|
Download docs/en/use/knowledge-base.md from qa1145/astrbbbb: direct link, hf CLI and curl.
- Browser
- Download file 2.72 kB
-
https://huggingface.co/spaces/qa1145/astrbbbb/resolve/main/docs/en/use/knowledge-base.md
- Command line
-
hf download hf://spaces/qa1145/astrbbbb/docs/en/use/knowledge-base.md
-
curl -L -o knowledge-base.md https://huggingface.co/spaces/qa1145/astrbbbb/resolve/main/docs/en/use/knowledge-base.md
2.72 kB
| # AstrBot Knowledge Base | |
| > [!TIP] | |
| > Requires AstrBot version >= 4.5.0. | |
|  | |
| ## Configuring Embedding Model | |
| Open the service provider page, click "Add Service Provider", and select Embedding. | |
| Currently, AstrBot supports embedding vector services compatible with OpenAI API and Gemini API. | |
| Click on the provider card above to enter the configuration page and fill in the configuration. | |
| After completing the configuration, click Save. | |
| ## Configuring Reranker Model (Optional) | |
| A reranker model can improve the precision of final retrieval results to some extent. | |
| Similar to configuring the embedding model, open the service provider page, click "Add Service Provider", and select Reranker. For more information about reranker models, please refer to online resources. | |
| ## Creating a Knowledge Base | |
| AstrBot supports multiple knowledge base management. During chat, you can **freely specify which knowledge base to use**. | |
| Enter the knowledge base page and click "Create Knowledge Base", as shown below: | |
|  | |
| Fill in the relevant information. In the embedding model dropdown menu, you will see the embedding model and reranker model you just created (reranker model is optional). | |
| > [!TIP] | |
| > Once you've selected an embedding model for a knowledge base, do not modify the **model** or **vector dimension information** of that provider, as this will **seriously affect** the retrieval accuracy of the knowledge base or even **cause errors**. | |
| ## Uploading Files | |
| After creating a knowledge base, you can upload documents to it. Up to 10 files can be uploaded simultaneously, with a maximum size of 128 MB per file. | |
|  | |
| ## Using the Knowledge Base | |
| In the configuration file, you can specify different knowledge bases for different configuration profiles. | |
| ## Appendix 2: Applying for Free Embedding Models | |
| ### PPIO Cloud | |
| 1. Open the [PPIO Cloud website](https://ppio.cn/user/register?invited_by=AIOONE) and register an account (accounts registered through this link will receive a 15 RMB voucher). | |
| 2. Go to the [Model Marketplace](https://ppio.cn/model-api/console) and click on Embedding Models. | |
| 3. Click on BAAI:BGE-M3 (as of 2025-06-02, this model is free on this platform). | |
| 4. Find the API integration guide and apply for a Key. | |
| 5. Fill in the AstrBot OpenAI Embedding model provider configuration: | |
| 1. API Key is the PPIO API Key you just applied for | |
| 2. embedding api base: enter `https://api.ppinfra.com/v3/openai` | |
| 3. model: enter the model you selected, in this example `baai/bge-m3`. | |