Breexe-8x7B is a language model family that builds on top of
Mixtral-8x7B
,
specifically intended for Traditional Chinese use.
Breexe-8x7B-Base is the base model for the Breexe-8x7B series. Breexe-8x7B-Base expands the original vocabulary with additional
30,000 Traditional Chinese tokens. With the expanded vocabulary, Breexe-8x7B operates at twice the inference speed for Traditional
Chinese to Mixtral-8x7B. [See
Inference Performance
.]
Breexe-8x7B-Instruct
derives from the base model Breexe-8x7B-Base,
making the resulting model amenable to be used as-is for commonly seen tasks, such as Q&A, RAG, multi-round chat, and summarization.
Breexe-8x7B-Instruct demonstrates impressive performance in benchmarks for Traditional Chinese and English, on par with OpenAI's gpt-3.5-turbo-1106.
[See
Chat Model Performance
.]
The current release version of Breexe-8x7B is v0.1.
The models were trained on Nvidia's Taipei-1. Special thanks for Nvidia's technical support.
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.
BreeXe API
We offer a trial API for business integration and academic benchmarking.
API service open time: 14:00 - 20:00 (from 2024/4/3 to 2024/5/3)
The API is in beta testing. If you are experiencing connectivity issues, please bear with us.
Free trial API key:
'eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJ1c2VyX2lkIjoyLCJ1c2VybmFtZSI6ImdlbmVyYWxfcHVibGljIiwibm90ZSI6ImdlbmVyYWwgcHVibGljIn0.kCp68nRw3RSh3jbMm8FvhG0NIkStflgI1wTHLviRPQE'
Get started here:
import time
from openai import OpenAI
API_KEY = <API_KEY_HERE>
BASE_URL = 'https://api-mtkresearch.com/v1'
MODEL_NAME = 'BreeXe-8x7B'
client = OpenAI(base_url=BASE_URL, api_key=API_KEY)
defask_breexe(messages):
completion = client.chat.completions.create(
model=MODEL_NAME,
messages=messages,
temperature=0.01,
top_p=0.01,
max_tokens=512
)
response = completion.choices[0].message.content
time.sleep(3) # due to a rate limit of 200 requests per 10 minutesreturn response
messages = [
{
'role': 'system',
'content': 'You are a helpful AI assistant built by MediaTek Research. The user you are helping speaks Traditional Chinese and comes from Taiwan.'
},
{
'role': 'user',
'content': '請問在量子物理領域中,疊加原理是什麼?它又是如何與量子糾纏現象相關聯的?'
}
]
response = ask_breexe(messages)
print(response)
Some examples are shown below.
(System Prompt:
You are a helpful AI assistant built by MediaTek Research. The user you are helping speaks Traditional Chinese and comes from Taiwan.
, Temperature: 0.01, Top-P: 0.01, Repetition-Penalty: 1.1)
1️⃣
修正以下英語句子中的錯誤
On the other words, making friends with smarter people can help people make money.
2️⃣
請將以下英語文本翻譯成繁體中文
Many studies have shown that time outdoors is good for our physical health. Three wilderness education experts explain why periods alone in nature also provide mental and spiritual benefits.
Breexe
1️⃣
修正以下英語句子中的錯誤:
在另一方面,與更聰明的人交朋友可以幫助人賺錢。
正確的句子是:"On the other hand, making friends with smarter people can help people make money."
假設有兩個已排序陣列 A 和 B,其長度分別為 m 和 n (m <= n)。我們將利用以下步驟來計算中位數:
1.找出兩個陣列的中間元素索引 i = (m + n - 1)/2 和 j = m + n/2。注意,如果 n 是奇數,則取整後的 i 和 j 可能相同;如果 n 是偶數,i 和 j 分別表示兩個中位數的位置。
2.比較 A[i-1] 和 B[j-1]、A[i] 和 B[j]。
3.如果 A[i-1] > B[j-1],則中位數位於 B[0..j-1] 區間,因此可以忽略 A[i..m] 區間。反之,如果 A[i] < B[j],則可以忽略 B[j+1..n] 區間。
4.重複上述步驟,直到剩下一個區間。這個區間包含兩個陣列的中位數。
5.最後,計算中位數。如果 n 是奇數,中位數就是該區間的中間元素;如果 n 是偶數,中位數就是該區間的前兩個元素之平均值。
以下是 Python 程式碼範例:
```
def find_median(A, B):
m, n = len(A), len(B)
if m > n:
A, B, m, n = B, A, n, m # Ensure m <= n
imin, imax = 0, m
while imin <= imax:
i = (imin + imax) // 2
j = (m + n + 1) // 2 - i
# If i is too small or too large, then adjust it accordingly
if i < m and B[j-1] > A[i]:
imin = i + 1
elif i > 0 and A[i-1] > B[j]:
imax = i - 1
else:
max_left_A = float('-inf') if i == 0 else A[i-1]
min_right_A = float('inf') if i == m else A[i]
max_left_B = float('-inf') if j == 0 else B[j-1]
min_right_B = float('inf') if j == n else B[j]
Expanding the vocabulary dictionary size from 32k to 62k to better support Traditional Chinese
8k-token context length
Multi-turn dialogue (without special handling for harmfulness)
Sparse mixture of experts (MoE)
Inference Performance
In this test, we use the first 700 characters of the
web article
as the input and ask the model to write the same article again.
All inferences run on 4 RTX A6000 GPUs (using
vllm
, with a tensor-parallel size of 4).
Breexe-8x7B-Instruct-v0_1 huggingface.co is an AI model on huggingface.co that provides Breexe-8x7B-Instruct-v0_1's model effect (), which can be used instantly with this MediaTek-Research Breexe-8x7B-Instruct-v0_1 model. huggingface.co supports a free trial of the Breexe-8x7B-Instruct-v0_1 model, and also provides paid use of the Breexe-8x7B-Instruct-v0_1. Support call Breexe-8x7B-Instruct-v0_1 model through api, including Node.js, Python, http.
Breexe-8x7B-Instruct-v0_1 huggingface.co is an online trial and call api platform, which integrates Breexe-8x7B-Instruct-v0_1's modeling effects, including api services, and provides a free online trial of Breexe-8x7B-Instruct-v0_1, you can try Breexe-8x7B-Instruct-v0_1 online for free by clicking the link below.
MediaTek-Research Breexe-8x7B-Instruct-v0_1 online free url in huggingface.co:
Breexe-8x7B-Instruct-v0_1 is an open source model from GitHub that offers a free installation service, and any user can find Breexe-8x7B-Instruct-v0_1 on GitHub to install. At the same time, huggingface.co provides the effect of Breexe-8x7B-Instruct-v0_1 install, users can directly use Breexe-8x7B-Instruct-v0_1 installed effect in huggingface.co for debugging and trial. It also supports api for free installation.
Breexe-8x7B-Instruct-v0_1 install url in huggingface.co: