File size: 15,873 Bytes
28dfa1c
 
b7bec6b
 
 
28dfa1c
 
d0c2b23
28dfa1c
f22a64e
 
abb39c0
f22a64e
 
abb39c0
f22a64e
 
abb39c0
f22a64e
 
abb39c0
 
 
 
f22a64e
abb39c0
 
 
 
f22a64e
 
731c46c
f436e2b
aba6083
f576567
f436e2b
 
7cff80c
a6038f2
7cff80c
 
a6038f2
7cff80c
 
 
a6038f2
7cff80c
 
28dfa1c
 
a1b5bd3
 
 
 
 
 
 
 
 
 
d7b7220
a1b5bd3
 
28dfa1c
d0d7824
dbdef5f
29e7be5
 
 
 
 
 
277e69e
e1f5660
 
 
 
 
 
 
 
 
b2afa7e
2359491
c033882
b2afa7e
2359491
29e7be5
47300a8
7a9887f
ddd7854
 
 
 
ec61a06
ddd7854
37b7280
277e69e
29e7be5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
277e69e
 
5bc7700
 
 
 
 
 
 
 
 
29e7be5
 
 
277e69e
29e7be5
277e69e
 
 
 
 
 
 
 
 
29e7be5
 
 
37b3542
2164897
65792a4
37b3542
277e69e
37b3542
277e69e
 
 
 
 
 
 
 
 
37b3542
c57d110
 
b7bec6b
c57d110
d0d7824
7472408
b7bec6b
7472408
f7fbd24
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f205daf
f7fbd24
f205daf
f7fbd24
c85462e
f7fbd24
 
 
8256038
29e7be5
277e69e
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
---
pipeline_tag: text-generation
license: apache-2.0
language:
- zh
---

# Model Card for Breeze-7B-Instruct-v0.1

Breeze-7B is a language model family that builds on top of [Mistral-7B](https://huggingface.co/mistralai/Mistral-7B-v0.1). 
By additionally pretraining Mistral 7B with 250GB of Traditional Chinese content, Breeze is specifically intended for Traditional Chinese use.

[Breeze-7B-Base](https://huggingface.co/MediaTek-Research/Breeze-7B-Base-v0.1) is the base model for the Breeze 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://huggingface.co/MediaTek-Research/Breeze-7B-Instruct-v0.1) derives from the base model Breeze-7B-Base and has 
undergone supervised fine-tuning with over 1 million instances, making the resulting model amenable to be used as-is for commonly seen tasks.

[Breeze-7B-Instruct-64k](https://huggingface.co/MediaTek-Research/Breeze-7B-Instruct-64k-v0.1) is a slightly modified version of 
Breeze-7B-Instruct to enable a 64k-token context length. Roughly speaking, that is equivalent to 88k Traditional Chinese characters.

The current release version of Breeze is v0.1.

Practicality-wise:
- Breeze expands the original vocabulary with additional 30,000 Traditional Chinese tokens. With the expanded vocabulary, everything else being equal, Breeze operates at twice the inference speed for Traditional Chinese to Mistral-7B and Llama 7B. [See [Inference Performance](#inference-performance).]
- Breeze-Instruct can be used as is for common tasks such as Q&A, RAG, multi-round chat, and summarization.
- In particular, Breeze-Instruct-64k can perform tasks at a document level, not a chapter level.

Performance-wise:
- Breeze demonstrates impressive performance in benchmarks for Traditional Chinese, when compared to similar sized open-source contemporaries such as Taiwan-LLM, QWen, and Yi. [See [Chat Model Performance](#chat-model-performance).]
- Breeze shows comparable results to Mistral-7B-Instruct-v0.1 on the MMLU and MT-Bench benchmarks. [See [Chat Model Performance](#chat-model-performance).]


*A project by the members (in alphabetical order): Chan-Jan Hsu 許湛然, Chang-Le Liu 劉昶樂, 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-Base-v0.1
  - Expanding the vocabulary dictionary size from 32k to 62k to better support Traditional Chinese
  - 8k tokens context length
- Breeze-7B-Instruct-v0.1
  - Expanding the vocabulary dictionary size from 32k to 62k to better support Traditional Chinese 
  - 8k tokens context length
  - Multi-turn dialogue (without special handling for harmfulness)
- Breeze-7B-Instruct-64k-v0.1
  - Expanding the vocabulary dictionary size from 32k to 62k to better support Traditional Chinese
  - 64k tokens context length
  - Multi-turn dialogue (without special handling for harmfulness)

## Model Details

- Breeze-7B-Base-v0.1
  - Finetuned from: [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1)
  - Model type: Causal decoder-only transformer language model
  - Language: English and Traditional Chinese (zh-tw)
- Breeze-7B-Instruct-v0.1
  - Finetuned from: [MediaTek-Research/Breeze-7B-Base-v0.1](https://huggingface.co/MediaTek-Research/Breeze-7B-Base-v0.1)
  - Model type: Causal decoder-only transformer language model
  - Language: English and Traditional Chinese (zh-tw)
- Breeze-7B-Instruct-64k-v0.1
  - Finetuned from: [MediaTek-Research/Breeze-7B-Instruct-v0.1](https://huggingface.co/MediaTek-Research/Breeze-7B-Instruct-v0.1)
  - Model type: Causal decoder-only transformer language model
  - Language: English and Traditional Chinese (zh-tw)

## Base Model Performance

**TMMLU+**, **DRCD**, and **Table** source from [MediaTek-Research/TCEval-v2](https://huggingface.co/datasets/MediaTek-Research/TCEval-v2).
[MediaTek-Research/TCEval-v2](https://huggingface.co/datasets/MediaTek-Research/TCEval-v2) derives from [TCEval-v1](https://github.com/mtkresearch/MR-Models/tree/main/TC-Eval)
 and [ikala/tmmluplus](https://huggingface.co/datasets/ikala/tmmluplus). **MMLU** sources from [hails/mmlu_no_train](https://huggingface.co/datasets/hails/mmlu_no_train).
 We use the code revised from [EleutherAI/lm-evaluation-harness](https://github.com/EleutherAI/lm-evaluation-harness) to evaluate **TMMLU+**, **DRCD**, **Table**, and **MMLU**. 


| Models                                       |        |↑ TMMLU+ (ACC) | DRCD (EM)   | Table (ACC) | MMLU (ACC) |
|----------------------------------------------|--------|--------------|-------------|-------------|------------|
|                                              |        |TC, Knowledge |TC, Reasoning|TC, Reasoning|EN, Knowledge|
|                                              |        | 5 shot       | 3 shot      | 5 shot      | 5 shot     |
| [Yi-34B](https://huggingface.co/01-ai/Yi-34B)| 34B    | 63.10        | 84.57       | 49.31  | 77.42      |
| [Qwen-14B](https://huggingface.co/01-ai/Qwen/Qwen-14B)| 14B    | 51.30        | 16.95 *     | 50.69  | 68.83      |
| [Yi-6B](https://huggingface.co/01-ai/Yi-6B) | 6B     | 49.63        | 76.61       | 34.72  | 65.35      |
| [Qwen-7B](https://huggingface.co/01-ai/Qwen/Qwen-7B)| 7B     | 42.84        | 0.0 *       | 39.58  | 61.00      |
| [**Breeze-7B-Base-v0.1**](https://huggingface.co/MediaTek-Research/Breeze-7B-Base-v0.1)       | 7B     | 40.35        | 81.13        | 28.47  | 61.63      |
| [Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1)| 7B     | 36.93        | 79.27        | 27.78 | 64.89      |


\* Few-shot learning cannot effectively guide the model to generate the proper answer.


## Chat Model Performance

**TMMLU+**, **DRCD**, **Table**, and **MT-Bench-tw** source from [MediaTek-Research/TCEval-v2](https://huggingface.co/datasets/MediaTek-Research/TCEval-v2).
[MediaTek-Research/TCEval-v2](https://huggingface.co/datasets/MediaTek-Research/TCEval-v2) derives from [TCEval-v1](https://github.com/mtkresearch/MR-Models/tree/main/TC-Eval)
 and [ikala/tmmluplus](https://huggingface.co/datasets/ikala/tmmluplus). **MMLU** sources from [hails/mmlu_no_train](https://huggingface.co/datasets/hails/mmlu_no_train).
 **MT-Bench** source from [lmsys/mt_bench_human_judgments](https://huggingface.co/datasets/lmsys/mt_bench_human_judgments).
 We use the code revised from [EleutherAI/lm-evaluation-harness](https://github.com/EleutherAI/lm-evaluation-harness) to evaluate **TMMLU+**, **DRCD**, **Table**, and **MMLU**. 
 We use the code revised from [fastchat llm_judge](https://github.com/lm-sys/FastChat/tree/main/fastchat/llm_judge) (GPT4 as judge) to evaluate **MT-Bench-tw** and **MT-Bench**.


| Models                                                                                                  |        |↑ MT-Bench-tw (Score)| TMMLU+ (ACC) | TMMLU+ (ACC) | DRCD (EM)   | Table (ACC) | MT-Bench (Score) | MMLU (ACC)  | MMLU (ACC)  | 
|---------------------------------------------------------------------------------------------------------|--------|--------------------|--------------|--------------|-------------|-------------|------------------|-------------|-------------|
|                                                                                                         |        |TC, Chat            |TC, Knowledge |TC, Knowledge |TC, Reasoning|TC, Reasoning|EN, Chat          |EN, Knowledge|EN, Knowledge|
|                                                                                                         |        |0 shot              | 0 shot       | 5 shot       | 3 shot      | 0 shot      |0 shot            |  0 shot     | 5 shot      | 
| [gpt-3.5-turbo](https://openai.com)                                                                     |        |7.1                 | 41.76        |              |             |             |7.9               |  70.00      |             |    
| [Yi-34B-Chat](https://huggingface.co/01-ai/Yi-34B-Chat)                                                 | 34B    |6.9                 | 54.87        |              |             | 36.81       |7.6               |   71.04     |             |    
| [Qwen-14B-Chat](https://huggingface.co/Qwen/Qwen-14B-Chat)                                              | 14B    |6.4                 | 48.41        |              |             | 41.67       |7.2               |    64.91    |             |    
| [**Breeze-7B-Instruct-v0.1**](https://huggingface.co/MediaTek-Research/Breeze-7B-Instruct-v0.1)         | 7B     |5.7                 | 41.61        |              |             | 45.83       |7.1               |    63.26    |             |    
| [**Breeze-7B-Instruct-64k-v0.1**](https://huggingface.co/MediaTek-Research/Breeze-7B-Instruct-64k-v0.1) | 7B     |5.5                 | 40.99        |              |             | 36.11       |7.1               |    63.68    |             |    
| [Qwen-7B-Chat](https://huggingface.co/Qwen/Qwen-7B-Chat)                                                | 7B     |5.4                 | 40.02        |              |             | 33.33       |6.2               |    55.94    |             |    
| [Yi-6B-Chat](https://huggingface.co/01-ai/Yi-6B-Chat)                                                   | 6B     |5.0                 | 44.79        |              |             | 25.69       |6.0               |    59.45    |             |    
| [Taiwan-LLM-13B-v2.0-chat](https://huggingface.co/yentinglin/Taiwan-LLM-13B-v2.0-chat)                  | 13B    |5.0                 | 29.47        |              |             | 23.61       |-*                |    50.50    |             |     
| [Taiwan-LLM-7B-v2.1-chat](https://huggingface.co/yentinglin/Taiwan-LLM-7B-v2.1-chat)                    | 7B     |4.2                 | 28.08        |              |             | 31.25       | -*               |    42.72    |             |    

\* Taiwan-LLM models responds to multi-turn questions (English) in Traditional Chinese.

**Category Score of MT-Bench-tw (0 shot)**

| Models                                              | STEM    |Extraction|Reasoning| Math   | Coding  | Roleplay| Writing |Humanities|↑ AVG   |
|-----------------------------------------------------|---------|---------|---------|---------|---------|---------|---------|---------|---------|
| gpt-3.5-turbo                                       |  7.8    |  6.1    |   5.1   |   6.4   |  6.2    |   8.7   |   7.4   |   9.3   |   7.1   |
| Yi-34B-Chat                                         |  9.0    |  4.8    |   5.7   |   4.0   |  4.7    |   8.5   |   8.7   |   9.8   |   6.9   |
| Qwen-14B-Chat                                       |  7.6    |  5.7    |   4.5   |   4.2   |  5.3    |   7.5   |   7.3   |   9.1   |   6.4   |
| **Breeze-7B-Instruct-v0.1**                         |  6.5    |  5.6    |   3.9   |   3.6   |  4.3    |   6.9   |   5.7   |   9.3   |   5.7   |
| **Breeze-7B-Instruct-64k-v0.1**                     |  6.1    |  5.3    |   3.7   |   2.9   |  4.2    |   7.0   |   6.7   |   8.3   |   5.5   |
| Qwen-7B-Chat                                        |  6.6    |  4.5    |   4.8   |   2.9   |  3.6    |   6.2   |   6.8   |   8.2   |   5.4   |
| Yi-6B-Chat                                          |  7.3    |  2.7    |   3.1   |   3.3   |  2.3    |   7.2   |   5.2   |   8.8   |   5.0   |
| Taiwan-LLM-13B-v2.0-chat                            |  6.1    |  3.4    |   4.1   |   2.3   |  3.1    |   7.4   |   6.6   |   6.8   |   5.0   |
| Taiwan-LLM-7B-v2.1-chat                             |  5.2    |  2.6    |   2.3   |   1.2   |  3.4    |   6.6   |   5.7   |   6.8   |   4.2   |

**Category ACC of TMMLU+ (0 shot)**   

| Model                                               | STEM         | Social Science | Humanities | Other      | ↑ AVG   |
|-----------------------------------------------------|--------------|----------------|------------|------------|---------|
| Yi-34B-Chat                                         | 47.65        | 64.25          | 52.73      | 54.91      | 54.87   |
| Qwen-14B-Chat                                       | 43.83        | 55.00          | 48.55      | 46.22      | 48.41   |
| Yi-6B-Chat                                          | 37.80        | 51.74          | 45.36      | 44.25      | 44.79   |
| gpt-3.5-turbo                                       | 41.56        | 46.72          | 36.73      | 42.03      | 41.76   |
| **Breeze-7B-Instruct-v0.1**                         | 37.41        | 46.81          | 42.06      | 40.16      | 41.61   |
| **Breeze-7B-Instruct-64k-v0.1**                     | 37.88        | 46.35          | 40.31      | 39.40      | 40.99   |
| Qwen-7B-Chat                                        | 35.44        | 46.22          | 38.35      | 40.06      | 40.02   |
| Taiwan-LLM-13B-v2.0-chat                            | 27.74        | 33.69          | 27.03      | 29.43      | 29.47   |
| Taiwan-LLM-7B-v2.1-chat                             | 25.58        | 31.76          | 27.36      | 27.61      | 28.08   |



## Inference Performance
In this test, we use the first 700 characters of the [web article](https://health.udn.com/health/story/5976/7699252?from=udn_ch1005_main_index) as the input and ask the model to write the same article again.
All inferences run on 2 RTX A6000 GPUs (using `vllm`, with a tensor-parallel size of 2).

| Models                                                             | ↓ Inference Time (sec)|Estimated Max Input Length (Char)|
|--------------------------------------------------------------------|-------------------|--------------------------|
| Yi-6B                                                              |   10.62  |   5.2k                |
| **Breeze-7B-Instruct-v0.1**                                        |  10.74  |    11.1k                 |
| **Breeze-7B-Instruct-64k-v0.1**                                    | 10.74       |  88.8k            |
| Qwen-7B                                                            |   10.86         |    9.8k                  |
| Qwen-14B                                                           |   18.89  |    9.8k                  |
| Mistral-7B-v0.1                                                    |  20.48   |    5.1k                 |
| Taiwan-LLM-7B-v2.1-base                                            |   26.26          |    2.2k                  |
| Taiwan-LLM-13B-v2.0-base                                           |   36.80          |    2.2k                  |
| Yi-34B                                                             |  43.71   |    4.5k                  |

## Long-context Performance

TBD

## Examples

TBD

## Use in Transformers

First install direct dependencies:
```
pip install transformers torch accelerate
```
If you want faster inference using flash-attention2, you need to install these dependencies:
```bash
pip install packaging ninja
pip install flash-attn
```
Then load the model in transformers:
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

model = AutoModelForCausalLM.from_pretrained(
    model="MediaTek-Research/Breeze-7B-Instruct-v0.1",
    device_map="auto",
    torch_dtype=torch.bfloat16,
    use_flash_attn_2=True # optional
)
```

The structure of the query template follows that of Mistral-7B-Instruct, as shown below.
```txt
<s> SYS_PROMPT   [INST] QUERY1 [/INST] RESPONSE1 [INST] QUERY2 [/INST]
```
where `SYS_PROMPT`, `QUERY1`, `RESPONSE1`, and `QUERY2` can be provided by the user.

The suggested default `SYS_PROMPT` is 
```txt
You are a helpful AI assistant built by MediaTek Research. The user you are helping speaks Traditional Chinese and comes from Taiwan.
```

## Citation

```
@article{breeze7b2024,
  title={},
  author={},
  journal={arXiv},
  year={2024}
}
```