feat: update README
Browse files
README.md
CHANGED
|
@@ -112,6 +112,13 @@ tags:
|
|
| 112 |
</p>
|
| 113 |
|
| 114 |
# Jina-ColBERT-v2
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 115 |
|
| 116 |
## Usage
|
| 117 |
|
|
|
|
| 112 |
</p>
|
| 113 |
|
| 114 |
# Jina-ColBERT-v2
|
| 115 |
+
Jina ColBERT v2 (`jina-colbert-v2`) is a new model based on the [Jina-ColBERT architecture](https://jina.ai/news/what-is-colbert-and-late-interaction-and-why-they-matter-in-search/) that expands on the capabilities and performance of the `jina-colbert-v1-en` model. Like the previous release, it has Jina AI’s 8192 token input context and the [improved efficiency, performance](https://jina.ai/news/what-is-colbert-and-late-interaction-and-why-they-matter-in-search/), and [explainability](https://jina.ai/news/ai-explainability-made-easy-how-late-interaction-makes-jina-colbert-transparent/) of token-level embeddings and late interaction.
|
| 116 |
+
|
| 117 |
+
This new release adds new functionality and performance improvements:
|
| 118 |
+
|
| 119 |
+
- Multilingual support for dozens of languages, with strong performance on major global languages.
|
| 120 |
+
- [Matryoshka embeddings](https://huggingface.co/blog/matryoshka), which allow users to trade between efficiency and precision flexibly.
|
| 121 |
+
- Superior retrieval performance when compared to the English-only `jina-colbert-v1-en`.
|
| 122 |
|
| 123 |
## Usage
|
| 124 |
|