuning data to tune it for your specific use case.

[Breeze-7B-Instruct](https://hugging-face.cn/MediaTek-Research/Breeze-7B-Instruct-v1_0) derives from the base model Breeze-7B-Base, making the resulting model amenable to be used as-is for commonly seen tasks.

[Breeze-7B-32k-Base](https://hugging-face.cn/MediaTek-Research/Breeze-7B-32k-Base-v1_0) is extended from the base model with more data, base change, and the disabling of the sliding window. 
Roughly speaking, that is equivalent to 44k Traditional Chinese characters.

[Breeze-7B-32k-Instruct](https://hugging-face.cn/MediaTek-Research/Breeze-7B-32k-Instruct-v1_0) derives from the base model Breeze-7B-32k-Base, making the resulting model amenable to be used as-is for commonly seen tasks.

Practicality-wise:
- Breeze-7B-Base expands the original vocabulary with additional 30,000 Traditional Chinese tokens. With the expanded vocabulary, everything else being equal, Breeze-7B operates at twice the inference speed for Traditional Chinese to Mistral-7B and Llama 7B. [See [Inference Performance](#inference-performance).]
- Breeze-7B-Instruct can be used as is for common tasks such as Q&A, RAG, multi-round chat, and summarization.
- Breeze-7B-32k-Instruct can perform tasks at a document level (For Chinese, 20 ~ 40 pages).

*A project by the members (in alphabetical order): Chan-Jan Hsu 許湛然, Feng-Ting Liao 廖峰挺, Po-Chun Hsu 許博竣, Yi-Chang Chen 陳宜昌, and the supervisor Da-Shan Shiu 許大山.*

## Features

- Breeze-7B-32k-Base-v1_0
  - Expanding the vocabulary dictionary size from 32k to 62k to better support Traditional Chinese
  - 32k-token context length
 
- Breeze-7B-32k-Instruct-v1_0
  - Expanding the vocabulary dictionary size from 32k to 62k to better support Traditional Chinese
  - 32k-token context length
  - Multi-turn dialogue (without special handling for harmfulness)

## Model Details

- Breeze-7B-32k-Base-v1_0
  - Pretrained from: [Breeze-7B-Base](https://hugging-face.cn/MediaTek-Research/Breeze-7B-Base-v1_0)
  - Model type: Causal decoder-only transformer language model
  - Language: English and Traditional Chinese (zh-tw)
- Breeze-7B-32k-Instruct-v1_0
  - Finetuned from: [Breeze-7B-32k-Base](https://hugging-face.cn/MediaTek-Research/Breeze-7B-32k-Base-v1_0)
  - Model type: Causal decoder-only transformer language model
  - Language: English and Traditional Chinese (zh-tw)

## Long-context Performance

#### Needle-in-a-haystack Performance

We use the passkey retrieval task to test the model's ability to attend to different various depths in a given sequence.
A key in placed within a long context distracting document for the model to retrieve.
The key position is binned into 16 bins, and there are 20 testcases for each bin.
Breeze-7B-32k-Base clears the tasks with 90+% accuracy, shown in the figure below.
![Needle-in-a-haystack Performance](https://hugging-face.cn/MediaTek-Research/Breeze-7B-32k-Base-v1_0/resolve/main/needle-in-a-haystack-performance.png)

#### Long-DRCD Performance

| **Model/Performance(EM)** | **DRCD** | **DRCD-16k** | **DRCD-32k** |
|---------------------------|----------|--------------|--------------|
| **Breeze-7B-32k-Instruct-v1\_0** |  76.9   |   54.82   |  44.26     |       
| **Breeze-7B-32k-Base-v1\_0** | 79.73    | 69.68        | 61.55        |
| **Breeze-7B-Base-v1\_0**      | 80.61    | 21.79        | 15.29        |

#### Short-Benchmark Performance

| **Model/Performance(EM)** | **TMMLU+** | **MMLU** | **TABLE** | **MT-Bench-tw** | **MT-Bench** |
|---------------------------|----------|--------------|--------------|-----|-----|
| **Breeze-7B-32k-Instruct-v1\_0** | 41.37    |   61.34   |   34    | 5.8 | 7.4 |
| **Breeze-7B-Instruct-v1\_0**      |  42.67  |   62.73    |   39.58 | 6.0 | 7.4 |

## Citation

```
@article{MediaTek-Research2024breeze7b,
      title={Breeze-7B Technical Report}, 
      author={Chan-Jan Hsu and Chang-Le Liu and Feng-Ting Liao and Po-Chun Hsu and Yi-Chang Chen and Da-Shan Shiu},
      year={2024},
      eprint={2403.02712},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}
```

Paste, drop or click to upload images (.png, .jpeg, .jpg, .svg, .gif)