[Doc]: 请教：关于打造自己的知识库

[369663841]

README 中的表述如下：

1.将个人知识文件或者文件夹放入pilot/datasets目录中 ... 3.在tools目录执行知识入库脚本（） 如果是选择默认知识库，不需要指定 --vector_name, 默认default python tools/knowledge_init.py 如果选择新增知识库，在界面上新增知识库输入你的知识库名, python tools/knowledge_init.py --vector_name = yourname --vector_name: vector_name default_value:default 就可以根据你的知识库进行问答

没有看明白的是： 1、界面上的上传文档功能 与 『1.将个人知识文件或者文件夹放入pilot/datasets目录中』操作，两者 是否实现同样的作用？

2、执行 python tools/knowledge_init.py --vector_name = yourname ，是在『如果选择新增知识库，在界面上新增知识库输入你的知识库名,』 操作之后吗？ 即在界面输入了新的知识库名，才执行 python tools/knowledge_init.py --vector_name = yourname ？

3、在界面上输入新的知识库名，然后按『Add as new documents』 ，界面上没有反应，多按几下，界面会显示报错，但没有具体的错误信息，但执行 python tools/knowledge_init.py --vector_name = yourname 是成功的。请问是何原因？

[Aries-ckt]

README 中的表述如下：

1.将个人知识文件或者文件夹放入pilot/datasets目录中 ... 3.在tools目录执行知识入库脚本（） 如果是选择默认知识库，不需要指定 --vector_name, 默认default python tools/knowledge_init.py 如果选择新增知识库，在界面上新增知识库输入你的知识库名, python tools/knowledge_init.py --vector_name = yourname --vector_name: vector_name default_value:default 就可以根据你的知识库进行问答

没有看明白的是： 1、界面上的上传文档功能 与 『1.将个人知识文件或者文件夹放入pilot/datasets目录中』操作，两者 是否实现同样的作用？

2、执行 python tools/knowledge_init.py --vector_name = yourname ，是在『如果选择新增知识库，在界面上新增知识库输入你的知识库名,』 操作之后吗？ 即在界面输入了新的知识库名，才执行 python tools/knowledge_init.py --vector_name = yourname ？

3、在界面上输入新的知识库名，然后按『Add as new documents』 ，界面上没有反应，多按几下，界面会显示报错，但没有具体的错误信息，但执行 python tools/knowledge_init.py --vector_name = yourname 是成功的。请问是何原因？

Uploading documents on the interface is only limited to some small documents. If there are many and large documents involved, it is recommended to use the python tools/knowledge_init.py --vector_name = yourname command.

[369663841]

README 中的表述如下：

1.将个人知识文件或者文件夹放入pilot/datasets目录中 ... 3.在tools目录执行知识入库脚本（） 如果是选择默认知识库，不需要指定 --vector_name, 默认default python tools/knowledge_init.py 如果选择新增知识库，在界面上新增知识库输入你的知识库名, python tools/knowledge_init.py --vector_name = yourname --vector_name: vector_name default_value:default 就可以根据你的知识库进行问答

没有看明白的是： 1、界面上的上传文档功能 与 『1.将个人知识文件或者文件夹放入pilot/datasets目录中』操作，两者 是否实现同样的作用？ 2、执行 python tools/knowledge_init.py --vector_name = yourname ，是在『如果选择新增知识库，在界面上新增知识库输入你的知识库名,』 操作之后吗？ 即在界面输入了新的知识库名，才执行 python tools/knowledge_init.py --vector_name = yourname ？ 3、在界面上输入新的知识库名，然后按『Add as new documents』 ，界面上没有反应，多按几下，界面会显示报错，但没有具体的错误信息，但执行 python tools/knowledge_init.py --vector_name = yourname 是成功的。请问是何原因？

Uploading documents on the interface is only limited to some small documents. If there are many and large documents involved, it is recommended to use the python tools/knowledge_init.py --vector_name = yourname command.

Thank you for your reply! My steps of operation are as follows:

1. I placed the documents of relevant materials in the pilot/datasets directory.
2. I executed python tools/knowledge_init.py --vector_name=yourname (the actual file name is knowlege_init.py, not knowledge_init.py), and it executed successfully.
3. On the interface, I selected "New documents", and in the "New document name" input box under Configure Documents, I entered the actual name for 'yourname' (I did not press the "Add as new documents" button, and there was no response when I pressed it).
4. Then I asked questions based on the content of my materials, but the responses were completely unrelated to the questions.

Could you please let me know if there are any issues with my steps of operation?

[Aries-ckt]

README 中的表述如下：

1.将个人知识文件或者文件夹放入pilot/datasets目录中 ... 3.在tools目录执行知识入库脚本（） 如果是选择默认知识库，不需要指定 --vector_name, 默认default python tools/knowledge_init.py 如果选择新增知识库，在界面上新增知识库输入你的知识库名, python tools/knowledge_init.py --vector_name = yourname --vector_name: vector_name default_value:default 就可以根据你的知识库进行问答

没有看明白的是： 1、界面上的上传文档功能 与 『1.将个人知识文件或者文件夹放入pilot/datasets目录中』操作，两者 是否实现同样的作用？ 2、执行 python tools/knowledge_init.py --vector_name = yourname ，是在『如果选择新增知识库，在界面上新增知识库输入你的知识库名,』 操作之后吗？ 即在界面输入了新的知识库名，才执行 python tools/knowledge_init.py --vector_name = yourname ？ 3、在界面上输入新的知识库名，然后按『Add as new documents』 ，界面上没有反应，多按几下，界面会显示报错，但没有具体的错误信息，但执行 python tools/knowledge_init.py --vector_name = yourname 是成功的。请问是何原因？

Uploading documents on the interface is only limited to some small documents. If there are many and large documents involved, it is recommended to use the python tools/knowledge_init.py --vector_name = yourname command.

Thank you for your reply! My steps of operation are as follows:

1. I placed the documents of relevant materials in the pilot/datasets directory.
2. I executed python tools/knowledge_init.py --vector_name=yourname (the actual file name is knowlege_init.py, not knowledge_init.py), and it executed successfully.
3. On the interface, I selected "New documents", and in the "New document name" input box under Configure Documents, I entered the actual name for 'yourname' (I did not press the "Add as new documents" button, and there was no response when I pressed it).
4. Then I asked questions based on the content of my materials, but the responses were completely unrelated to the questions.

Could you please let me know if there are any issues with my steps of operation?

Have there been any error reports in the logs of your webserver backend? Can you see logs such as 'prompt': 'You are an AI designed to answer human questions, please follow the prompts and conventions of the system's input for your answers###human:'?

[369663841]

README 中的表述如下：

1.将个人知识文件或者文件夹放入pilot/datasets目录中 ... 3.在tools目录执行知识入库脚本（） 如果是选择默认知识库，不需要指定 --vector_name, 默认default python tools/knowledge_init.py 如果选择新增知识库，在界面上新增知识库输入你的知识库名, python tools/knowledge_init.py --vector_name = yourname --vector_name: vector_name default_value:default 就可以根据你的知识库进行问答

没有看明白的是： 1、界面上的上传文档功能 与 『1.将个人知识文件或者文件夹放入pilot/datasets目录中』操作，两者 是否实现同样的作用？ 2、执行 python tools/knowledge_init.py --vector_name = yourname ，是在『如果选择新增知识库，在界面上新增知识库输入你的知识库名,』 操作之后吗？ 即在界面输入了新的知识库名，才执行 python tools/knowledge_init.py --vector_name = yourname ？ 3、在界面上输入新的知识库名，然后按『Add as new documents』 ，界面上没有反应，多按几下，界面会显示报错，但没有具体的错误信息，但执行 python tools/knowledge_init.py --vector_name = yourname 是成功的。请问是何原因？

Uploading documents on the interface is only limited to some small documents. If there are many and large documents involved, it is recommended to use the python tools/knowledge_init.py --vector_name = yourname command.

Thank you for your reply! My steps of operation are as follows:

1. I placed the documents of relevant materials in the pilot/datasets directory.
2. I executed python tools/knowledge_init.py --vector_name=yourname (the actual file name is knowlege_init.py, not knowledge_init.py), and it executed successfully.
3. On the interface, I selected "New documents", and in the "New document name" input box under Configure Documents, I entered the actual name for 'yourname' (I did not press the "Add as new documents" button, and there was no response when I pressed it).
4. Then I asked questions based on the content of my materials, but the responses were completely unrelated to the questions.

Could you please let me know if there are any issues with my steps of operation?

Have there been any error reports in the logs of your webserver backend? Can you see logs such as 'prompt': 'You are an AI designed to answer human questions, please follow the prompts and conventions of the system's input for your answers###human:'?

Thank you! I only encountered an error previously when executing "python tools/knowledge_init.py --vector_name=yourname" . Later, after installing text2vec-large-chinese and changing knowledge_init.py to knowlede_init.py, the command executed normally. In the subsequent questions and answers, I didn't encounter any error messages. There were no error reports in the web backend either, and I could see logs with "###human:".

The issue is that the content of the responses is completely unrelated to the questions. I am not sure whether it’s because my knowledge base wasn't set up properly, or if I didn't correctly select the knowledge base that I added when asking questions. I have already entered the name of my knowledge base in the "New document name" input box under "New documents" in Configure Documents.

[Aries-ckt]

README 中的表述如下：

1.将个人知识文件或者文件夹放入pilot/datasets目录中 ... 3.在tools目录执行知识入库脚本（） 如果是选择默认知识库，不需要指定 --vector_name, 默认default python tools/knowledge_init.py 如果选择新增知识库，在界面上新增知识库输入你的知识库名, python tools/knowledge_init.py --vector_name = yourname --vector_name: vector_name default_value:default 就可以根据你的知识库进行问答

没有看明白的是： 1、界面上的上传文档功能 与 『1.将个人知识文件或者文件夹放入pilot/datasets目录中』操作，两者 是否实现同样的作用？ 2、执行 python tools/knowledge_init.py --vector_name = yourname ，是在『如果选择新增知识库，在界面上新增知识库输入你的知识库名,』 操作之后吗？ 即在界面输入了新的知识库名，才执行 python tools/knowledge_init.py --vector_name = yourname ？ 3、在界面上输入新的知识库名，然后按『Add as new documents』 ，界面上没有反应，多按几下，界面会显示报错，但没有具体的错误信息，但执行 python tools/knowledge_init.py --vector_name = yourname 是成功的。请问是何原因？

Uploading documents on the interface is only limited to some small documents. If there are many and large documents involved, it is recommended to use the python tools/knowledge_init.py --vector_name = yourname command.

Thank you for your reply! My steps of operation are as follows:

1. I placed the documents of relevant materials in the pilot/datasets directory.
2. I executed python tools/knowledge_init.py --vector_name=yourname (the actual file name is knowlege_init.py, not knowledge_init.py), and it executed successfully.
3. On the interface, I selected "New documents", and in the "New document name" input box under Configure Documents, I entered the actual name for 'yourname' (I did not press the "Add as new documents" button, and there was no response when I pressed it).
4. Then I asked questions based on the content of my materials, but the responses were completely unrelated to the questions.

Could you please let me know if there are any issues with my steps of operation?

Have there been any error reports in the logs of your webserver backend? Can you see logs such as 'prompt': 'You are an AI designed to answer human questions, please follow the prompts and conventions of the system's input for your answers###human:'?

Thank you! I only encountered an error previously when executing "python tools/knowledge_init.py --vector_name=yourname" . Later, after installing text2vec-large-chinese and changing knowledge_init.py to knowlede_init.py, the command executed normally. In the subsequent questions and answers, I didn't encounter any error messages. There were no error reports in the web backend either, and I could see logs with "###human:".

The issue is that the content of the responses is completely unrelated to the questions. I am not sure whether it’s because my knowledge base wasn't set up properly, or if I didn't correctly select the knowledge base that I added when asking questions. I have already entered the name of my knowledge base in the "New document name" input box under "New documents" in Configure Documents.

ok, you can check the webserver log, the prompt('prompt': 'You are an AI designed to answer human questions, please follow the prompts and conventions of the system's input for your answers###human:'? ) will contains the topK query answers,then the prompt will send to model server and wait for model server response.

[369663841]

README 中的表述如下：

1.将个人知识文件或者文件夹放入pilot/datasets目录中 ... 3.在tools目录执行知识入库脚本（） 如果是选择默认知识库，不需要指定 --vector_name, 默认default python tools/knowledge_init.py 如果选择新增知识库，在界面上新增知识库输入你的知识库名, python tools/knowledge_init.py --vector_name = yourname --vector_name: vector_name default_value:default 就可以根据你的知识库进行问答

没有看明白的是： 1、界面上的上传文档功能 与 『1.将个人知识文件或者文件夹放入pilot/datasets目录中』操作，两者 是否实现同样的作用？ 2、执行 python tools/knowledge_init.py --vector_name = yourname ，是在『如果选择新增知识库，在界面上新增知识库输入你的知识库名,』 操作之后吗？ 即在界面输入了新的知识库名，才执行 python tools/knowledge_init.py --vector_name = yourname ？ 3、在界面上输入新的知识库名，然后按『Add as new documents』 ，界面上没有反应，多按几下，界面会显示报错，但没有具体的错误信息，但执行 python tools/knowledge_init.py --vector_name = yourname 是成功的。请问是何原因？

Uploading documents on the interface is only limited to some small documents. If there are many and large documents involved, it is recommended to use the python tools/knowledge_init.py --vector_name = yourname command.

Thank you for your reply! My steps of operation are as follows:

1. I placed the documents of relevant materials in the pilot/datasets directory.
2. I executed python tools/knowledge_init.py --vector_name=yourname (the actual file name is knowlege_init.py, not knowledge_init.py), and it executed successfully.
3. On the interface, I selected "New documents", and in the "New document name" input box under Configure Documents, I entered the actual name for 'yourname' (I did not press the "Add as new documents" button, and there was no response when I pressed it).
4. Then I asked questions based on the content of my materials, but the responses were completely unrelated to the questions.

Could you please let me know if there are any issues with my steps of operation?

Have there been any error reports in the logs of your webserver backend? Can you see logs such as 'prompt': 'You are an AI designed to answer human questions, please follow the prompts and conventions of the system's input for your answers###human:'?

Thank you! I only encountered an error previously when executing "python tools/knowledge_init.py --vector_name=yourname" . Later, after installing text2vec-large-chinese and changing knowledge_init.py to knowlede_init.py, the command executed normally. In the subsequent questions and answers, I didn't encounter any error messages. There were no error reports in the web backend either, and I could see logs with "###human:". The issue is that the content of the responses is completely unrelated to the questions. I am not sure whether it’s because my knowledge base wasn't set up properly, or if I didn't correctly select the knowledge base that I added when asking questions. I have already entered the name of my knowledge base in the "New document name" input box under "New documents" in Configure Documents.

ok, you can check the webserver log, the prompt('prompt': 'You are an AI designed to answer human questions, please follow the prompts and conventions of the system's input for your answers###human:'? ) will contains the topK query answers,then the prompt will send to model server and wait for model server response.

Thank you! I have placed pdf documents titled "User Operation Manual" and "Training Handouts" related to the "＊＊System" in the pilot/datasets directory. These documents contain information about the features included in the "＊＊System". When I asked the question, "What features does the ＊＊System have? (请问＊＊系统有什么功能？)", the web backend log is as follows:

output: You are an AI designed to answer human questions, please follow the prompts and conventions of the system's input for your answers###human:请问＊＊系统有什么功能？###system: Based on the known information below, provide users with professional and concise answers to their questions. If the answer cannot be obtained from the provided content, please say: "The information provided in the knowledge base is not sufficient to answer this question." It is forbidden to make up information randomly. known information: ['它包括一个向量存储索引和一个简单的目录阅读器，可以帮助您处理和操作数据。'] question: 请问＊＊系统有什么功能？ ###system: ＊＊系统是一个数据处理和操作工具，它包括一个向量存储索引和一个简单的目录阅读器。这些功能可以帮助您处理和操作 数据。

The response it gave, "＊＊System is a data processing and operation tool, which includes a vector storage index and a simple directory reader. These features can help you process and manipulate data.", is completely unrelated to the contents of the "User Operation Manual" and "Training Handouts" documents that I provided. Does this indicate that my knowledge base was not successfully established? Or does it mean that it did not base its answer on my knowledge base, but used a default knowledge base instead?

[Aries-ckt]

README 中的表述如下：

1.将个人知识文件或者文件夹放入pilot/datasets目录中 ... 3.在tools目录执行知识入库脚本（） 如果是选择默认知识库，不需要指定 --vector_name, 默认default python tools/knowledge_init.py 如果选择新增知识库，在界面上新增知识库输入你的知识库名, python tools/knowledge_init.py --vector_name = yourname --vector_name: vector_name default_value:default 就可以根据你的知识库进行问答

没有看明白的是： 1、界面上的上传文档功能 与 『1.将个人知识文件或者文件夹放入pilot/datasets目录中』操作，两者 是否实现同样的作用？ 2、执行 python tools/knowledge_init.py --vector_name = yourname ，是在『如果选择新增知识库，在界面上新增知识库输入你的知识库名,』 操作之后吗？ 即在界面输入了新的知识库名，才执行 python tools/knowledge_init.py --vector_name = yourname ？ 3、在界面上输入新的知识库名，然后按『Add as new documents』 ，界面上没有反应，多按几下，界面会显示报错，但没有具体的错误信息，但执行 python tools/knowledge_init.py --vector_name = yourname 是成功的。请问是何原因？

Uploading documents on the interface is only limited to some small documents. If there are many and large documents involved, it is recommended to use the python tools/knowledge_init.py --vector_name = yourname command.

Thank you for your reply! My steps of operation are as follows:

1. I placed the documents of relevant materials in the pilot/datasets directory.
2. I executed python tools/knowledge_init.py --vector_name=yourname (the actual file name is knowlege_init.py, not knowledge_init.py), and it executed successfully.
3. On the interface, I selected "New documents", and in the "New document name" input box under Configure Documents, I entered the actual name for 'yourname' (I did not press the "Add as new documents" button, and there was no response when I pressed it).
4. Then I asked questions based on the content of my materials, but the responses were completely unrelated to the questions.

Could you please let me know if there are any issues with my steps of operation?

Have there been any error reports in the logs of your webserver backend? Can you see logs such as 'prompt': 'You are an AI designed to answer human questions, please follow the prompts and conventions of the system's input for your answers###human:'?

Thank you! I only encountered an error previously when executing "python tools/knowledge_init.py --vector_name=yourname" . Later, after installing text2vec-large-chinese and changing knowledge_init.py to knowlede_init.py, the command executed normally. In the subsequent questions and answers, I didn't encounter any error messages. There were no error reports in the web backend either, and I could see logs with "###human:". The issue is that the content of the responses is completely unrelated to the questions. I am not sure whether it’s because my knowledge base wasn't set up properly, or if I didn't correctly select the knowledge base that I added when asking questions. I have already entered the name of my knowledge base in the "New document name" input box under "New documents" in Configure Documents.

ok, you can check the webserver log, the prompt('prompt': 'You are an AI designed to answer human questions, please follow the prompts and conventions of the system's input for your answers###human:'? ) will contains the topK query answers,then the prompt will send to model server and wait for model server response.

Thank you! I have placed pdf documents titled "User Operation Manual" and "Training Handouts" related to the "＊＊System" in the pilot/datasets directory. These documents contain information about the features included in the "＊＊System". When I asked the question, "What features does the ＊＊System have? (请问＊＊系统有什么功能？)", the web backend log is as follows:

output: You are an AI designed to answer human questions, please follow the prompts and conventions of the system's input for your answers###human:请问＊＊系统有什么功能？###system: Based on the known information below, provide users with professional and concise answers to their questions. If the answer cannot be obtained from the provided content, please say: "The information provided in the knowledge base is not sufficient to answer this question." It is forbidden to make up information randomly. known information: ['它包括一个向量存储索引和一个简单的目录阅读器，可以帮助您处理和操作数据。'] question: 请问＊＊系统有什么功能？ ###system: ＊＊系统是一个数据处理和操作工具，它包括一个向量存储索引和一个简单的目录阅读器。这些功能可以帮助您处理和操作 数据。

The response it gave, "＊＊System is a data processing and operation tool, which includes a vector storage index and a simple directory reader. These features can help you process and manipulate data.", is completely unrelated to the contents of the "User Operation Manual" and "Training Handouts" documents that I provided. Does this indicate that my knowledge base was not successfully established? Or does it mean that it did not base its answer on my knowledge base, but used a default knowledge base instead?

what's your embedding EMBEDDINGS_MODEL? and what's your LLM_MODEL?

[369663841]

README 中的表述如下：

1.将个人知识文件或者文件夹放入pilot/datasets目录中 ... 3.在tools目录执行知识入库脚本（） 如果是选择默认知识库，不需要指定 --vector_name, 默认default python tools/knowledge_init.py 如果选择新增知识库，在界面上新增知识库输入你的知识库名, python tools/knowledge_init.py --vector_name = yourname --vector_name: vector_name default_value:default 就可以根据你的知识库进行问答

没有看明白的是： 1、界面上的上传文档功能 与 『1.将个人知识文件或者文件夹放入pilot/datasets目录中』操作，两者 是否实现同样的作用？ 2、执行 python tools/knowledge_init.py --vector_name = yourname ，是在『如果选择新增知识库，在界面上新增知识库输入你的知识库名,』 操作之后吗？ 即在界面输入了新的知识库名，才执行 python tools/knowledge_init.py --vector_name = yourname ？ 3、在界面上输入新的知识库名，然后按『Add as new documents』 ，界面上没有反应，多按几下，界面会显示报错，但没有具体的错误信息，但执行 python tools/knowledge_init.py --vector_name = yourname 是成功的。请问是何原因？

Uploading documents on the interface is only limited to some small documents. If there are many and large documents involved, it is recommended to use the python tools/knowledge_init.py --vector_name = yourname command.

Thank you for your reply! My steps of operation are as follows:

1. I placed the documents of relevant materials in the pilot/datasets directory.
2. I executed python tools/knowledge_init.py --vector_name=yourname (the actual file name is knowlege_init.py, not knowledge_init.py), and it executed successfully.
3. On the interface, I selected "New documents", and in the "New document name" input box under Configure Documents, I entered the actual name for 'yourname' (I did not press the "Add as new documents" button, and there was no response when I pressed it).
4. Then I asked questions based on the content of my materials, but the responses were completely unrelated to the questions.

Could you please let me know if there are any issues with my steps of operation?

Have there been any error reports in the logs of your webserver backend? Can you see logs such as 'prompt': 'You are an AI designed to answer human questions, please follow the prompts and conventions of the system's input for your answers###human:'?

Thank you! I only encountered an error previously when executing "python tools/knowledge_init.py --vector_name=yourname" . Later, after installing text2vec-large-chinese and changing knowledge_init.py to knowlede_init.py, the command executed normally. In the subsequent questions and answers, I didn't encounter any error messages. There were no error reports in the web backend either, and I could see logs with "###human:". The issue is that the content of the responses is completely unrelated to the questions. I am not sure whether it’s because my knowledge base wasn't set up properly, or if I didn't correctly select the knowledge base that I added when asking questions. I have already entered the name of my knowledge base in the "New document name" input box under "New documents" in Configure Documents.

ok, you can check the webserver log, the prompt('prompt': 'You are an AI designed to answer human questions, please follow the prompts and conventions of the system's input for your answers###human:'? ) will contains the topK query answers,then the prompt will send to model server and wait for model server response.

Thank you! I have placed pdf documents titled "User Operation Manual" and "Training Handouts" related to the "＊＊System" in the pilot/datasets directory. These documents contain information about the features included in the "＊＊System". When I asked the question, "What features does the ＊＊System have? (请问＊＊系统有什么功能？)", the web backend log is as follows: output: You are an AI designed to answer human questions, please follow the prompts and conventions of the system's input for your answers###human:请问＊＊系统有什么功能？###system: Based on the known information below, provide users with professional and concise answers to their questions. If the answer cannot be obtained from the provided content, please say: "The information provided in the knowledge base is not sufficient to answer this question." It is forbidden to make up information randomly. known information: ['它包括一个向量存储索引和一个简单的目录阅读器，可以帮助您处理和操作数据。'] question: 请问＊＊系统有什么功能？ ###system: ＊＊系统是一个数据处理和操作工具，它包括一个向量存储索引和一个简单的目录阅读器。这些功能可以帮助您处理和操作 数据。 The response it gave, "＊＊System is a data processing and operation tool, which includes a vector storage index and a simple directory reader. These features can help you process and manipulate data.", is completely unrelated to the contents of the "User Operation Manual" and "Training Handouts" documents that I provided. Does this indicate that my knowledge base was not successfully established? Or does it mean that it did not base its answer on my knowledge base, but used a default knowledge base instead?

what's your embedding EMBEDDINGS_MODEL? and what's your LLM_MODEL?

Thank you! In the .env file, LLM_MODEL=vicuna-13b and EMBEDDING_MODEL=text2vec

[csunny]

这个issue将被关闭，欢迎体验我们最新的V0.4.0版本。