OpenSearch Vector Search Edition supports built-in models and custom models. You can view and manage the models on the Models page of the OpenSearch console.
Custom models
If you want to call a custom model, you can add the model in the OpenSearch console and select the model when you configure tables. For more information, see Custom models.
Built-in models
Model ID | Model name | Model type | Model description |
clip | General-purpose Image-to-Vector Model - 512 Dimensions | Image vectorization | Serves as a general-purpose image-to-vector model that supports both image-to-image and text-to-image search scenarios. The model has an output vector dimension of 512. |
clip_ecom | Enhanced Image-to-Vector Model for E-commerce - 512 Dimensions | Image vectorization | Serves as an image-to-vector model specifically enhanced for e-commerce scenarios. The model supports both image-to-image and text-to-image search scenarios and has an output vector dimension of 512. |
ops-text-embedding-1024-000-20231001 | Enhanced Chinese Text-to-Vector Model - 768 Dimensions | Text vectorization | Supports the conversion of Chinese text into vector representations. The model accepts Chinese text with a maximum length of 1,024 tokens and has an output vector dimension of 768. |
ops-text-embedding-512-000-20231001 | Enhanced Chinese Text-to-Vector Model - 768 Dimensions | Text vectorization | Supports the conversion of Chinese text into vector representations. The model accepts Chinese text with a maximum length of 512 tokens and has an output vector dimension of 768. |
ops-text-embedding-128-000-20231001 | Enhanced Chinese Text-to-Vector Model - 768 Dimensions | Text vectorization | Supports the conversion of Chinese text into vector representations. The model accepts Chinese text with a maximum length of 128 tokens and has an output vector dimension of 768. |
ops-text-embedding-512-en-000-20231001 | Enhanced English Text-to-Vector Model - 768 Dimensions | Text vectorization | Supports the conversion of English text into vector representations. The model accepts English text with a maximum length of 512 tokens and has an output vector dimension of 768. |
ops-text-embedding-128-en-000-20231001 | Enhanced English Text-to-Vector Model - 768 Dimensions | Text vectorization | Supports the conversion of English text into vector representations. The model accepts English text with a maximum length of 128 tokens and has an output vector dimension of 768. |