From 7d0853de4674d95c9d5df9a36a3328b3776962f1 Mon Sep 17 00:00:00 2001
From: ftgreat <ftgreat@163.com>
Date: Wed, 19 Jul 2023 08:23:58 +0000
Subject: [PATCH 1/5] docs updated

Signed-off-by: ftgreat <ftgreat@163.com>
---
 examples/Aquila/Aquila-chat/README.md        | 16 +++++++--------
 examples/Aquila/Aquila-chat/README_en.md     | 10 +++++-----
 examples/Aquila/Aquila-code/README.md        | 21 +++++++++++---------
 examples/Aquila/Aquila-code/README_en.md     | 16 +++++++--------
 examples/Aquila/Aquila-code/generate_code.py |  7 ++++---
 examples/Aquila/Aquila-pretrain/README.md    | 16 +++++++--------
 examples/Aquila/Aquila-pretrain/README_en.md | 10 +++++-----
 examples/Aquila/README.md                    | 16 +++++++--------
 examples/Aquila/README_en.md                 | 10 +++++-----
 examples/Aquila/changelog.md                 |  5 +++++
 examples/Aquila/changelog_zh.md              |  5 +++++
 examples/glm_title_generation/data_load.py   |  2 +-
 flagai/auto_model/auto_loader.py             |  2 ++
 flagai/model/aquila_model.py                 |  1 -
 flagai/model/predictor/aquila.py             |  3 ++-
 15 files changed, 78 insertions(+), 62 deletions(-)
 mode change 100644 => 100755 examples/glm_title_generation/data_load.py

diff --git a/examples/Aquila/Aquila-chat/README.md b/examples/Aquila/Aquila-chat/README.md
index f15db90c..4c8226dc 100755
--- a/examples/Aquila/Aquila-chat/README.md
+++ b/examples/Aquila/Aquila-chat/README.md
@@ -21,20 +21,20 @@
 | :---------------- | :------- | :-- |:-- |   :-- | :-- | :-- | 
 | Aquila-7B         | 基础模型，70亿参数  |   **Aquila 基础模型**在技术上继承了 GPT-3、LLaMA 等的架构设计优点，替换了一批更高效的底层算子实现、重新设计实现了中英双语的 tokenizer，升级了 BMTrain 并行训练方法，实现了比 Magtron+DeepSpeed ZeRO-2 将近８倍的训练效率。   | [./examples/Aquila/Aquila-pretrain](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-pretrain)  | [下载Aquila-7B](http://model.baai.ac.cn/model-detail/100098) | 已发布 | Nvidia-A100 |
 | Aquila-33B          |基础模型，330亿参数  |    同上    | ——  | ——  | **敬请期待** | Nvidia-A100 | 
-| AquilaChat-7B          |SFT model，基于 Aquila-7B 进行微调和强化学习  |    **AquilaChat 对话模型**支持流畅的文本对话及多种语言类生成任务，通过定义可扩展的特殊指令规范，实现 AquilaChat对其它模型和工具的调用，且易于扩展。 <br><br>例如，调用智源开源的 **[AltDiffusion](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/AltDiffusion-m18) 多语言文图生成模型**，实现了流畅的文图生成能力。配合智源 **InstructFace 多步可控文生图模型**，轻松实现对人脸图像的多步可控编辑。  |   [./examples/Aquila/Aquila-chat](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-chat)  | [下载AquilaChat-7B](https://model.baai.ac.cn/model-detail/100101) | 已发布  | Nvidia-A100  | 
-| AquilaChat-33B           |SFT model，基于 Aquila-33B 进行微调和强化学习 |   同上    |   ——    |——  | **敬请期待** | Nvidia-A100 | 
-| AquilaCode-7B-NV          | 基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练，在英伟达芯片完成训练  |   AquilaCode-7B 以小数据集、小参数量，实现高性能，是目前支持中英双语的、性能最好的开源代码模型，经过了高质量过滤、使用有合规开源许可的训练代码数据进行训练。<br><br> AquilaCode-7B 分别在英伟达和国产芯片上完成了代码模型的训练。  | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  |[下载AquilaCode-7B-NV](https://model.baai.ac.cn/model-detail/100102)  | 已发布  | Nvidia-A100 | 
-| AquilaCode-7B-TS           |基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练，在天数智芯芯片上完成训练  |    同上    | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  | [下载AquilaCode-7B-TS](https://model.baai.ac.cn/model-detail/100099)  | 已发布  | Tianshu-BI-V100  | 
+| AquilaChat-7B          |SFT 模型，基于 Aquila-7B 进行微调和强化学习  |    **AquilaChat 对话模型**支持流畅的文本对话及多种语言类生成任务，通过定义可扩展的特殊指令规范，实现 AquilaChat对其它模型和工具的调用，且易于扩展。 <br><br>例如，调用智源开源的 **[AltDiffusion](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/AltDiffusion-m18) 多语言文图生成模型**，实现了流畅的文图生成能力。配合智源 **InstructFace 多步可控文生图模型**，轻松实现对人脸图像的多步可控编辑。  |   [./examples/Aquila/Aquila-chat](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-chat)  | [下载AquilaChat-7B](https://model.baai.ac.cn/model-detail/100101) | 已发布  | Nvidia-A100  | 
+| AquilaChat-33B           |SFT 模型，基于 Aquila-33B 进行微调和强化学习 |   同上    |   ——    |——  | **敬请期待** | Nvidia-A100 | 
+| AquilaCode-multi         | 基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练  |   AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型来定制自己的代码助手。  | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  |[下载AquilaCode-7B-multi](https://model.baai.ac.cn/model-detail/100102)  | 已发布  | Nvidia-A100 | 
+| AquilaCode-py           |基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练。  |    AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型定制自己的代码助手。    | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  | [下载AquilaCode-py](https://model.baai.ac.cn/model-detail/100099)  | 已发布  | Nvidia-A100  |   
 
 悟道·天鹰Aquila系列模型将持续开源更优版本，大家可以先删除原来目录下的`checkpoints_in/aquilachat-7b`，再下载新权重，其他使用方式不变。
 
-- 2023/07/13 ：发布权重文件 v0.8，开源了 Aquila-7B、AquilaChat-7B 最新权重，AquilaCode 权重无更新。
+- 2023/07/19 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
-  - AquilaCode-7B-NV md5：91115e72a7fc7f780b410696eae6259c
-  - AquilaCode-7B-TS md5：5dae2486bc5a885279be87c13872cd5c
+  - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
+  - AquilaCode-py md5：3faa85fc03d8fda70a73064f48d02d85
     
-其中AquilaChat-7B v0.8 在 FlagEval 大模型评测中（ “主观+客观”）相比0.7的版本整体稍有提，其在Chinese-MMLU上提升10%左右，详细评测结果请通过 http://flageval.baai.ac.cn 网站查看。历史版本变更详情见：[变更日志](https://github.com/FlagAI-Open/FlagAI/blob/master/examples/Aquila/changelog.md) 。
+其中AquilaChat-7B新版本模型在 FlagEval 大模型评测中（ “主观+客观”）相比0.7的版本整体稍有提，其在Chinese-MMLU上提升10%左右，详细评测结果请通过 http://flageval.baai.ac.cn 网站查看。历史版本变更详情见：[变更日志](https://github.com/FlagAI-Open/FlagAI/blob/master/examples/Aquila/changelog.md) 。
 
 
 <br>如有使用问题请先查看 [FAQ](https://github.com/FlagAI-Open/FlagAI/issues/371)，若不能解决，请直接提交 [issue](https://github.com/FlagAI-Open/FlagAI/issues) ~
diff --git a/examples/Aquila/Aquila-chat/README_en.md b/examples/Aquila/Aquila-chat/README_en.md
index 30dcbde6..8554b2e9 100755
--- a/examples/Aquila/Aquila-chat/README_en.md
+++ b/examples/Aquila/Aquila-chat/README_en.md
@@ -24,17 +24,17 @@ The additional details of the Aquila model will be presented in the official tec
 | Aquila-33B         | Base model, 33 billion parameters   | Same as above      | ——      | ——    | Coming soon                                               | Nvidia-A100   |
 | AquilaChat-7B      | SFT model, fine-tuned and RL based on Aquila-7B  | **AquilaChat Dialog Model** supports fluent text dialogue and multiple language generation tasks, and realizes the call of AquilaChat to other models and tools by defining an expandable special instruction specification, which is easy to extend. For example, calling the open source **[AltDiffusion](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/AltDiffusion-m18) multimodal language image generation model** of Flagship Intelligence achieved smooth image generation capability. Together with Flagship Intelligence's **InstructFace multi-step controllable text-picture model**, it is easy to achieve multi-step controllable editing of human face images. | [./examples/Aquila/Aquila-chat](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-chat) | [Download AquilaChat-7B](https://model.baai.ac.cn/model-detail/100101) | Released    | Nvidia-A100   |
 | AquilaChat-33B     | SFT model, fine-tuned and RL based on Aquila-33B  | Same as above| ——   | ——   |Coming soon                                               | Nvidia-A100   |
-| AquilaCode-7B-NV   | Base model, "text-code" generation model, further pre-trained based on Aquila-7B, trained on Nvidia  | AquilaCode-7B achieves high performance with small data sets and parameters, and is currently the best open source code model that supports both Chinese and English, trained using training code data with compliant open source licenses after high-quality filtering. AquilaCode-7B has been trained on both Nvidia and domestic chips for code models. | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-7B-NV](https://model.baai.ac.cn/model-detail/100102) | Released | Nvidia-A100  |
-| AquilaCode-7B-TS   | Base model, "text-code" generation model, further pre-trained based on Aquila-7B, trained on Horizon Robotics chips | Same as above | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-7B-TS](https://model.baai.ac.cn/model-detail/100099)  | Released        | Tianshu-BI-V100 |
+| AquilaCode-multi  | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B.  | AquilaCode utilizes high-quality, filtered, and compliant open-source code data for training, with a dataset size of approximately 10-40% compared to other open-source code generation models. By following the provided official guidelines, developers can harness the power of the AquilaCode model to customize their own code assistant. | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-multi](https://model.baai.ac.cn/model-detail/100102) | Released | Nvidia-A100  |
+| AquilaCode-py   | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B, trained on Horizon Robotics chips | Same as above | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-py](https://model.baai.ac.cn/model-detail/100099)  | Released        | Nvidia-A100 |
 
 
 We will continue to release improved versions of Aquila model as open source. You can start by deleting the `checkpoints_in/aquilachat-7b` in the original directory and then download the new weights. Other usage methods remain unchanged. For more details, please refer to the folloing change log:
 
-- 2023/07/13 ：Released v0.8 checkpoint files，The latest weights of Aquila-7B and AquilaChat-7B have been open sourced, but there are no updates for the weights of AquilaCode.
+- 2023/07/19 ：Released v0.8 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
-  - AquilaCode-7B-NV md5：91115e72a7fc7f780b410696eae6259c
-  - AquilaCode-7B-TS md5：5dae2486bc5a885279be87c13872cd5c
+  - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
+  - AquilaCode-py md5：3faa85fc03d8fda70a73064f48d02d85
 
 In the FlagEval large model evaluation ("Subjective + Objective"), AquilaChat-7B v0.8 has shown a slight overall improvement compared to version 0.7. It achieved an improvement of around 10% on the Chinese-MMLU dataset. For detailed evaluation results, please refer to the website http://flageval.baai.ac.cn.
 For detailed version change history, see [Change Log](https://github.com/FlagAI-Open/FlagAI/blob/master/examples/Aquila/changelog.md).
diff --git a/examples/Aquila/Aquila-code/README.md b/examples/Aquila/Aquila-code/README.md
index 70541bcf..1b23390f 100755
--- a/examples/Aquila/Aquila-code/README.md
+++ b/examples/Aquila/Aquila-code/README.md
@@ -21,19 +21,22 @@
 | :---------------- | :------- | :-- |:-- |   :-- | :-- | :-- | 
 | Aquila-7B         | 基础模型，70亿参数  |   **Aquila 基础模型**在技术上继承了 GPT-3、LLaMA 等的架构设计优点，替换了一批更高效的底层算子实现、重新设计实现了中英双语的 tokenizer，升级了 BMTrain 并行训练方法，实现了比 Magtron+DeepSpeed ZeRO-2 将近８倍的训练效率。   | [./examples/Aquila/Aquila-pretrain](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-pretrain)  | [下载Aquila-7B](http://model.baai.ac.cn/model-detail/100098) | 已发布 | Nvidia-A100 |
 | Aquila-33B          |基础模型，330亿参数  |    同上    | ——  | ——  | **敬请期待** | Nvidia-A100 | 
-| AquilaChat-7B          |SFT model，基于 Aquila-7B 进行微调和强化学习  |    **AquilaChat 对话模型**支持流畅的文本对话及多种语言类生成任务，通过定义可扩展的特殊指令规范，实现 AquilaChat对其它模型和工具的调用，且易于扩展。 <br><br>例如，调用智源开源的 **[AltDiffusion](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/AltDiffusion-m18) 多语言文图生成模型**，实现了流畅的文图生成能力。配合智源 **InstructFace 多步可控文生图模型**，轻松实现对人脸图像的多步可控编辑。  |   [./examples/Aquila/Aquila-chat](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-chat)  | [下载AquilaChat-7B](https://model.baai.ac.cn/model-detail/100101) | 已发布  | Nvidia-A100  | 
-| AquilaChat-33B           |SFT model，基于 Aquila-33B 进行微调和强化学习 |   同上    |   ——    |——  | **敬请期待** | Nvidia-A100 | 
-| AquilaCode-7B-NV          | 基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练，在英伟达芯片完成训练  |   AquilaCode-7B 以小数据集、小参数量，实现高性能，是目前支持中英双语的、性能最好的开源代码模型，经过了高质量过滤、使用有合规开源许可的训练代码数据进行训练。<br><br> AquilaCode-7B 分别在英伟达和国产芯片上完成了代码模型的训练。  | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  |[下载AquilaCode-7B-NV](https://model.baai.ac.cn/model-detail/100102)  | 已发布  | Nvidia-A100 | 
-| AquilaCode-7B-TS           |基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练，在天数智芯芯片上完成训练  |    同上    | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  | [下载AquilaCode-7B-TS](https://model.baai.ac.cn/model-detail/100099)  | 已发布  | Tianshu-BI-V100  | 
+| AquilaChat-7B          |SFT 模型，基于 Aquila-7B 进行微调和强化学习  |    **AquilaChat 对话模型**支持流畅的文本对话及多种语言类生成任务，通过定义可扩展的特殊指令规范，实现 AquilaChat对其它模型和工具的调用，且易于扩展。 <br><br>例如，调用智源开源的 **[AltDiffusion](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/AltDiffusion-m18) 多语言文图生成模型**，实现了流畅的文图生成能力。配合智源 **InstructFace 多步可控文生图模型**，轻松实现对人脸图像的多步可控编辑。  |   [./examples/Aquila/Aquila-chat](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-chat)  | [下载AquilaChat-7B](https://model.baai.ac.cn/model-detail/100101) | 已发布  | Nvidia-A100  | 
+| AquilaChat-33B           |SFT 模型，基于 Aquila-33B 进行微调和强化学习 |   同上    |   ——    |——  | **敬请期待** | Nvidia-A100 | 
+| AquilaCode-multi         | 基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练  |   AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型来定制自己的代码助手。  | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  |[下载AquilaCode-7B-multi](https://model.baai.ac.cn/model-detail/100102)  | 已发布  | Nvidia-A100 | 
+| AquilaCode-py           |基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练。  |    AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型定制自己的代码助手。    | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  | [下载AquilaCode-py](https://model.baai.ac.cn/model-detail/100099)  | 已发布  | Nvidia-A100  | 
 
 悟道·天鹰Aquila系列模型将持续开源更优版本，大家可以先删除原来目录下的code模型路径，再下载新权重，其他使用方式不变。
 
-- 2023/07/13 ：发布权重文件 v0.8，开源了 Aquila-7B、AquilaChat-7B 最新权重，AquilaCode 权重无更新。
+- 2023/07/19 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
-  - AquilaCode-7B-NV md5：91115e72a7fc7f780b410696eae6259c
-  - AquilaCode-7B-TS md5：5dae2486bc5a885279be87c13872cd5c
-    
+  - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
+  - AquilaCode-py md5：3faa85fc03d8fda70a73064f48d02d85
+
+AquilaCode-multi多语言模型支持多种编程语言高准确率生成，包括  Python/C++/Java/Javascript/Go 等。HumanEval（Python）评测结果 Pass@1、Pass@10、Pass@100 分别为 26/45.7/71.6，并在 HumanEval-X 多语言代码生成评测中显著领先于其他同参数等级的开源模型（2023.7.19）。
+
+AquilaCode-py单语言python版模型专注于 Python 代码生成，HumanEval 评测结果 Pass@1、Pass@10、Pass@100 分别为 28.8 / 50.6 / 76.9（2023.7.19）。
 
 <br>如有使用问题请先查看 [FAQ](https://github.com/FlagAI-Open/FlagAI/issues/371)，若不能解决，请直接提交 [issue](https://github.com/FlagAI-Open/FlagAI/issues) ~
 
@@ -178,5 +181,5 @@ python generate_code_bminf.py
 
 ## 证书
 
-AquilaCode-7B-NV和AquilaCode-7B-TS开源模型使用 [智源Aquila系列模型许可协议](../../../BAAI_Aquila_Model_License.pdf), 原始代码基于[Apache Licence 2.0](https://www.apache.org/licenses/LICENSE-2.0)。
+AquilaCode-7B-multi和AquilaCode-py开源模型使用 [智源Aquila系列模型许可协议](../../../BAAI_Aquila_Model_License.pdf), 原始代码基于[Apache Licence 2.0](https://www.apache.org/licenses/LICENSE-2.0)。
 
diff --git a/examples/Aquila/Aquila-code/README_en.md b/examples/Aquila/Aquila-code/README_en.md
index 9db2b97b..bac01a65 100755
--- a/examples/Aquila/Aquila-code/README_en.md
+++ b/examples/Aquila/Aquila-code/README_en.md
@@ -24,17 +24,17 @@ The additional details of the Aquila model will be presented in the official tec
 | Aquila-33B         | Base model, 33 billion parameters   | Same as above      | ——      | ——    | Coming soon                                               | Nvidia-A100   |
 | AquilaChat-7B      | SFT model, fine-tuned and RL based on Aquila-7B  | **AquilaChat Dialog Model** supports fluent text dialogue and multiple language generation tasks, and realizes the call of AquilaChat to other models and tools by defining an expandable special instruction specification, which is easy to extend. For example, calling the open source **[AltDiffusion](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/AltDiffusion-m18) multimodal language image generation model** of Flagship Intelligence achieved smooth image generation capability. Together with Flagship Intelligence's **InstructFace multi-step controllable text-picture model**, it is easy to achieve multi-step controllable editing of human face images. | [./examples/Aquila/Aquila-chat](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-chat) | [Download AquilaChat-7B](https://model.baai.ac.cn/model-detail/100101) | Released    | Nvidia-A100   |
 | AquilaChat-33B     | SFT model, fine-tuned and RL based on Aquila-33B  | Same as above| ——   | ——   |Coming soon                                               | Nvidia-A100   |
-| AquilaCode-7B-NV   | Base model, "text-code" generation model, further pre-trained based on Aquila-7B, trained on Nvidia  | AquilaCode-7B achieves high performance with small data sets and parameters, and is currently the best open source code model that supports both Chinese and English, trained using training code data with compliant open source licenses after high-quality filtering. AquilaCode-7B has been trained on both Nvidia and domestic chips for code models. | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-7B-NV](https://model.baai.ac.cn/model-detail/100102) | Released | Nvidia-A100  |
-| AquilaCode-7B-TS   | Base model, "text-code" generation model, further pre-trained based on Aquila-7B, trained on Horizon Robotics chips | Same as above | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-7B-TS](https://model.baai.ac.cn/model-detail/100099)  | Released        | Tianshu-BI-V100 |
+| AquilaCode-multi  | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B.  | AquilaCode utilizes high-quality, filtered, and compliant open-source code data for training, with a dataset size of approximately 10-40% compared to other open-source code generation models. By following the provided official guidelines, developers can harness the power of the AquilaCode model to customize their own code assistant. | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-multi](https://model.baai.ac.cn/model-detail/100102) | Released | Nvidia-A100  |
+| AquilaCode-py   | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B, trained on Horizon Robotics chips | Same as above | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-py](https://model.baai.ac.cn/model-detail/100099)  | Released        | Nvidia-A100 |
 
 
 We will continue to release improved versions of Aquila model as open source. You can start by deleting the checkpoint file in the original directory and then download the new weights. Other usage methods remain unchanged. For more details, please refer to the folloing change log:
 
-- 2023/07/13 ：Released v0.8 checkpoint files，The latest weights of Aquila-7B and AquilaChat-7B have been open sourced, but there are no updates for the weights of AquilaCode.
+- 2023/07/19 ：Released v0.8 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
-  - AquilaCode-7B-NV md5：91115e72a7fc7f780b410696eae6259c
-  - AquilaCode-7B-TS md5：5dae2486bc5a885279be87c13872cd5c
+  - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
+  - AquilaCode-py md5：3faa85fc03d8fda70a73064f48d02d85
 
 
  <br>If you have any question, please refer to the [FAQ](https://github.com/FlagAI-Open/FlagAI/issues/371) first. If you cannot solve them, please submit an [issue](https://github.com/FlagAI-Open/FlagAI/issues) directly.
@@ -77,7 +77,7 @@ Under default parameters, GPU memory consumption is approximately 4.3GB. You can
 
 ![bminf](../img/bminf.png)
 
-After running the inference program, the AquilaCode-7B-NV/the AquilaCode-7B-TS model will be automatically downloaded to ./`checkpoints_in`.
+After running the inference program, the AquilaCode-7B-multi/the AquilaCode-7B-TS model will be automatically downloaded to ./`checkpoints_in`.
 
 <details><summary>Example output:：</summary>
 
@@ -168,9 +168,9 @@ Complete parameter information can be found in https://github.com/FlagAI-Open/Fl
 | max_length     | int   | 200           | To avoid generating infinite length text, we need to limit the length of the generated text. The max_length parameter controls the maximum length of the generated text. Once this length is reached, the model stops generating. The maximum length of the Aquila series models is 2048 tokens. |
 
 - v0.5   
-md5 value of AquilaCode-7B-NV：91115e72a7fc7f780b410696eae6259c
+md5 value of AquilaCode-7B-multi：91115e72a7fc7f780b410696eae6259c
 md5 value of AquilaCode-7B-TS：5dae2486bc5a885279be87c13872cd5c
 
 ## License
 
-AquilaCode-7B-NV and AquilaCode-7B-TS open-source model is licensed under [ BAAI Aquila Model Licence Agreement](../../BAAI_Aquila_Model_License.pdf). The source code is under [Apache Licence 2.0](https://www.apache.org/licenses/LICENSE-2.0)
+AquilaCode-multiand AquilaCode-7B-TS open-source model is licensed under [ BAAI Aquila Model Licence Agreement](../../BAAI_Aquila_Model_License.pdf). The source code is under [Apache Licence 2.0](https://www.apache.org/licenses/LICENSE-2.0)
diff --git a/examples/Aquila/Aquila-code/generate_code.py b/examples/Aquila/Aquila-code/generate_code.py
index 06e8640e..d0b704d0 100755
--- a/examples/Aquila/Aquila-code/generate_code.py
+++ b/examples/Aquila/Aquila-code/generate_code.py
@@ -8,6 +8,7 @@
 from flagai.auto_model.auto_loader import AutoLoader
 import random
 import numpy as np
+
 from flagai.model.predictor.predictor import Predictor
 from flagai.data.tokenizer import Tokenizer
 
@@ -16,16 +17,16 @@
 
 print(f"building model...")
 loader = AutoLoader("lm",
-                    model_name="aquilacode-7b-nv",
+                    model_name="aquilacode-multi",
                     use_cache=True,
                     fp16=True,
+                    device=device,
                     model_dir=model_dir)
 
 model = loader.get_model()
 tokenizer = loader.get_tokenizer()
 model.half()
 model.eval()
-model.cuda()
 model.to(device)
 
 vocab = tokenizer.get_vocab()
@@ -44,5 +45,5 @@
     max_length = input_length + max_new_tokens
     with torch.no_grad():
         res = predictor.predict_generate_randomsample(
-            text, out_max_length=max_length, top_p=0.95, temperature=0.7)
+            text, out_max_length=max_length, top_p=0.95, temperature=0.1)
         print(res)
diff --git a/examples/Aquila/Aquila-pretrain/README.md b/examples/Aquila/Aquila-pretrain/README.md
index 3c2a80e2..f32d4c82 100755
--- a/examples/Aquila/Aquila-pretrain/README.md
+++ b/examples/Aquila/Aquila-pretrain/README.md
@@ -20,20 +20,20 @@
 | :---------------- | :------- | :-- |:-- |   :-- | :-- | :-- | 
 | Aquila-7B         | 基础模型，70亿参数  |   **Aquila 基础模型**在技术上继承了 GPT-3、LLaMA 等的架构设计优点，替换了一批更高效的底层算子实现、重新设计实现了中英双语的 tokenizer，升级了 BMTrain 并行训练方法，实现了比 Magtron+DeepSpeed ZeRO-2 将近８倍的训练效率。   | [./examples/Aquila/Aquila-pretrain](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-pretrain)  | [下载Aquila-7B](http://model.baai.ac.cn/model-detail/100098) | 已发布 | Nvidia-A100 |
 | Aquila-33B          |基础模型，330亿参数  |    同上    | ——  | ——  | **敬请期待** | Nvidia-A100 | 
-| AquilaChat-7B          |SFT model，基于 Aquila-7B 进行微调和强化学习  |    **AquilaChat 对话模型**支持流畅的文本对话及多种语言类生成任务，通过定义可扩展的特殊指令规范，实现 AquilaChat对其它模型和工具的调用，且易于扩展。 <br><br>例如，调用智源开源的 **[AltDiffusion](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/AltDiffusion-m18) 多语言文图生成模型**，实现了流畅的文图生成能力。配合智源 **InstructFace 多步可控文生图模型**，轻松实现对人脸图像的多步可控编辑。  |   [./examples/Aquila/Aquila-chat](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-chat)  | [下载AquilaChat-7B](https://model.baai.ac.cn/model-detail/100101) | 已发布  | Nvidia-A100  | 
-| AquilaChat-33B           |SFT model，基于 Aquila-33B 进行微调和强化学习 |   同上    |   ——    |——  | **敬请期待** | Nvidia-A100 | 
-| AquilaCode-7B-NV          | 基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练，在英伟达芯片完成训练  |   AquilaCode-7B 以小数据集、小参数量，实现高性能，是目前支持中英双语的、性能最好的开源代码模型，经过了高质量过滤、使用有合规开源许可的训练代码数据进行训练。<br><br> AquilaCode-7B 分别在英伟达和国产芯片上完成了代码模型的训练。  | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  |[下载AquilaCode-7B-NV](https://model.baai.ac.cn/model-detail/100102)  | 已发布  | Nvidia-A100 | 
-| AquilaCode-7B-TS           |基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练，在天数智芯芯片上完成训练  |    同上    | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  | [下载AquilaCode-7B-TS](https://model.baai.ac.cn/model-detail/100099)  | 已发布  | Tianshu-BI-V100  | 
+| AquilaChat-7B          |SFT 模型，基于 Aquila-7B 进行微调和强化学习  |    **AquilaChat 对话模型**支持流畅的文本对话及多种语言类生成任务，通过定义可扩展的特殊指令规范，实现 AquilaChat对其它模型和工具的调用，且易于扩展。 <br><br>例如，调用智源开源的 **[AltDiffusion](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/AltDiffusion-m18) 多语言文图生成模型**，实现了流畅的文图生成能力。配合智源 **InstructFace 多步可控文生图模型**，轻松实现对人脸图像的多步可控编辑。  |   [./examples/Aquila/Aquila-chat](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-chat)  | [下载AquilaChat-7B](https://model.baai.ac.cn/model-detail/100101) | 已发布  | Nvidia-A100  | 
+| AquilaChat-33B           |SFT 模型，基于 Aquila-33B 进行微调和强化学习 |   同上    |   ——    |——  | **敬请期待** | Nvidia-A100 | 
+| AquilaCode-multi         | 基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练  |   AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型来定制自己的代码助手。  | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  |[下载AquilaCode-7B-multi](https://model.baai.ac.cn/model-detail/100102)  | 已发布  | Nvidia-A100 | 
+| AquilaCode-py           |基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练。  |    AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型定制自己的代码助手。    | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  | [下载AquilaCode-py](https://model.baai.ac.cn/model-detail/100099)  | 已发布  | Nvidia-A100  | 
 
 悟道·天鹰Aquila系列模型将持续开源更优版本，大家可以先删除原来目录下的`checkpoints_in/aquila-7b`，再下载新权重，其他使用方式不变。
 
-- 2023/07/13 ：发布权重文件 v0.8，开源了 Aquila-7B、AquilaChat-7B 最新权重，AquilaCode 权重无更新。
+- 2023/07/19 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
-  - AquilaCode-7B-NV md5：91115e72a7fc7f780b410696eae6259c
-  - AquilaCode-7B-TS md5：5dae2486bc5a885279be87c13872cd5c
+  - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
+  - AquilaCode-py md5：3faa85fc03d8fda70a73064f48d02d85
 
-Aquila-7B v0.8 在 FlagEval 大模型评测中（ “客观”）相比0.7的版本在MMLU_Chinese、TruthfulQA、MMLU上分别提升10.07%，14.84%和7.94%。详细评测结果请通过 http://flageval.baai.ac.cn 网站查看。历史版本变更详情见：[变更日志](https://github.com/FlagAI-Open/FlagAI/blob/master/examples/Aquila/changelog.md) 。
+Aquila-7B 新版本模型 在 FlagEval 大模型评测中（ “客观”）相比0.7的版本在MMLU_Chinese、TruthfulQA、MMLU上分别提升10.07%，14.84%和7.94%。详细评测结果请通过 http://flageval.baai.ac.cn 网站查看。历史版本变更详情见：[变更日志](https://github.com/FlagAI-Open/FlagAI/blob/master/examples/Aquila/changelog.md) 。
 
 
 <br>如有使用问题请先查看 [FAQ](https://github.com/FlagAI-Open/FlagAI/issues/371)，若不能解决，请直接提交 [issue](https://github.com/FlagAI-Open/FlagAI/issues) ~
diff --git a/examples/Aquila/Aquila-pretrain/README_en.md b/examples/Aquila/Aquila-pretrain/README_en.md
index 2fb51904..bc2fb7f3 100755
--- a/examples/Aquila/Aquila-pretrain/README_en.md
+++ b/examples/Aquila/Aquila-pretrain/README_en.md
@@ -23,17 +23,17 @@ The additional details of the Aquila model will be presented in the official tec
 | Aquila-33B         | Base model, 33 billion parameters   | Same as above      | ——      | ——    | Coming soon                                               | Nvidia-A100   |
 | AquilaChat-7B      | SFT model, fine-tuned and RL based on Aquila-7B  | **AquilaChat Dialog Model** supports fluent text dialogue and multiple language generation tasks, and realizes the call of AquilaChat to other models and tools by defining an expandable special instruction specification, which is easy to extend. For example, calling the open source **[AltDiffusion](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/AltDiffusion-m18) multimodal language image generation model** of Flagship Intelligence achieved smooth image generation capability. Together with Flagship Intelligence's **InstructFace multi-step controllable text-picture model**, it is easy to achieve multi-step controllable editing of human face images. | [./examples/Aquila/Aquila-chat](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-chat) | [Download AquilaChat-7B](https://model.baai.ac.cn/model-detail/100101) | Released    | Nvidia-A100   |
 | AquilaChat-33B     | SFT model, fine-tuned and RL based on Aquila-33B  | Same as above| ——   | ——   |Coming soon                                               | Nvidia-A100   |
-| AquilaCode-7B-NV   | Base model, "text-code" generation model, further pre-trained based on Aquila-7B, trained on Nvidia  | AquilaCode-7B achieves high performance with small data sets and parameters, and is currently the best open source code model that supports both Chinese and English, trained using training code data with compliant open source licenses after high-quality filtering. AquilaCode-7B has been trained on both Nvidia and domestic chips for code models. | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-7B-NV](https://model.baai.ac.cn/model-detail/100102) | Released | Nvidia-A100  |
-| AquilaCode-7B-TS   | Base model, "text-code" generation model, further pre-trained based on Aquila-7B, trained on Horizon Robotics chips | Same as above | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-7B-TS](https://model.baai.ac.cn/model-detail/100099)  | Released        | Tianshu-BI-V100 |
+| AquilaCode-multi  | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B.  | AquilaCode utilizes high-quality, filtered, and compliant open-source code data for training, with a dataset size of approximately 10-40% compared to other open-source code generation models. By following the provided official guidelines, developers can harness the power of the AquilaCode model to customize their own code assistant. | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-multi](https://model.baai.ac.cn/model-detail/100102) | Released | Nvidia-A100  |
+| AquilaCode-py   | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B, trained on Horizon Robotics chips | Same as above | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-py](https://model.baai.ac.cn/model-detail/100099)  | Released        | Nvidia-A100 |
 
 
 We will continue to release improved versions of Aquila model as open source. You can start by deleting the `checkpoints_in/aquila-7b` in the original directory and then download the new weights. Other usage methods remain unchanged. For more details, please refer to the folloing change log:
 
-- 2023/07/13 ：Released v0.8 checkpoint files，The latest weights of Aquila-7B and AquilaChat-7B have been open sourced, but there are no updates for the weights of AquilaCode.
+- 2023/07/19 ：Released v0.8 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
-  - AquilaCode-7B-NV md5：91115e72a7fc7f780b410696eae6259c
-  - AquilaCode-7B-TS md5：5dae2486bc5a885279be87c13872cd5c
+  - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
+  - AquilaCode-py md5：3faa85fc03d8fda70a73064f48d02d85
 
 Aquila-7B v0.8 has shown improvements in the FlagEval large model evaluation ("Objective") compared to version 0.7. It achieved improvements of approximately 10.07% on MMLU_Chinese, 14.84% on TruthfulQA, and 7.94% on MMLU datasets. For detailed evaluation results, please refer to the website http://flageval.baai.ac.cn.
 For detailed version change history, see [Change Log](https://github.com/FlagAI-Open/FlagAI/blob/master/examples/Aquila/changelog.md).
diff --git a/examples/Aquila/README.md b/examples/Aquila/README.md
index d0d20eae..41ff9cf7 100755
--- a/examples/Aquila/README.md
+++ b/examples/Aquila/README.md
@@ -21,20 +21,20 @@
 | :---------------- | :------- | :-- |:-- |   :-- | :-- | :-- | 
 | Aquila-7B         | 基础模型，70亿参数  |   **Aquila 基础模型**在技术上继承了 GPT-3、LLaMA 等的架构设计优点，替换了一批更高效的底层算子实现、重新设计实现了中英双语的 tokenizer，升级了 BMTrain 并行训练方法，实现了比 Magtron+DeepSpeed ZeRO-2 将近８倍的训练效率。   | [./examples/Aquila/Aquila-pretrain](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-pretrain)  | [下载Aquila-7B](http://model.baai.ac.cn/model-detail/100098) | 已发布 | Nvidia-A100 |
 | Aquila-33B          |基础模型，330亿参数  |    同上    | ——  | ——  | **敬请期待** | Nvidia-A100 | 
-| AquilaChat-7B          |SFT model，基于 Aquila-7B 进行微调和强化学习  |    **AquilaChat 对话模型**支持流畅的文本对话及多种语言类生成任务，通过定义可扩展的特殊指令规范，实现 AquilaChat对其它模型和工具的调用，且易于扩展。 <br><br>例如，调用智源开源的 **[AltDiffusion](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/AltDiffusion-m18) 多语言文图生成模型**，实现了流畅的文图生成能力。配合智源 **InstructFace 多步可控文生图模型**，轻松实现对人脸图像的多步可控编辑。  |   [./examples/Aquila/Aquila-chat](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-chat)  | [下载AquilaChat-7B](https://model.baai.ac.cn/model-detail/100101) | 已发布  | Nvidia-A100  | 
-| AquilaChat-33B           |SFT model，基于 Aquila-33B 进行微调和强化学习 |   同上    |   ——    |——  | **敬请期待** | Nvidia-A100 | 
-| AquilaCode-7B-NV          | 基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练，在英伟达芯片完成训练  |   AquilaCode-7B 以小数据集、小参数量，实现高性能，是目前支持中英双语的、性能最好的开源代码模型，经过了高质量过滤、使用有合规开源许可的训练代码数据进行训练。<br><br> AquilaCode-7B 分别在英伟达和国产芯片上完成了代码模型的训练。  | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  |[下载AquilaCode-7B-NV](https://model.baai.ac.cn/model-detail/100102)  | 已发布  | Nvidia-A100 | 
-| AquilaCode-7B-TS           |基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练，在天数智芯芯片上完成训练  |    同上    | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  | [下载AquilaCode-7B-TS](https://model.baai.ac.cn/model-detail/100099)  | 已发布  | Tianshu-BI-V100  | 
+| AquilaChat-7B          |SFT 模型，基于 Aquila-7B 进行微调和强化学习  |    **AquilaChat 对话模型**支持流畅的文本对话及多种语言类生成任务，通过定义可扩展的特殊指令规范，实现 AquilaChat对其它模型和工具的调用，且易于扩展。 <br><br>例如，调用智源开源的 **[AltDiffusion](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/AltDiffusion-m18) 多语言文图生成模型**，实现了流畅的文图生成能力。配合智源 **InstructFace 多步可控文生图模型**，轻松实现对人脸图像的多步可控编辑。  |   [./examples/Aquila/Aquila-chat](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-chat)  | [下载AquilaChat-7B](https://model.baai.ac.cn/model-detail/100101) | 已发布  | Nvidia-A100  | 
+| AquilaChat-33B           |SFT 模型，基于 Aquila-33B 进行微调和强化学习 |   同上    |   ——    |——  | **敬请期待** | Nvidia-A100 | 
+| AquilaCode-multi         | 基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练  |   AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型来定制自己的代码助手。  | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  |[下载AquilaCode-7B-multi](https://model.baai.ac.cn/model-detail/100102)  | 已发布  | Nvidia-A100 | 
+| AquilaCode-py           |基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练。  |    AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型定制自己的代码助手。    | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  | [下载AquilaCode-py](https://model.baai.ac.cn/model-detail/100099)  | 已发布  | Nvidia-A100  |   
 
 悟道·天鹰Aquila系列模型将持续开源更优版本，大家可以先删除原来目录下的`checkpoints_in/aquila-7b`，再下载新权重，其他使用方式不变。
 
-- 2023/07/13 ：发布权重文件 v0.8，开源了 Aquila-7B、AquilaChat-7B 最新权重，AquilaCode 权重无更新。
+- 2023/07/19 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
-  - AquilaCode-7B-NV md5：91115e72a7fc7f780b410696eae6259c
-  - AquilaCode-7B-TS md5：5dae2486bc5a885279be87c13872cd5c
+  - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
+  - AquilaCode-py md5：3faa85fc03d8fda70a73064f48d02d85
 
-Aquila-7B v0.8 在 FlagEval 大模型评测中（ “客观”）相比0.7的版本在MMLU_Chinese、TruthfulQA、MMLU上分别提升10.07%，14.84%和7.94%。详细评测结果请通过 http://flageval.baai.ac.cn 网站查看。历史版本变更详情见：[变更日志](https://github.com/FlagAI-Open/FlagAI/blob/master/examples/Aquila/changelog.md) 。
+Aquila-7B 新版本模型在 FlagEval 大模型评测中（ “客观”）相比0.7的版本在MMLU_Chinese、TruthfulQA、MMLU上分别提升10.07%，14.84%和7.94%。详细评测结果请通过 http://flageval.baai.ac.cn 网站查看。历史版本变更详情见：[变更日志](https://github.com/FlagAI-Open/FlagAI/blob/master/examples/Aquila/changelog.md) 。
 
 
 <br>如有使用问题请先查看 [FAQ](https://github.com/FlagAI-Open/FlagAI/issues/371)，若不能解决，请直接提交 [issue](https://github.com/FlagAI-Open/FlagAI/issues) ~
diff --git a/examples/Aquila/README_en.md b/examples/Aquila/README_en.md
index 74518de3..8c9a0efa 100755
--- a/examples/Aquila/README_en.md
+++ b/examples/Aquila/README_en.md
@@ -25,17 +25,17 @@ The additional details of the Aquila model will be presented in the official tec
 | Aquila-33B         | Base model, 33 billion parameters   | Same as above      | ——      | ——    | Coming soon                                               | Nvidia-A100   |
 | AquilaChat-7B      | SFT model, fine-tuned and RL based on Aquila-7B  | **AquilaChat Dialog Model** supports fluent text dialogue and multiple language generation tasks, and realizes the call of AquilaChat to other models and tools by defining an expandable special instruction specification, which is easy to extend. For example, calling the open source **[AltDiffusion](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/AltDiffusion-m18) multimodal language image generation model** of Flagship Intelligence achieved smooth image generation capability. Together with Flagship Intelligence's **InstructFace multi-step controllable text-picture model**, it is easy to achieve multi-step controllable editing of human face images. | [./examples/Aquila/Aquila-chat](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-chat) | [Download AquilaChat-7B](https://model.baai.ac.cn/model-detail/100101) | Released    | Nvidia-A100   |
 | AquilaChat-33B     | SFT model, fine-tuned and RL based on Aquila-33B  | Same as above| ——   | ——   |Coming soon                                               | Nvidia-A100   |
-| AquilaCode-7B-NV   | Base model, "text-code" generation model, further pre-trained based on Aquila-7B, trained on Nvidia  | AquilaCode-7B achieves high performance with small data sets and parameters, and is currently the best open source code model that supports both Chinese and English, trained using training code data with compliant open source licenses after high-quality filtering. AquilaCode-7B has been trained on both Nvidia and domestic chips for code models. | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-7B-NV](https://model.baai.ac.cn/model-detail/100102) | Released | Nvidia-A100  |
-| AquilaCode-7B-TS   | Base model, "text-code" generation model, further pre-trained based on Aquila-7B, trained on Horizon Robotics chips | Same as above | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-7B-TS](https://model.baai.ac.cn/model-detail/100099)  | Released        | Tianshu-BI-V100 |
+| AquilaCode-multi  | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B.  | AquilaCode utilizes high-quality, filtered, and compliant open-source code data for training, with a dataset size of approximately 10-40% compared to other open-source code generation models. By following the provided official guidelines, developers can harness the power of the AquilaCode model to customize their own code assistant. | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-multi](https://model.baai.ac.cn/model-detail/100102) | Released | Nvidia-A100  |
+| AquilaCode-py   | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B, trained on Horizon Robotics chips | Same as above | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-py](https://model.baai.ac.cn/model-detail/100099)  | Released        | Nvidia-A100 |
 
 
 We will continue to release improved versions of Aquila model as open source. You can start by deleting the `checkpoints_in/aquila-7b` in the original directory and then download the new weights. Other usage methods remain unchanged. For more details, please refer to the folloing change log:
 
-- 2023/07/13 ：Released v0.8 checkpoint files，The latest weights of Aquila-7B and AquilaChat-7B have been open sourced, but there are no updates for the weights of AquilaCode.
+- 2023/07/19 ：Released v0.8 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
-  - AquilaCode-7B-NV md5：91115e72a7fc7f780b410696eae6259c
-  - AquilaCode-7B-TS md5：5dae2486bc5a885279be87c13872cd5c
+  - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
+  - AquilaCode-py md5：3faa85fc03d8fda70a73064f48d02d85
 
 Aquila-7B v0.8 has shown improvements in the FlagEval large model evaluation ("Objective") compared to version 0.7. It achieved improvements of approximately 10.07% on MMLU_Chinese, 14.84% on TruthfulQA, and 7.94% on MMLU datasets. For detailed evaluation results, please refer to the website http://flageval.baai.ac.cn.
 For detailed version change history, see [Change Log](https://github.com/FlagAI-Open/FlagAI/blob/master/examples/Aquila/changelog.md).
diff --git a/examples/Aquila/changelog.md b/examples/Aquila/changelog.md
index 1ad12bbd..6a5a9431 100755
--- a/examples/Aquila/changelog.md
+++ b/examples/Aquila/changelog.md
@@ -1,3 +1,8 @@
+- 2023/07/19 ：Released v0.8 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
+  - Aquila-7B md5: 18eac56434db0198494b22b321633785
+  - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
+  - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
+  - AquilaCode-py md5：3faa85fc03d8fda70a73064f48d02d85
 - 2023/07/13 ：Released v0.8 checkpoint files，The latest weights of Aquila-7B and AquilaChat-7B have been open sourced, but there are no updates for the weights of AquilaCode.
   - Aquila-7B  md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
diff --git a/examples/Aquila/changelog_zh.md b/examples/Aquila/changelog_zh.md
index acc8247c..7d4f4cda 100755
--- a/examples/Aquila/changelog_zh.md
+++ b/examples/Aquila/changelog_zh.md
@@ -1,3 +1,8 @@
+- 2023/07/19 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
+  - Aquila-7B md5: 18eac56434db0198494b22b321633785
+  - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
+  - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
+  - AquilaCode-py md5：3faa85fc03d8fda70a73064f48d02d85
 - 2023/07/13 ：发布权重文件 v0.8，开源了 Aquila-7B、AquilaChat-7B 最新权重，AquilaCode 权重无更新。
   - Aquila-7B  md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
diff --git a/examples/glm_title_generation/data_load.py b/examples/glm_title_generation/data_load.py
old mode 100644
new mode 100755
index 82415f3f..cdd3136b
--- a/examples/glm_title_generation/data_load.py
+++ b/examples/glm_title_generation/data_load.py
@@ -20,7 +20,7 @@
 from torch.utils.data.distributed import DistributedSampler
 from webdataset.filters import _shuffle
 from webdataset.tariterators import base_plus_ext, url_opener, tar_file_expander, valid_sample
-import pdb
+
 
 try:
     import horovod.torch as hvd
diff --git a/flagai/auto_model/auto_loader.py b/flagai/auto_model/auto_loader.py
index 739ce4ea..fe5b16ae 100755
--- a/flagai/auto_model/auto_loader.py
+++ b/flagai/auto_model/auto_loader.py
@@ -105,6 +105,8 @@ def __getattr__(self, name):
     "aquila-33b": ["flagai.model.aquila_model", "AQUILAModel", "aquila", "nlp"],
     "aquilacode-7b-nv": ["flagai.model.aquila_model", "AQUILAModel", "aquila", "nlp"],
     "aquilacode-7b-ts": ["flagai.model.aquila_model", "AQUILAModel", "aquila", "nlp"],
+    "aquilacode-multi": ["flagai.model.aquila_model", "AQUILAModel", "aquila", "nlp"],
+    "aquilacode-python": ["flagai.model.aquila_model", "AQUILAModel", "aquila", "nlp"],
     "vit-base-p16-224":
         ["flagai.model.vision.vit", "VisionTransformer", "vit", "vision"],
     "vit-base-p16-384":
diff --git a/flagai/model/aquila_model.py b/flagai/model/aquila_model.py
index 0bc44269..f8f9f868 100755
--- a/flagai/model/aquila_model.py
+++ b/flagai/model/aquila_model.py
@@ -186,7 +186,6 @@ def forward(self, input_ids: torch.Tensor, start_pos=0, labels=None, **kwargs):
                 layer.start_pos = start_pos
                 h = layer(h, freqs_cis, mask)
                 
-        # import pdb;pdb.set_trace()
         h = self.norm(h)
         if labels is not None:
             h = self.output(h)
diff --git a/flagai/model/predictor/aquila.py b/flagai/model/predictor/aquila.py
index fc67a745..d47bba66 100755
--- a/flagai/model/predictor/aquila.py
+++ b/flagai/model/predictor/aquila.py
@@ -29,7 +29,8 @@ def aquila_generate(
 
         total_len = min(2048, max_gen_len + max_prompt_size)
 
-        tokens = torch.full((bsz, total_len), 0).cuda().long()
+        # tokens = torch.full((bsz, total_len), 0).cuda().long()
+        tokens = torch.full((bsz, total_len), 0).to("cuda:5").long()
         for k, t in enumerate(prompt_tokens):
             tokens[k, : len(t)] = t.clone().detach().long()
         input_text_mask = tokens != 0

From eca533fe4132c1e484f1acb54316c8ea6ba0d6e3 Mon Sep 17 00:00:00 2001
From: ftgreat <ftgreat@163.com>
Date: Fri, 21 Jul 2023 14:44:55 +0800
Subject: [PATCH 2/5] updated date

Signed-off-by: ftgreat <ftgreat@163.com>
---
 examples/Aquila/Aquila-chat/README.md        |  10 +-
 examples/Aquila/Aquila-chat/README_en.md     |   2 +-
 examples/Aquila/Aquila-code/README.md        |   4 +-
 examples/Aquila/Aquila-code/README_en.md     |   9 +-
 examples/Aquila/Aquila-pretrain/README.md    |   4 +-
 examples/Aquila/Aquila-pretrain/README_en.md |   2 +-
 examples/Aquila/README.md                    |   4 +-
 examples/Aquila/README_en.md                 |   2 +-
 examples/Aquila/changelog.md                 |   2 +-
 examples/Aquila/changelog_zh.md              |   2 +-
 examples/Aquila/vscode_inference_service.md  | 121 +++++++++++++++++++
 11 files changed, 148 insertions(+), 14 deletions(-)
 create mode 100644 examples/Aquila/vscode_inference_service.md

diff --git a/examples/Aquila/Aquila-chat/README.md b/examples/Aquila/Aquila-chat/README.md
index 4c8226dc..1eb5b6e5 100755
--- a/examples/Aquila/Aquila-chat/README.md
+++ b/examples/Aquila/Aquila-chat/README.md
@@ -28,13 +28,13 @@
 
 悟道·天鹰Aquila系列模型将持续开源更优版本，大家可以先删除原来目录下的`checkpoints_in/aquilachat-7b`，再下载新权重，其他使用方式不变。
 
-- 2023/07/19 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
+- 2023/07/21 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
   - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
   - AquilaCode-py md5：3faa85fc03d8fda70a73064f48d02d85
     
-其中AquilaChat-7B新版本模型在 FlagEval 大模型评测中（ “主观+客观”）相比0.7的版本整体稍有提，其在Chinese-MMLU上提升10%左右，详细评测结果请通过 http://flageval.baai.ac.cn 网站查看。历史版本变更详情见：[变更日志](https://github.com/FlagAI-Open/FlagAI/blob/master/examples/Aquila/changelog.md) 。
+其中AquilaChat-7B新版本模型在 FlagEval 大模型评测中（ “主观+客观”）相比0.7的版本整体稍有提，其在Chinese-MMLU上提升10%左右，详细评测结果请通过 http://flageval.baai.ac.cn 网站查看。历史版本变更详情见：[变更日志](https://github.com/FlagAI-Open/FlagAI/blob/master/examples/Aquila/changelog_zh.md) 。
 
 
 <br>如有使用问题请先查看 [FAQ](https://github.com/FlagAI-Open/FlagAI/issues/371)，若不能解决，请直接提交 [issue](https://github.com/FlagAI-Open/FlagAI/issues) ~
@@ -42,10 +42,14 @@
 
 ## 快速开始使用 AquilaChat-7B 对话模型
 
+### VScode上部署推理线上服务
+
+<br>具体部署流程见[FAQ](../vscode_inference_service.md)
+
 ### 基础模型的环境准备
 
 1. 在本地克隆FlagAI github仓库
-   
+
     ```
     git clone https://github.com/FlagAI-Open/FlagAI.git
     ```
diff --git a/examples/Aquila/Aquila-chat/README_en.md b/examples/Aquila/Aquila-chat/README_en.md
index 8554b2e9..aea58576 100755
--- a/examples/Aquila/Aquila-chat/README_en.md
+++ b/examples/Aquila/Aquila-chat/README_en.md
@@ -30,7 +30,7 @@ The additional details of the Aquila model will be presented in the official tec
 
 We will continue to release improved versions of Aquila model as open source. You can start by deleting the `checkpoints_in/aquilachat-7b` in the original directory and then download the new weights. Other usage methods remain unchanged. For more details, please refer to the folloing change log:
 
-- 2023/07/19 ：Released v0.8 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
+- 2023/07/21 ：Released v0.8 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
   - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
diff --git a/examples/Aquila/Aquila-code/README.md b/examples/Aquila/Aquila-code/README.md
index 1b23390f..cd3bab37 100755
--- a/examples/Aquila/Aquila-code/README.md
+++ b/examples/Aquila/Aquila-code/README.md
@@ -28,7 +28,7 @@
 
 悟道·天鹰Aquila系列模型将持续开源更优版本，大家可以先删除原来目录下的code模型路径，再下载新权重，其他使用方式不变。
 
-- 2023/07/19 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
+- 2023/07/21 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
   - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
@@ -38,6 +38,8 @@ AquilaCode-multi多语言模型支持多种编程语言高准确率生成，包
 
 AquilaCode-py单语言python版模型专注于 Python 代码生成，HumanEval 评测结果 Pass@1、Pass@10、Pass@100 分别为 28.8 / 50.6 / 76.9（2023.7.19）。
 
+历史版本变更详情见：[变更日志](https://github.com/FlagAI-Open/FlagAI/blob/master/examples/Aquila/changelog_zh.md) 。
+
 <br>如有使用问题请先查看 [FAQ](https://github.com/FlagAI-Open/FlagAI/issues/371)，若不能解决，请直接提交 [issue](https://github.com/FlagAI-Open/FlagAI/issues) ~
 
 
diff --git a/examples/Aquila/Aquila-code/README_en.md b/examples/Aquila/Aquila-code/README_en.md
index bac01a65..0c2782f7 100755
--- a/examples/Aquila/Aquila-code/README_en.md
+++ b/examples/Aquila/Aquila-code/README_en.md
@@ -30,13 +30,20 @@ The additional details of the Aquila model will be presented in the official tec
 
 We will continue to release improved versions of Aquila model as open source. You can start by deleting the checkpoint file in the original directory and then download the new weights. Other usage methods remain unchanged. For more details, please refer to the folloing change log:
 
-- 2023/07/19 ：Released v0.8 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
+- 2023/07/21 ：Released v0.8 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
   - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
   - AquilaCode-py md5：3faa85fc03d8fda70a73064f48d02d85
 
 
+AquilaCode-multi is a multi-language model that supports high-accuracy code generation for various programming languages, including Python/C++/Java/Javascript/Go, etc. It has achieved impressive results in HumanEval (Python) evaluation, with Pass@1, Pass@10, and Pass@100 scores of 26/45.7/71.6, respectively. In the HumanEval-X multi-language code generation evaluation, it significantly outperforms other open-source models with similar parameters (as of July 19, 2023).
+
+AquilaCode-py, on the other hand, is a single-language Python version of the model that focuses on Python code generation. It has also demonstrated excellent performance in HumanEval evaluation, with Pass@1, Pass@10, and Pass@100 scores of 28.8/50.6/76.9 (as of July 19, 2023).
+
+For more details about the version history and changes, you can refer to the [changelog](https://github.com/FlagAI-Open/FlagAI/blob/master/examples/Aquila/changelog.md).
+
+
  <br>If you have any question, please refer to the [FAQ](https://github.com/FlagAI-Open/FlagAI/issues/371) first. If you cannot solve them, please submit an [issue](https://github.com/FlagAI-Open/FlagAI/issues) directly.
 
 
diff --git a/examples/Aquila/Aquila-pretrain/README.md b/examples/Aquila/Aquila-pretrain/README.md
index f32d4c82..63fb9057 100755
--- a/examples/Aquila/Aquila-pretrain/README.md
+++ b/examples/Aquila/Aquila-pretrain/README.md
@@ -27,13 +27,13 @@
 
 悟道·天鹰Aquila系列模型将持续开源更优版本，大家可以先删除原来目录下的`checkpoints_in/aquila-7b`，再下载新权重，其他使用方式不变。
 
-- 2023/07/19 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
+- 2023/07/21 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
   - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
   - AquilaCode-py md5：3faa85fc03d8fda70a73064f48d02d85
 
-Aquila-7B 新版本模型 在 FlagEval 大模型评测中（ “客观”）相比0.7的版本在MMLU_Chinese、TruthfulQA、MMLU上分别提升10.07%，14.84%和7.94%。详细评测结果请通过 http://flageval.baai.ac.cn 网站查看。历史版本变更详情见：[变更日志](https://github.com/FlagAI-Open/FlagAI/blob/master/examples/Aquila/changelog.md) 。
+Aquila-7B 新版本模型 在 FlagEval 大模型评测中（ “客观”）相比0.7的版本在MMLU_Chinese、TruthfulQA、MMLU上分别提升10.07%，14.84%和7.94%。详细评测结果请通过 http://flageval.baai.ac.cn 网站查看。历史版本变更详情见：[变更日志](https://github.com/FlagAI-Open/FlagAI/blob/master/examples/Aquila/changelog_zh.md) 。
 
 
 <br>如有使用问题请先查看 [FAQ](https://github.com/FlagAI-Open/FlagAI/issues/371)，若不能解决，请直接提交 [issue](https://github.com/FlagAI-Open/FlagAI/issues) ~
diff --git a/examples/Aquila/Aquila-pretrain/README_en.md b/examples/Aquila/Aquila-pretrain/README_en.md
index bc2fb7f3..93d1f75f 100755
--- a/examples/Aquila/Aquila-pretrain/README_en.md
+++ b/examples/Aquila/Aquila-pretrain/README_en.md
@@ -29,7 +29,7 @@ The additional details of the Aquila model will be presented in the official tec
 
 We will continue to release improved versions of Aquila model as open source. You can start by deleting the `checkpoints_in/aquila-7b` in the original directory and then download the new weights. Other usage methods remain unchanged. For more details, please refer to the folloing change log:
 
-- 2023/07/19 ：Released v0.8 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
+- 2023/07/21 ：Released v0.8 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
   - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
diff --git a/examples/Aquila/README.md b/examples/Aquila/README.md
index 41ff9cf7..b738ef07 100755
--- a/examples/Aquila/README.md
+++ b/examples/Aquila/README.md
@@ -28,13 +28,13 @@
 
 悟道·天鹰Aquila系列模型将持续开源更优版本，大家可以先删除原来目录下的`checkpoints_in/aquila-7b`，再下载新权重，其他使用方式不变。
 
-- 2023/07/19 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
+- 2023/07/21 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
   - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
   - AquilaCode-py md5：3faa85fc03d8fda70a73064f48d02d85
 
-Aquila-7B 新版本模型在 FlagEval 大模型评测中（ “客观”）相比0.7的版本在MMLU_Chinese、TruthfulQA、MMLU上分别提升10.07%，14.84%和7.94%。详细评测结果请通过 http://flageval.baai.ac.cn 网站查看。历史版本变更详情见：[变更日志](https://github.com/FlagAI-Open/FlagAI/blob/master/examples/Aquila/changelog.md) 。
+Aquila-7B 新版本模型在 FlagEval 大模型评测中（ “客观”）相比0.7的版本在MMLU_Chinese、TruthfulQA、MMLU上分别提升10.07%，14.84%和7.94%。详细评测结果请通过 http://flageval.baai.ac.cn 网站查看。历史版本变更详情见：[变更日志](https://github.com/FlagAI-Open/FlagAI/blob/master/examples/Aquila/changelog_zh.md) 。
 
 
 <br>如有使用问题请先查看 [FAQ](https://github.com/FlagAI-Open/FlagAI/issues/371)，若不能解决，请直接提交 [issue](https://github.com/FlagAI-Open/FlagAI/issues) ~
diff --git a/examples/Aquila/README_en.md b/examples/Aquila/README_en.md
index 8c9a0efa..f56d8b3a 100755
--- a/examples/Aquila/README_en.md
+++ b/examples/Aquila/README_en.md
@@ -31,7 +31,7 @@ The additional details of the Aquila model will be presented in the official tec
 
 We will continue to release improved versions of Aquila model as open source. You can start by deleting the `checkpoints_in/aquila-7b` in the original directory and then download the new weights. Other usage methods remain unchanged. For more details, please refer to the folloing change log:
 
-- 2023/07/19 ：Released v0.8 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
+- 2023/07/21 ：Released v0.8 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
   - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
diff --git a/examples/Aquila/changelog.md b/examples/Aquila/changelog.md
index 6a5a9431..5e117592 100755
--- a/examples/Aquila/changelog.md
+++ b/examples/Aquila/changelog.md
@@ -1,4 +1,4 @@
-- 2023/07/19 ：Released v0.8 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
+- 2023/07/21 ：Released v0.9 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
   - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
diff --git a/examples/Aquila/changelog_zh.md b/examples/Aquila/changelog_zh.md
index 7d4f4cda..7d6ebff5 100755
--- a/examples/Aquila/changelog_zh.md
+++ b/examples/Aquila/changelog_zh.md
@@ -1,4 +1,4 @@
-- 2023/07/19 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
+- 2023/07/21 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
   - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
diff --git a/examples/Aquila/vscode_inference_service.md b/examples/Aquila/vscode_inference_service.md
new file mode 100644
index 00000000..ab918cc1
--- /dev/null
+++ b/examples/Aquila/vscode_inference_service.md
@@ -0,0 +1,121 @@
+## 该项目是基于[CodeGeeX](https://github.com/CodeGeeX/codegeex-vscode-extension) master分支-tag：1.1.2完成的开发，具体修改如下：
+
+* packagge.json中contributes.configuration.properties添加插件配置信息
+```
+"CodeGeeX.DecodingStrategies.max_seq_len": {
+    "type": "number",
+    "default": 50,
+    "maximum": 300,
+    "minimum": 20,
+    "description": "custom the max_seq_len,rang [20, 300]."
+},
+"CodeGeeX.DecodingStrategies.url": {
+    "type": "string",
+    "default": "请求模型接口地址",
+    "description": "Replace plug-in interface calls"
+},
+```
+
+* src/params/configures增加获取配置信息
+```
+const defaultConfig = {
+    temp: 0.8,
+    topp: 0.95,
+    topk: 0,
+    url: '默认配置模型输出接口',
+    max_seq_len: 50,  // 新增配置
+};
+const modelConfig = configuration.get("DecodingStrategies", defaultConfig);
+export const temp = modelConfig.temp;
+export const topk = modelConfig.topk;
+export const topp = modelConfig.topp;
+export const API_URL = modelConfig.url;  // 获取配置
+export const max_seq_len = modelConfig.max_seq_len;  // 获取配置
+```
+
+* src/provider/inlineCompletionProvider中去掉权限密钥校验相关代码
+```
+// rs = await getCodeCompletions(
+//     textBeforeCursor,
+//     num,
+//     lang,
+//     apiKey,
+//     apiSecret,
+//     "inlinecompletion"
+// );
+let timestart = new Date().getTime();
+let timeend = new Date().getTime();
+const completions = Array<string>();
+let commandid = "";
+rs = { completions, commandid };
+```
+
+* src/utils/getCodeCompletions 不同模式下接口地址配置，这里我们修改为调用同一个地址，并且地址在插件中可以配置
+```
+import {
+    temp, topp, topk, API_URL,
+    max_seq_len, } from "../param/configures";
+去掉相关选择代码
+ // let API_URL = "";
+// if (mode === "prompt") {
+//     API_URL = `${apiHref}/multilingual_code_generate_block`;
+// } else if (mode === "interactive") {
+//     API_URL = `${apiHref}/multilingual_code_generate_adapt`;
+// } else {
+//     if (generationPreference === "line by line") {
+//         API_URL = `${apiHref}/multilingual_code_generate`;
+//     } else {
+//         API_URL = `${apiHref}/multilingual_code_generate_adapt`;
+//     }
+// }
+payload参数也进行修改
+let payload = {
+    ability: "seo_article_creation",
+    context: prompt,
+    temperature: temp,
+    top_k: topk,
+    top_p: topp,
+    max_seq_len: max_seq_len,
+    len_penalty: 1.0,
+    repetition_penalty: 1.0,
+    presence_penalty: 1.0,
+    frequency_penalty: 1.0,
+    end_tokens: [],
+};
+接口返回数据格式根据需要自定义
+// 原代码位置
+ if (res?.data.status === 0) {
+    let codeArray = res?.data.result.output.code;
+    const completions = Array<string>();
+    for (let i = 0; i < codeArray.length; i++) {
+        const completion = codeArray[i];
+        let tmpstr = completion;
+        if (tmpstr.trim() === "") continue;
+        if (completions.includes(completion)) continue;
+        completions.push(completion);
+    }
+    let timeEnd = new Date().getTime();
+    console.log(timeEnd - time1, timeEnd - time2);
+    resolve({ completions, commandid });
+} else {
+    try {
+        await getEndData(commandid, res.data.message, "No");
+    } catch (err) {
+        console.log(err);
+    }
+    reject(res.data.message);
+}
+替换
+let codeArray = res?.data.generated;
+const completions = Array<string>();
+completions.push(codeArray);
+let timeEnd = new Date().getTime();
+resolve({ completions, commandid });
+```
+
+## 项目运行以及打包
+* 首先电脑安装前端项目启动所需环境：node。安装完成之后通过node -v和npm -v，安装成功应该可以看到对应的版本号
+* 终端目录指向项目根目录，执行`npm install`安装相关以来,安装完成根目录会多出一个package-lock.json文件和node_modules文件夹
+* vscode中运行和调试左上角执行`Run Extentensions`,修改src目录下的代码会进行实时编译到out文件夹中进行本地调试
+* 安装全局模块vsce [`npm install --global @vscode/vsce`](https://www.npmjs.com/package/@vscode/vsce)
+* 修改完代码执行`vsce package`打包插件，根目录会生成.vsix文件，然后在vscode插件中选择从VSIX安装，选择刚刚打包好的文件进行安装试用
\ No newline at end of file

From 260a95cb605593caf2f03785305c6b8098ded8ef Mon Sep 17 00:00:00 2001
From: ftgreat <ftgreat@163.com>
Date: Fri, 21 Jul 2023 16:21:53 +0800
Subject: [PATCH 3/5] updated requirements

Signed-off-by: ftgreat <ftgreat@163.com>
---
 setup.py | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/setup.py b/setup.py
index af042349..ab4a8553 100755
--- a/setup.py
+++ b/setup.py
@@ -17,10 +17,10 @@
     include_package_data=True,
     python_requires=">=3.8",
     install_requires=[
-        'nltk==3.6.7',
+        'nltk>=3.6.7',
         'sentencepiece>=0.1.96',
         'boto3==1.17.32',
-        'pandas==1.3.5',
+        'pandas>=1.3.5',
         'jieba==0.42.1',
         'scikit-learn==1.0.2',
         'tensorboard>=2.9.0',
@@ -30,12 +30,12 @@
         'protobuf==3.19.6',
         'ftfy',
         'Pillow>=9.3.0',
-        'einops==0.3.0',
+        'einops>=0.3.0',
         'diffusers==0.7.2',
         'pytorch-lightning==1.6.5',
         'taming-transformers-rom1504==0.0.6',
         'rouge-score',
-        'sacrebleu==2.3.1',
+        'sacrebleu>=2.3.1',
         'jsonlines',
         'accelerate',
         'PyYAML==5.4.1',

From 1dc4998c9b9eb5eaab89d74bcaee8848f235e8c9 Mon Sep 17 00:00:00 2001
From: ftgreat <ftgreat@163.com>
Date: Mon, 24 Jul 2023 08:55:39 +0800
Subject: [PATCH 4/5] updated to the lastest time

Signed-off-by: ftgreat <ftgreat@163.com>
---
 examples/Aquila/Aquila-chat/README.md        | 2 +-
 examples/Aquila/Aquila-chat/README_en.md     | 2 +-
 examples/Aquila/Aquila-code/README.md        | 2 +-
 examples/Aquila/Aquila-code/README_en.md     | 2 +-
 examples/Aquila/Aquila-pretrain/README.md    | 2 +-
 examples/Aquila/Aquila-pretrain/README_en.md | 2 +-
 examples/Aquila/README.md                    | 2 +-
 examples/Aquila/README_en.md                 | 2 +-
 examples/Aquila/changelog.md                 | 2 +-
 examples/Aquila/changelog_zh.md              | 2 +-
 10 files changed, 10 insertions(+), 10 deletions(-)

diff --git a/examples/Aquila/Aquila-chat/README.md b/examples/Aquila/Aquila-chat/README.md
index 1eb5b6e5..09541806 100755
--- a/examples/Aquila/Aquila-chat/README.md
+++ b/examples/Aquila/Aquila-chat/README.md
@@ -28,7 +28,7 @@
 
 悟道·天鹰Aquila系列模型将持续开源更优版本，大家可以先删除原来目录下的`checkpoints_in/aquilachat-7b`，再下载新权重，其他使用方式不变。
 
-- 2023/07/21 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
+- 2023/07/24 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
   - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
diff --git a/examples/Aquila/Aquila-chat/README_en.md b/examples/Aquila/Aquila-chat/README_en.md
index aea58576..b43a550d 100755
--- a/examples/Aquila/Aquila-chat/README_en.md
+++ b/examples/Aquila/Aquila-chat/README_en.md
@@ -30,7 +30,7 @@ The additional details of the Aquila model will be presented in the official tec
 
 We will continue to release improved versions of Aquila model as open source. You can start by deleting the `checkpoints_in/aquilachat-7b` in the original directory and then download the new weights. Other usage methods remain unchanged. For more details, please refer to the folloing change log:
 
-- 2023/07/21 ：Released v0.8 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
+- 2023/07/24 ：Released v0.8 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
   - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
diff --git a/examples/Aquila/Aquila-code/README.md b/examples/Aquila/Aquila-code/README.md
index cd3bab37..a2642210 100755
--- a/examples/Aquila/Aquila-code/README.md
+++ b/examples/Aquila/Aquila-code/README.md
@@ -28,7 +28,7 @@
 
 悟道·天鹰Aquila系列模型将持续开源更优版本，大家可以先删除原来目录下的code模型路径，再下载新权重，其他使用方式不变。
 
-- 2023/07/21 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
+- 2023/07/24 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
   - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
diff --git a/examples/Aquila/Aquila-code/README_en.md b/examples/Aquila/Aquila-code/README_en.md
index 0c2782f7..2b6d16df 100755
--- a/examples/Aquila/Aquila-code/README_en.md
+++ b/examples/Aquila/Aquila-code/README_en.md
@@ -30,7 +30,7 @@ The additional details of the Aquila model will be presented in the official tec
 
 We will continue to release improved versions of Aquila model as open source. You can start by deleting the checkpoint file in the original directory and then download the new weights. Other usage methods remain unchanged. For more details, please refer to the folloing change log:
 
-- 2023/07/21 ：Released v0.8 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
+- 2023/07/24 ：Released v0.8 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
   - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
diff --git a/examples/Aquila/Aquila-pretrain/README.md b/examples/Aquila/Aquila-pretrain/README.md
index 63fb9057..eed216e1 100755
--- a/examples/Aquila/Aquila-pretrain/README.md
+++ b/examples/Aquila/Aquila-pretrain/README.md
@@ -27,7 +27,7 @@
 
 悟道·天鹰Aquila系列模型将持续开源更优版本，大家可以先删除原来目录下的`checkpoints_in/aquila-7b`，再下载新权重，其他使用方式不变。
 
-- 2023/07/21 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
+- 2023/07/24 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
   - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
diff --git a/examples/Aquila/Aquila-pretrain/README_en.md b/examples/Aquila/Aquila-pretrain/README_en.md
index 93d1f75f..1ba5c32b 100755
--- a/examples/Aquila/Aquila-pretrain/README_en.md
+++ b/examples/Aquila/Aquila-pretrain/README_en.md
@@ -29,7 +29,7 @@ The additional details of the Aquila model will be presented in the official tec
 
 We will continue to release improved versions of Aquila model as open source. You can start by deleting the `checkpoints_in/aquila-7b` in the original directory and then download the new weights. Other usage methods remain unchanged. For more details, please refer to the folloing change log:
 
-- 2023/07/21 ：Released v0.8 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
+- 2023/07/24 ：Released v0.8 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
   - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
diff --git a/examples/Aquila/README.md b/examples/Aquila/README.md
index b738ef07..7c468ed7 100755
--- a/examples/Aquila/README.md
+++ b/examples/Aquila/README.md
@@ -28,7 +28,7 @@
 
 悟道·天鹰Aquila系列模型将持续开源更优版本，大家可以先删除原来目录下的`checkpoints_in/aquila-7b`，再下载新权重，其他使用方式不变。
 
-- 2023/07/21 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
+- 2023/07/24 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
   - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
diff --git a/examples/Aquila/README_en.md b/examples/Aquila/README_en.md
index f56d8b3a..a2da3faf 100755
--- a/examples/Aquila/README_en.md
+++ b/examples/Aquila/README_en.md
@@ -31,7 +31,7 @@ The additional details of the Aquila model will be presented in the official tec
 
 We will continue to release improved versions of Aquila model as open source. You can start by deleting the `checkpoints_in/aquila-7b` in the original directory and then download the new weights. Other usage methods remain unchanged. For more details, please refer to the folloing change log:
 
-- 2023/07/21 ：Released v0.8 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
+- 2023/07/24 ：Released v0.8 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
   - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
diff --git a/examples/Aquila/changelog.md b/examples/Aquila/changelog.md
index 5e117592..7affe97c 100755
--- a/examples/Aquila/changelog.md
+++ b/examples/Aquila/changelog.md
@@ -1,4 +1,4 @@
-- 2023/07/21 ：Released v0.9 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
+- 2023/07/24 ：Released v0.9 checkpoint files，AquilaCode-multi and AquilaCode-python have been released while AquilaCode-7B-NV and AquilaCode-7B-TS are temporarily not maintained. There are no updates for the weights of Aquila-7B and AquilaChat-7B.
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
   - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286
diff --git a/examples/Aquila/changelog_zh.md b/examples/Aquila/changelog_zh.md
index 7d6ebff5..75d06ae4 100755
--- a/examples/Aquila/changelog_zh.md
+++ b/examples/Aquila/changelog_zh.md
@@ -1,4 +1,4 @@
-- 2023/07/21 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
+- 2023/07/24 ：发布权重文件 v0.9，开源了 AquilaCode-multi、AquilaCode-py。 AquilaChat-7B和Aquila-7B权重无更新, AquilaCode-7B-NV和AquilaCode-7B-TS权重暂时不会有更新计划。
   - Aquila-7B md5: 18eac56434db0198494b22b321633785
   - AquilaChat-7B md5: 465683009c8b536ef4cca85febb0227c
   - AquilaCode-multi md5：07cfce9440a0fa1ac2768b39d2cf4286

From 8cec8645e76f951a33efcb519564929ff427e5b4 Mon Sep 17 00:00:00 2001
From: ftgreat <ftgreat@163.com>
Date: Mon, 24 Jul 2023 08:58:38 +0800
Subject: [PATCH 5/5] updated modelhub address

Signed-off-by: ftgreat <ftgreat@163.com>
---
 examples/Aquila/Aquila-chat/README.md        | 4 ++--
 examples/Aquila/Aquila-chat/README_en.md     | 4 ++--
 examples/Aquila/Aquila-code/README.md        | 4 ++--
 examples/Aquila/Aquila-code/README_en.md     | 4 ++--
 examples/Aquila/Aquila-pretrain/README.md    | 4 ++--
 examples/Aquila/Aquila-pretrain/README_en.md | 4 ++--
 examples/Aquila/README.md                    | 4 ++--
 examples/Aquila/README_en.md                 | 4 ++--
 8 files changed, 16 insertions(+), 16 deletions(-)

diff --git a/examples/Aquila/Aquila-chat/README.md b/examples/Aquila/Aquila-chat/README.md
index 09541806..399fa885 100755
--- a/examples/Aquila/Aquila-chat/README.md
+++ b/examples/Aquila/Aquila-chat/README.md
@@ -23,8 +23,8 @@
 | Aquila-33B          |基础模型，330亿参数  |    同上    | ——  | ——  | **敬请期待** | Nvidia-A100 | 
 | AquilaChat-7B          |SFT 模型，基于 Aquila-7B 进行微调和强化学习  |    **AquilaChat 对话模型**支持流畅的文本对话及多种语言类生成任务，通过定义可扩展的特殊指令规范，实现 AquilaChat对其它模型和工具的调用，且易于扩展。 <br><br>例如，调用智源开源的 **[AltDiffusion](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/AltDiffusion-m18) 多语言文图生成模型**，实现了流畅的文图生成能力。配合智源 **InstructFace 多步可控文生图模型**，轻松实现对人脸图像的多步可控编辑。  |   [./examples/Aquila/Aquila-chat](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-chat)  | [下载AquilaChat-7B](https://model.baai.ac.cn/model-detail/100101) | 已发布  | Nvidia-A100  | 
 | AquilaChat-33B           |SFT 模型，基于 Aquila-33B 进行微调和强化学习 |   同上    |   ——    |——  | **敬请期待** | Nvidia-A100 | 
-| AquilaCode-multi         | 基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练  |   AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型来定制自己的代码助手。  | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  |[下载AquilaCode-7B-multi](https://model.baai.ac.cn/model-detail/100102)  | 已发布  | Nvidia-A100 | 
-| AquilaCode-py           |基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练。  |    AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型定制自己的代码助手。    | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  | [下载AquilaCode-py](https://model.baai.ac.cn/model-detail/100099)  | 已发布  | Nvidia-A100  |   
+| AquilaCode-multi         | 基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练  |   AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型来定制自己的代码助手。  | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  |[下载AquilaCode-7B-multi](https://model.baai.ac.cn/model-detail/100104)  | 已发布  | Nvidia-A100 | 
+| AquilaCode-py           |基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练。  |    AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型定制自己的代码助手。    | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  | [下载AquilaCode-py](https://model.baai.ac.cn/model-detail/100103)  | 已发布  | Nvidia-A100  |   
 
 悟道·天鹰Aquila系列模型将持续开源更优版本，大家可以先删除原来目录下的`checkpoints_in/aquilachat-7b`，再下载新权重，其他使用方式不变。
 
diff --git a/examples/Aquila/Aquila-chat/README_en.md b/examples/Aquila/Aquila-chat/README_en.md
index b43a550d..f7a66859 100755
--- a/examples/Aquila/Aquila-chat/README_en.md
+++ b/examples/Aquila/Aquila-chat/README_en.md
@@ -24,8 +24,8 @@ The additional details of the Aquila model will be presented in the official tec
 | Aquila-33B         | Base model, 33 billion parameters   | Same as above      | ——      | ——    | Coming soon                                               | Nvidia-A100   |
 | AquilaChat-7B      | SFT model, fine-tuned and RL based on Aquila-7B  | **AquilaChat Dialog Model** supports fluent text dialogue and multiple language generation tasks, and realizes the call of AquilaChat to other models and tools by defining an expandable special instruction specification, which is easy to extend. For example, calling the open source **[AltDiffusion](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/AltDiffusion-m18) multimodal language image generation model** of Flagship Intelligence achieved smooth image generation capability. Together with Flagship Intelligence's **InstructFace multi-step controllable text-picture model**, it is easy to achieve multi-step controllable editing of human face images. | [./examples/Aquila/Aquila-chat](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-chat) | [Download AquilaChat-7B](https://model.baai.ac.cn/model-detail/100101) | Released    | Nvidia-A100   |
 | AquilaChat-33B     | SFT model, fine-tuned and RL based on Aquila-33B  | Same as above| ——   | ——   |Coming soon                                               | Nvidia-A100   |
-| AquilaCode-multi  | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B.  | AquilaCode utilizes high-quality, filtered, and compliant open-source code data for training, with a dataset size of approximately 10-40% compared to other open-source code generation models. By following the provided official guidelines, developers can harness the power of the AquilaCode model to customize their own code assistant. | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-multi](https://model.baai.ac.cn/model-detail/100102) | Released | Nvidia-A100  |
-| AquilaCode-py   | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B, trained on Horizon Robotics chips | Same as above | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-py](https://model.baai.ac.cn/model-detail/100099)  | Released        | Nvidia-A100 |
+| AquilaCode-multi  | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B.  | AquilaCode utilizes high-quality, filtered, and compliant open-source code data for training, with a dataset size of approximately 10-40% compared to other open-source code generation models. By following the provided official guidelines, developers can harness the power of the AquilaCode model to customize their own code assistant. | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-multi](https://model.baai.ac.cn/model-detail/100104) | Released | Nvidia-A100  |
+| AquilaCode-py   | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B, trained on Horizon Robotics chips | Same as above | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-py](https://model.baai.ac.cn/model-detail/100103)  | Released        | Nvidia-A100 |
 
 
 We will continue to release improved versions of Aquila model as open source. You can start by deleting the `checkpoints_in/aquilachat-7b` in the original directory and then download the new weights. Other usage methods remain unchanged. For more details, please refer to the folloing change log:
diff --git a/examples/Aquila/Aquila-code/README.md b/examples/Aquila/Aquila-code/README.md
index a2642210..61fd1a57 100755
--- a/examples/Aquila/Aquila-code/README.md
+++ b/examples/Aquila/Aquila-code/README.md
@@ -23,8 +23,8 @@
 | Aquila-33B          |基础模型，330亿参数  |    同上    | ——  | ——  | **敬请期待** | Nvidia-A100 | 
 | AquilaChat-7B          |SFT 模型，基于 Aquila-7B 进行微调和强化学习  |    **AquilaChat 对话模型**支持流畅的文本对话及多种语言类生成任务，通过定义可扩展的特殊指令规范，实现 AquilaChat对其它模型和工具的调用，且易于扩展。 <br><br>例如，调用智源开源的 **[AltDiffusion](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/AltDiffusion-m18) 多语言文图生成模型**，实现了流畅的文图生成能力。配合智源 **InstructFace 多步可控文生图模型**，轻松实现对人脸图像的多步可控编辑。  |   [./examples/Aquila/Aquila-chat](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-chat)  | [下载AquilaChat-7B](https://model.baai.ac.cn/model-detail/100101) | 已发布  | Nvidia-A100  | 
 | AquilaChat-33B           |SFT 模型，基于 Aquila-33B 进行微调和强化学习 |   同上    |   ——    |——  | **敬请期待** | Nvidia-A100 | 
-| AquilaCode-multi         | 基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练  |   AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型来定制自己的代码助手。  | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  |[下载AquilaCode-7B-multi](https://model.baai.ac.cn/model-detail/100102)  | 已发布  | Nvidia-A100 | 
-| AquilaCode-py           |基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练。  |    AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型定制自己的代码助手。    | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  | [下载AquilaCode-py](https://model.baai.ac.cn/model-detail/100099)  | 已发布  | Nvidia-A100  | 
+| AquilaCode-multi         | 基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练  |   AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型来定制自己的代码助手。  | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  |[下载AquilaCode-7B-multi](https://model.baai.ac.cn/model-detail/100104)  | 已发布  | Nvidia-A100 | 
+| AquilaCode-py           |基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练。  |    AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型定制自己的代码助手。    | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  | [下载AquilaCode-py](https://model.baai.ac.cn/model-detail/100103)  | 已发布  | Nvidia-A100  | 
 
 悟道·天鹰Aquila系列模型将持续开源更优版本，大家可以先删除原来目录下的code模型路径，再下载新权重，其他使用方式不变。
 
diff --git a/examples/Aquila/Aquila-code/README_en.md b/examples/Aquila/Aquila-code/README_en.md
index 2b6d16df..9c4e4bac 100755
--- a/examples/Aquila/Aquila-code/README_en.md
+++ b/examples/Aquila/Aquila-code/README_en.md
@@ -24,8 +24,8 @@ The additional details of the Aquila model will be presented in the official tec
 | Aquila-33B         | Base model, 33 billion parameters   | Same as above      | ——      | ——    | Coming soon                                               | Nvidia-A100   |
 | AquilaChat-7B      | SFT model, fine-tuned and RL based on Aquila-7B  | **AquilaChat Dialog Model** supports fluent text dialogue and multiple language generation tasks, and realizes the call of AquilaChat to other models and tools by defining an expandable special instruction specification, which is easy to extend. For example, calling the open source **[AltDiffusion](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/AltDiffusion-m18) multimodal language image generation model** of Flagship Intelligence achieved smooth image generation capability. Together with Flagship Intelligence's **InstructFace multi-step controllable text-picture model**, it is easy to achieve multi-step controllable editing of human face images. | [./examples/Aquila/Aquila-chat](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-chat) | [Download AquilaChat-7B](https://model.baai.ac.cn/model-detail/100101) | Released    | Nvidia-A100   |
 | AquilaChat-33B     | SFT model, fine-tuned and RL based on Aquila-33B  | Same as above| ——   | ——   |Coming soon                                               | Nvidia-A100   |
-| AquilaCode-multi  | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B.  | AquilaCode utilizes high-quality, filtered, and compliant open-source code data for training, with a dataset size of approximately 10-40% compared to other open-source code generation models. By following the provided official guidelines, developers can harness the power of the AquilaCode model to customize their own code assistant. | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-multi](https://model.baai.ac.cn/model-detail/100102) | Released | Nvidia-A100  |
-| AquilaCode-py   | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B, trained on Horizon Robotics chips | Same as above | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-py](https://model.baai.ac.cn/model-detail/100099)  | Released        | Nvidia-A100 |
+| AquilaCode-multi  | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B.  | AquilaCode utilizes high-quality, filtered, and compliant open-source code data for training, with a dataset size of approximately 10-40% compared to other open-source code generation models. By following the provided official guidelines, developers can harness the power of the AquilaCode model to customize their own code assistant. | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-multi](https://model.baai.ac.cn/model-detail/100104) | Released | Nvidia-A100  |
+| AquilaCode-py   | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B, trained on Horizon Robotics chips | Same as above | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-py](https://model.baai.ac.cn/model-detail/100103)  | Released        | Nvidia-A100 |
 
 
 We will continue to release improved versions of Aquila model as open source. You can start by deleting the checkpoint file in the original directory and then download the new weights. Other usage methods remain unchanged. For more details, please refer to the folloing change log:
diff --git a/examples/Aquila/Aquila-pretrain/README.md b/examples/Aquila/Aquila-pretrain/README.md
index eed216e1..743b6c18 100755
--- a/examples/Aquila/Aquila-pretrain/README.md
+++ b/examples/Aquila/Aquila-pretrain/README.md
@@ -22,8 +22,8 @@
 | Aquila-33B          |基础模型，330亿参数  |    同上    | ——  | ——  | **敬请期待** | Nvidia-A100 | 
 | AquilaChat-7B          |SFT 模型，基于 Aquila-7B 进行微调和强化学习  |    **AquilaChat 对话模型**支持流畅的文本对话及多种语言类生成任务，通过定义可扩展的特殊指令规范，实现 AquilaChat对其它模型和工具的调用，且易于扩展。 <br><br>例如，调用智源开源的 **[AltDiffusion](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/AltDiffusion-m18) 多语言文图生成模型**，实现了流畅的文图生成能力。配合智源 **InstructFace 多步可控文生图模型**，轻松实现对人脸图像的多步可控编辑。  |   [./examples/Aquila/Aquila-chat](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-chat)  | [下载AquilaChat-7B](https://model.baai.ac.cn/model-detail/100101) | 已发布  | Nvidia-A100  | 
 | AquilaChat-33B           |SFT 模型，基于 Aquila-33B 进行微调和强化学习 |   同上    |   ——    |——  | **敬请期待** | Nvidia-A100 | 
-| AquilaCode-multi         | 基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练  |   AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型来定制自己的代码助手。  | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  |[下载AquilaCode-7B-multi](https://model.baai.ac.cn/model-detail/100102)  | 已发布  | Nvidia-A100 | 
-| AquilaCode-py           |基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练。  |    AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型定制自己的代码助手。    | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  | [下载AquilaCode-py](https://model.baai.ac.cn/model-detail/100099)  | 已发布  | Nvidia-A100  | 
+| AquilaCode-multi         | 基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练  |   AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型来定制自己的代码助手。  | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  |[下载AquilaCode-7B-multi](https://model.baai.ac.cn/model-detail/100104)  | 已发布  | Nvidia-A100 | 
+| AquilaCode-py           |基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练。  |    AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型定制自己的代码助手。    | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  | [下载AquilaCode-py](https://model.baai.ac.cn/model-detail/100103)  | 已发布  | Nvidia-A100  | 
 
 悟道·天鹰Aquila系列模型将持续开源更优版本，大家可以先删除原来目录下的`checkpoints_in/aquila-7b`，再下载新权重，其他使用方式不变。
 
diff --git a/examples/Aquila/Aquila-pretrain/README_en.md b/examples/Aquila/Aquila-pretrain/README_en.md
index 1ba5c32b..989d26b2 100755
--- a/examples/Aquila/Aquila-pretrain/README_en.md
+++ b/examples/Aquila/Aquila-pretrain/README_en.md
@@ -23,8 +23,8 @@ The additional details of the Aquila model will be presented in the official tec
 | Aquila-33B         | Base model, 33 billion parameters   | Same as above      | ——      | ——    | Coming soon                                               | Nvidia-A100   |
 | AquilaChat-7B      | SFT model, fine-tuned and RL based on Aquila-7B  | **AquilaChat Dialog Model** supports fluent text dialogue and multiple language generation tasks, and realizes the call of AquilaChat to other models and tools by defining an expandable special instruction specification, which is easy to extend. For example, calling the open source **[AltDiffusion](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/AltDiffusion-m18) multimodal language image generation model** of Flagship Intelligence achieved smooth image generation capability. Together with Flagship Intelligence's **InstructFace multi-step controllable text-picture model**, it is easy to achieve multi-step controllable editing of human face images. | [./examples/Aquila/Aquila-chat](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-chat) | [Download AquilaChat-7B](https://model.baai.ac.cn/model-detail/100101) | Released    | Nvidia-A100   |
 | AquilaChat-33B     | SFT model, fine-tuned and RL based on Aquila-33B  | Same as above| ——   | ——   |Coming soon                                               | Nvidia-A100   |
-| AquilaCode-multi  | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B.  | AquilaCode utilizes high-quality, filtered, and compliant open-source code data for training, with a dataset size of approximately 10-40% compared to other open-source code generation models. By following the provided official guidelines, developers can harness the power of the AquilaCode model to customize their own code assistant. | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-multi](https://model.baai.ac.cn/model-detail/100102) | Released | Nvidia-A100  |
-| AquilaCode-py   | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B, trained on Horizon Robotics chips | Same as above | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-py](https://model.baai.ac.cn/model-detail/100099)  | Released        | Nvidia-A100 |
+| AquilaCode-multi  | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B.  | AquilaCode utilizes high-quality, filtered, and compliant open-source code data for training, with a dataset size of approximately 10-40% compared to other open-source code generation models. By following the provided official guidelines, developers can harness the power of the AquilaCode model to customize their own code assistant. | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-multi](https://model.baai.ac.cn/model-detail/100104) | Released | Nvidia-A100  |
+| AquilaCode-py   | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B, trained on Horizon Robotics chips | Same as above | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-py](https://model.baai.ac.cn/model-detail/100103)  | Released        | Nvidia-A100 |
 
 
 We will continue to release improved versions of Aquila model as open source. You can start by deleting the `checkpoints_in/aquila-7b` in the original directory and then download the new weights. Other usage methods remain unchanged. For more details, please refer to the folloing change log:
diff --git a/examples/Aquila/README.md b/examples/Aquila/README.md
index 7c468ed7..6228dbf6 100755
--- a/examples/Aquila/README.md
+++ b/examples/Aquila/README.md
@@ -23,8 +23,8 @@
 | Aquila-33B          |基础模型，330亿参数  |    同上    | ——  | ——  | **敬请期待** | Nvidia-A100 | 
 | AquilaChat-7B          |SFT 模型，基于 Aquila-7B 进行微调和强化学习  |    **AquilaChat 对话模型**支持流畅的文本对话及多种语言类生成任务，通过定义可扩展的特殊指令规范，实现 AquilaChat对其它模型和工具的调用，且易于扩展。 <br><br>例如，调用智源开源的 **[AltDiffusion](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/AltDiffusion-m18) 多语言文图生成模型**，实现了流畅的文图生成能力。配合智源 **InstructFace 多步可控文生图模型**，轻松实现对人脸图像的多步可控编辑。  |   [./examples/Aquila/Aquila-chat](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-chat)  | [下载AquilaChat-7B](https://model.baai.ac.cn/model-detail/100101) | 已发布  | Nvidia-A100  | 
 | AquilaChat-33B           |SFT 模型，基于 Aquila-33B 进行微调和强化学习 |   同上    |   ——    |——  | **敬请期待** | Nvidia-A100 | 
-| AquilaCode-multi         | 基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练  |   AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型来定制自己的代码助手。  | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  |[下载AquilaCode-7B-multi](https://model.baai.ac.cn/model-detail/100102)  | 已发布  | Nvidia-A100 | 
-| AquilaCode-py           |基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练。  |    AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型定制自己的代码助手。    | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  | [下载AquilaCode-py](https://model.baai.ac.cn/model-detail/100099)  | 已发布  | Nvidia-A100  |   
+| AquilaCode-multi         | 基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练  |   AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型来定制自己的代码助手。  | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  |[下载AquilaCode-7B-multi](https://model.baai.ac.cn/model-detail/100104)  | 已发布  | Nvidia-A100 | 
+| AquilaCode-py           |基础模型，“文本-代码”生成模型，基于 Aquila-7B继续预训练。  |    AquilaCode 使用经过高质量过滤且有合规开源许可的代码数据进行训练，数据量约为其他开源代码生成模型的 10～40%。通过参考官方提供的操作指南，开发者可以利用 AquilaCode 模型定制自己的代码助手。    | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code)  | [下载AquilaCode-py](https://model.baai.ac.cn/model-detail/100103)  | 已发布  | Nvidia-A100  |   
 
 悟道·天鹰Aquila系列模型将持续开源更优版本，大家可以先删除原来目录下的`checkpoints_in/aquila-7b`，再下载新权重，其他使用方式不变。
 
diff --git a/examples/Aquila/README_en.md b/examples/Aquila/README_en.md
index a2da3faf..5b6b7a94 100755
--- a/examples/Aquila/README_en.md
+++ b/examples/Aquila/README_en.md
@@ -25,8 +25,8 @@ The additional details of the Aquila model will be presented in the official tec
 | Aquila-33B         | Base model, 33 billion parameters   | Same as above      | ——      | ——    | Coming soon                                               | Nvidia-A100   |
 | AquilaChat-7B      | SFT model, fine-tuned and RL based on Aquila-7B  | **AquilaChat Dialog Model** supports fluent text dialogue and multiple language generation tasks, and realizes the call of AquilaChat to other models and tools by defining an expandable special instruction specification, which is easy to extend. For example, calling the open source **[AltDiffusion](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/AltDiffusion-m18) multimodal language image generation model** of Flagship Intelligence achieved smooth image generation capability. Together with Flagship Intelligence's **InstructFace multi-step controllable text-picture model**, it is easy to achieve multi-step controllable editing of human face images. | [./examples/Aquila/Aquila-chat](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-chat) | [Download AquilaChat-7B](https://model.baai.ac.cn/model-detail/100101) | Released    | Nvidia-A100   |
 | AquilaChat-33B     | SFT model, fine-tuned and RL based on Aquila-33B  | Same as above| ——   | ——   |Coming soon                                               | Nvidia-A100   |
-| AquilaCode-multi  | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B.  | AquilaCode utilizes high-quality, filtered, and compliant open-source code data for training, with a dataset size of approximately 10-40% compared to other open-source code generation models. By following the provided official guidelines, developers can harness the power of the AquilaCode model to customize their own code assistant. | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-multi](https://model.baai.ac.cn/model-detail/100102) | Released | Nvidia-A100  |
-| AquilaCode-py   | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B, trained on Horizon Robotics chips | Same as above | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-py](https://model.baai.ac.cn/model-detail/100099)  | Released        | Nvidia-A100 |
+| AquilaCode-multi  | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B.  | AquilaCode utilizes high-quality, filtered, and compliant open-source code data for training, with a dataset size of approximately 10-40% compared to other open-source code generation models. By following the provided official guidelines, developers can harness the power of the AquilaCode model to customize their own code assistant. | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-multi](https://model.baai.ac.cn/model-detail/100104) | Released | Nvidia-A100  |
+| AquilaCode-py   | Base model, "text-code" generation model, continue-pre-trained based on Aquila-7B, trained on Horizon Robotics chips | Same as above | [./examples/Aquila/Aquila-code](https://github.com/FlagAI-Open/FlagAI/tree/master/examples/Aquila/Aquila-code) | [Download AquilaCode-py](https://model.baai.ac.cn/model-detail/100103)  | Released        | Nvidia-A100 |
 
 
 We will continue to release improved versions of Aquila model as open source. You can start by deleting the `checkpoints_in/aquila-7b` in the original directory and then download the new weights. Other usage methods remain unchanged. For more details, please refer to the folloing change log: