Downloads · 30 days
0
Zhuhai-13B
emotibot-inc/Zhuhai-13B
Zhuhai-13B is a machine learning model from emotibot-inc. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
Hugging Face | GitHub | Model Scope | Emotibrain
Access
Public
Updated Aug 31, 2023
Repo size
—
Likes
1
Public
Repo makeup
Click a slice to open those files.
.png1.1 MB · 99%
At a glance
- Access
- Public
- Created
- Aug 30, 2023
- Updated
- Aug 31, 2023
- SHA
- 0ea9802e
Model card
From the Hugging Face model README
README
Zhuhai-13B
Hugging Face | GitHub | Model Scope | Emotibrain
模型介绍
"竹海-13B"是竹间智能继“竹海-7B”之后开发的一款拥大模语言模型，以下是“竹海-13B”的四个主要特点：
-
更大尺寸、更多数据：相比于“竹海-7B”，我们将参数量扩大到130亿，并在高质量语料上训练了1.2万亿tokens。Zhuhai-13B的上下文窗口长度为4096。
-
高效性能：基于Transformer结构，在大约1.2万亿tokens上训练出来的130亿参数模型，支持中英双语。
-
安全性：我们对“竹海-13B”进行了严格的安全控制和优化，确保其在实际应用中不会产生任何不适当或误导性的输出。通过精心设计和调整算法参数，“竹海-13B”可以有效地避免乱说话现象。
Model benchmark
中文评测 - CMMLU
Result
| Model 5-shot | STEM | Humanities | Social Science | Other | China-specific | Average |
|---|---|---|---|---|---|---|
| Multilingual-oriented | ||||||
| GPT4 | 65.23 | 72.11 | 72.06 | 74.79 | 66.12 | 70.95 |
| ChatGPT | 47.81 | 55.68 | 56.50 | 62.66 | 50.69 | 55.51 |
| Falcon-40B | 33.33 | 43.46 | 44.28 | 44.75 | 39.46 | 41.45 |
| LLaMA-65B | 34.47 | 40.24 | 41.55 | 42.88 | 37.00 | 39.80 |
| BLOOMZ-7B | 30.56 | 39.10 | 38.59 | 40.32 | 37.15 | 37.04 |
| Bactrian-LLaMA-13B | 27.52 | 32.47 | 32.27 | 35.77 | 31.56 | 31.88 |
| Chinese-oriented | ||||||
| Zhuzhi-6B | 40.30 | 48.08 | 46.72 | 47.41 | 45.51 | 45.60 |
| Zhuhai-13B | 42.39 | 61.57 | 60.48 | 58.57 | 55.68 | 55.74 |
| Baichuan-13B | 42.38 | 61.61 | 60.44 | 59.26 | 56.62 | 55.82 |
| ChatGLM2-6B | 42.55 | 50.98 | 50.99 | 50.80 | 48.37 | 48.80 |
| Baichuan-7B | 35.25 | 48.07 | 47.88 | 46.61 | 44.14 | 44.43 |
| ChatGLM-6B | 32.35 | 39.22 | 39.65 | 38.62 | 37.70 | 37.48 |
| BatGPT-15B | 34.96 | 35.45 | 36.31 | 42.14 | 37.89 | 37.16 |
| Chinese-LLaMA-13B | 27.12 | 33.18 | 34.87 | 35.10 | 32.97 | 32.63 |
| MOSS-SFT-16B | 27.23 | 30.41 | 28.84 | 32.56 | 28.68 | 29.57 |
| Chinese-GLM-10B | 25.49 | 27.05 | 27.42 | 29.21 | 28.05 | 27.26 |
| Random | 25.00 | 25.00 | 25.00 | 25.00 | 25.00 | 25.00 |
| Model 0-shot | STEM | Humanities | Social Science | Other | China-specific | Average |
|---|---|---|---|---|---|---|
| Multilingual-oriented | ||||||
| GPT4 | 63.16 | 69.19 | 70.26 | 73.16 | 63.47 | 68.9 |
| ChatGPT | 44.8 | 53.61 | 54.22 | 59.95 | 49.74 | 53.22 |
| BLOOMZ-7B | 33.03 | 45.74 | 45.74 | 46.25 | 41.58 | 42.8 |
| Falcon-40B | 31.11 | 41.3 | 40.87 | 40.61 | 36.05 | 38.5 |
| LLaMA-65B | 31.09 | 34.45 | 36.05 | 37.94 | 32.89 | 34.88 |
| Bactrian-LLaMA-13B | 26.46 | 29.36 | 31.81 | 31.55 | 29.17 | 30.06 |
| Chinese-oriented | ||||||
| Zhuzhi-6B | 42.51 | 48.91 | 48.85 | 50.25 | 47.57 | 47.62 |
| Zhuhai-13B | 42.37 | 60.97 | 59.71 | 56.35 | 54.81 | 54.84 |
| Baichuan-13B | 42.04 | 60.49 | 59.55 | 56.6 | 55.72 | 54.63 |
| ChatGLM2-6B | 41.28 | 52.85 | 53.37 | 52.24 | 50.58 | 49.95 |
| Baichuan-7B | 32.79 | 44.43 | 46.78 | 44.79 | 43.11 | 42.33 |
| ChatGLM-6B | 32.22 | 42.91 | 44.81 | 42.6 | 41.93 | 40.79 |
| BatGPT-15B | 33.72 | 36.53 | 38.07 | 46.94 | 38.32 | 38.51 |
| Chinese-LLaMA-13B | 26.76 | 26.57 | 27.42 | 28.33 | 26.73 | 27.34 |
| MOSS-SFT-16B | 25.68 | 26.35 | 27.21 | 27.92 | 26.7 | 26.88 |
| Chinese-GLM-10B | 25.57 | 25.01 | 26.33 | 25.94 | 25.81 | 25.8 |
| Random | 25 | 25 | 25 | 25 | 25 | 25 |
推理对话
您可以直接注册并登录竹间智能科技发布的大模型产品 Emotibrain，并选择 CoPilot（KKBot） 进行的在线测试，注册即可立即使用；
模型训练
您可以直接注册并登录竹间智能科技发布的大模型产品 Emotibrain，并选择 Fine-tune 进行 0 代码微调，注册即可立即使用；
详细的训练流程您可以浏览此文档：Emotibrain 快速入门（大约 5 分钟）
更多信息
若您想了解更多 大模型训练平台 的相关信息，请访问 Emotibrain 官网 进行了解；
- Created
- Aug 30, 2023
- Updated
- Aug 31, 2023