Update README.md

README.md

@@ -2766,6 +2766,87 @@ print(similarities)
 
 The API comes with native INT8 and binary quantization support! Check out the [docs](https://mixedbread.ai/docs) for more information.
 
+### Binary Matryoshka Representation Learning
+
+We offer both MRL and binary quantization through our API, and they're also supported through Sentence Transformers. Please see examples on how to use these features below:
+
+<CodeGroup title="Installation">
+    ```bash {{title: "Python (API)"}}
+    pip install -U mixedbread-ai
+    ```
+
+    ```bash {{title: "JavaScript (API)"}}
+    npm i @mixedbread-ai/sdk
+    ```
+
+    ```bash {{title: "Python"}}
+    pip install -U sentence-transformers
+    ```
+</CodeGroup>
+
+<CodeGroup title="Usage Example">
+```python {{title: "Python (API)"}}
+from mixedbread_ai.client import MixedbreadAI
+
+mxbai = MixedbreadAI(api_key="{MIXEDBREAD_API_KEY}")
+
+res = mxbai.embeddings(
+    model='mixedbread-ai/mxbai-embed-large-v1',
+    input=[
+        'Who is german and likes bread?',
+        'Everybody in Germany.'
+    ],
+    normalized=True, # this has to be True if you want to use binary with faiss
+    encoding_format=['ubinary', 'float'],
+    dimensions=512
+)
+
+res.data[0].embedding.ubinary
+res.data[0].embedding.float_
+res.dimensions
+```
+
+```javascript {{title: "JavaScript (API)"}}
+import { MixedbreadAIClient } from "@mixedbread-ai/sdk";
+
+const mxbai = new MixedbreadAIClient({
+  apiKey: "{MIXEDBREAD_API_KEY}"
+});
+
+const res = await mxbai.embeddings({
+  model: 'mixedbread-ai/mxbai-embed-large-v1',
+  input: [
+    'Who is german and likes bread?',
+    'Everybody in Germany.'
+  ],
+  normalized: true, // this has to be True if you want to use binary with faiss
+  encoding_format: ['ubinary', 'float'],
+  dimensions=512
+})
+
+console.log(res.data[0].embedding.ubinary, res.data[0].embedding.float, res.dimensions)
+```
+
+```python {{title: "Python"}}
+from sentence_transformers import SentenceTransformer
+from sentence_transformers.quantization import quantize_embeddings
+
+# 1. Load an embedding model
+model = SentenceTransformer("mixedbread-ai/mxbai-embed-large-v1")
+
+# 2. Encode some text and select MRL dimensions
+mrl_embeddings = model.encode(
+    ["Who is german and likes bread?", "Everybody in Germany."], normalize_embeddings=True)[..., :512] 
+
+# 3. Apply binary quantization
+binary_embeddings = quantize_embeddings(mrl_embeddings, precision="binary")
+```
+</CodeGroup>
+
+### Why is this relevant?
+
+Binary MRL allows for significant savings in memory usage and therefore money when using a vector database. Read more about the technology and its advantages in our [blog post](https://www.mixedbread.ai/blog/binary-mrl).
+
 ## Evaluation
 As of March 2024, our model archives SOTA performance for Bert-large sized models on the [MTEB](https://huggingface.co/spaces/mteb/leaderboard). It ourperforms commercial models like OpenAIs text-embedding-3-large and matches the performance of model 20x it's size like the [echo-mistral-7b](https://huggingface.co/jspringer/echo-mistral-7b-instruct-lasttoken). Our model was trained with no overlap of the MTEB data, which indicates that our model generalizes well across several domains, tasks and text length. We know there are some limitations with this model, which will be fixed in v2.