vivo-ai
/

BlueLM-7B-Chat-32K-AWQ

+模型许可协议/Model License Agreement
+1. 定义
+本协议项下的模型，是指vivo公司（维沃移动通信有限公司）为开发者学习和非商业用途之目的，公开发布的免费模型。如需商业使用，不适用本协议，请另外联系vivo公司以获取授权。
+2. 许可授予
+根据本许可的条款和条件，特此授予您非排他性、全球性、不可转让、不可再许可、可撤销、免费的版权许可。
+上述版权声明和本许可声明应包含在本模型的所有副本或重要部分中。
+3.限制
+您不得出于任何非法目的复制、修改、使用、发布本模型的全部或部分衍生作品。
+未经您所在国家或地区（如必要的审查或备案）的流程性许可，您不得将本模型用于任何需要许可的场合。
+4.免责声明
+本模型“按原样”提供，基于技术的原因，我们不提供任何明示或暗示的保证，包括但不限于对安全性、稳定性、适销性、特定用途的适用性和非侵权性的保证，我们也不对本模型及依据本模型输出、生成的内容承担任何形式的责任。
+我们也可能在没有通知和提前的情况下，基于各种原因，随时修改、下架本模型。您不应依赖本模型实施相关行为。
+5. 投诉反馈
+如您发现本模型存在违法或者不妥当处，请联系我们，我们将尽快处理。
+6.争议解决
+本协议的订立、效力、解释、履行、修改和终止，使用本模型以及争议的解决均适用中华人民共和国大陆地区（仅为本协议之目的，不包括香港、澳门和台湾）法律，并排除冲突法的适用。如产生诉讼纠纷，由中国广东省东莞市第二人民法院管辖。
+1. Definitions
+The model under this Agreement refers to the free model released publicly by vivo (vivo Mobile Communication Co., Ltd.) for the purpose of developer learning and non-commercial use. For commercial use, this agreement is not applicable. You are adviced to contact vivo for separate authorization.
+2 Grant of license
+Subject to the terms and conditions of this license, you are hereby granted a non-exclusive, worldwide, non-transferable, non-sublicensable, revocable, and free copyright license.
+The above copyright statement and this permission statement shall be included in all copies or important parts of this model.
+3. Restrictions
+You shall not copy, modify, use, or publish part of or all derivative works of this model for any illegal purpose.
+You shall not use this model in any situation that requires permission without obtaining procedural permission from your country or region (such as necessary review or filing).
+4. Disclaimer
+This model is provided "as is". For technical reasons, we do not provide any express or implied warranties, including but not limited to the warranties of security, stability, merchantability, fitness for a particular purpose and non-infringement. We also do not assume any form of responsibility for this model and the content output and generated based on this model.
+We may also modify or remove this model at any time for various reasons without advanced notice. You should not rely on this model to implement related behaviors.
+5. Complaints and feedback
+If you find that this model is illegal or inappropriate, please contact us and we will deal with it as soon as possible.
+6. Dispute settlement
+The formation, validity, interpretation, performance, modification and termination of this Agreement, the use of this model and the settlement of disputes shall be governed by the laws of the Chinese Mainland (excluding Hong Kong, Macao and Taiwan, for the purpose of this Agreement only), excluding application of conflict of laws. Any litigation or dispute shall be under the jurisdiction of the Dongguan No. 2 People's Court in Guangdong, China.

README.md CHANGED Viewed

@@ -1,3 +1,71 @@
 ---
 license: apache-2.0
 ---

 ---
 license: apache-2.0
+license: other
+language:
+- zh
+- en
 ---
+# BlueLM
+<p align="center">
+🖥 <a href="https://github.com/vivo-ai-lab/BlueLM" target="_blank">github</a>  • 📜 <a href="https://huggingface.co/vivo-ai/BlueLM-7B-Chat-32K-AWQ/blob/main/MODEL_LICENSE" target="_blank">LICENSE</a> • 🎯 <a href="https://developers.vivo.com/product/ai/bluelm" target="_blank">vivo Developers</a> • 🗨 <a href="https://github.com/vivo-ai-lab/BlueLM/blob/main/resources/wechat.png" target="_blank">WeChat</a>
+</p>
+## 模型介绍/Introduction
+BlueLM 是由 vivo AI 全球研究院自主研发的大规模预训练语言模型，本次发布包含 7B 基础模型和 7B 对话模型，同时我们开源了支持 **32K** 的长文本基础模型和对话模型。
+- **更大量的优质数据**：高质量语料库进行训练，规模达到了 **2.6 万亿** 的 token 数，该语料库包含中文、英文以及少量日韩数据。
+- **更优的效果**：其中 BlueLM-7B-Chat 在 **C-Eval** 和 **CMMLU** 上均取得领先结果，对比同尺寸开源模型中具有较强的竞争力。
+- **长文本支持**：BlueLM-7B-Base-32K 和 BlueLM-7B-Chat-32K 均支持 **32K** 长文本，在保持基础能力相当情况下，能够支持更长上下文理解。
+- **协议说明**：BlueLM 系列欢迎开发者进行学术研究和商业应用。
+BlueLM is a large-scale open-source language model independently developed by the vivo AI Lab. This release includes 2K and 32K context length versions for both Base and Chat models.
+- **High-quality Data**: BlueLM is trained on a high-quality data with 2.6 trillion tokens. Our train corpus mainly consists of Chinese and English data, with a small amount of Japanese and Korean data.
+- **Stronger Performance**: BlueLM-7B-Chat achieves a strong competitive performance in C-Eval and CMMLU benchmarks of the same size.
+- **Longer Context**: We have extended the context length of both BlueLM-7B-Base-32K and BlueLM-7B-Chat-32K models from 2K to 32K. The models can support longer context understanding while maintaining the same basic capabilities.
+- **Model License**: BlueLM weights are open for academic research and commercial use.
+本次发布基座模型下载链接见：
+The release versions and hugging face download links are listed in the table below:
+|     |          Base Model        |          Chat Model        |       4bits Quantized Chat Model        |
+|:---:|:--------------------:|:--------------------:|:--------------------------:|
+| 7B-2k  | [BlueLM-7B-Base](https://huggingface.co/vivo-ai/BlueLM-7B-Base)  | [BlueLM-7B-Chat](https://huggingface.co/vivo-ai/BlueLM-7B-Chat)  | [BlueLM-7B-Chat-4bits](https://huggingface.co/vivo-ai/BlueLM-7B-Chat-4bits)  |
+| 7B-32K | [BlueLM-7B-Base-32K](https://huggingface.co/vivo-ai/BlueLM-7B-Base-32K) | [BlueLM-7B-Chat-32K](https://huggingface.co/vivo-ai/BlueLM-7B-Chat-32K) | - |
+## 评测结果/Benchmark Results
+我们在 LongBench 评测集上对我们的 BlueLM-7B-Chat-32K 模型进行了测试，具体结果如下表所示：
+We tested our BlueLM-7B-Chat-32K  on the LongBench dataset and the results are shown in the table below:
+| Model                 | Average   | Summary  | Single-Doc QA | Multi-Doc QA  | Code  | Few-shot | Synthetic |
+|:----------------------|:-----|:---------|:--------------|:--------------|:------|:---------|:----------|
+| BlueLM-7B-Chat-32K    | 41.2 | 18.8     | 35.6          | 36.2          | 54.2  | 56.9     | 45.5      |
+## 推理部署/Inference and Deployment
+```python
+>>> import torch
+>>> from transformers import AutoModelForCausalLM, AutoTokenizer
+>>> tokenizer = AutoTokenizer.from_pretrained("vivo-ai/BlueLM-7B-Chat-32K-AWQ", trust_remote_code=True, use_fast=False)
+>>> model = AutoModelForCausalLM.from_pretrained("vivo-ai/BlueLM-7B-Chat-32K-AWQ", device_map="cuda:0", torch_dtype=torch.float16, trust_remote_code=True, low_cpu_mem_usage=True, use_cache=False)
+>>> model = model.eval()
+>>> inputs = tokenizer("[|Human|]:写一篇关于刘慈欣《三体》小说的读后感，1000字左右[|AI|]:", return_tensors="pt")
+>>> inputs = inputs.to("cuda:0")
+>>> pred = model.generate(**inputs, max_new_tokens=2048, repetition_penalty=1.1)
+>>> print(tokenizer.decode(pred.cpu()[0], skip_special_tokens=True))
+```
+更多使用说明，请参考我们的 [Github 仓库](https://github.com/vivo-ai-lab/BlueLM)。
+For more instructions, please refer to our [Github Repo](https://github.com/vivo-ai-lab/BlueLM).
+## 协议/License
+社区使用代码依照 [Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0) 协议开源，且使用 BlueLM 模型权重需要遵循 [vivo_BlueLM模型许可协议](https://huggingface.co/vivo-ai/BlueLM-7B-Chat-32K-AWQ/blob/main/MODEL_LICENSE)。
+Our code is licensed under the [Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0) and [Community License for BlueLM Model](https://huggingface.co/vivo-ai/BlueLM-7B-Chat-32K-AWQ/blob/main/MODEL_LICENSE).