Update README.md

README.md

@@ -4,4 +4,138 @@ pipeline_tag: feature-extraction
 ---
 
 
-`pplx-embed-1` and `pplx-embed-1-kontext`, state-of-the-art text embedding models built for real-world, web-scale retrieval.
+<p align="center">
+  <img src="assets/logo.svg" alt="Perplexity Logo" width="400">
+</p>
+
+<p align="center">pplx-embed-1: Diffusion-LM for Dense and Contextual Retrieval</p>
+
+`pplx-embed-1` and `pplx-embed-1-context` are state-of-the-art text embedding models optimized for real-world, web-scale retrieval tasks.
+
+![diag.png](assets/diag.png)
+
+## Models
+
+| Model | Dimensions | Context | MRL | Quantization | Instruction |
+|:-----:|:----------:|:-------:|:---:|:------------:|:-----------:|
+| `pplx-embed-1-0.6B` | 1024 | 32K tokens | Yes | INT8/BINARY | No |
+| `pplx-embed-1-4B` | 2560 | 32K tokens | Yes | INT8/BINARY | No |
+| `pplx-embed-1-context-0.6B` | 1024 | 32K tokens | Yes | INT8/BINARY | No |
+| `pplx-embed-1-context-4B` | 2560 | 32K tokens | Yes | INT8/BINARY | No |
+
+<sub>All models are built on diffusion continued pre-trained Qwen3 at Perplexity AI.</sub>
+
+<sub>Many modern embedding models rely on instruction tuning, where users prepend an instruction string to the text being embedded. This can yield a 2%-3% lift on benchmarks, but it also introduces prompt-selection overhead and can make indexing pipelines brittle (small instruction changes can shift embedding space). We deliberately **avoid** this requirement: you can embed the text you want to index directly, without having to choose or maintain an instruction prefix.</sub>
+
+## Quick Start
+
+<details>
+<summary>Standard Embeddings API</summary>
+
+```bash
+curl -X POST https://api.perplexity.ai/v1/embeddings \
+  -H "Authorization: Bearer YOUR_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "texts": ["What drives human curiosity?"],
+    "model": "pplx-embed-1-4B"
+  }'
+```
+
+</details>
+
+<details>
+<summary>Contextualized Embeddings API</summary>
+
+```bash
+curl -X POST https://api.perplexity.ai/v1/contextualizedembeddings \
+  -H "Authorization: Bearer YOUR_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "inputs": [
+      [
+        "Curiosity begins in childhood with endless questions.",
+        "As we grow, curiosity drives us to explore new ideas."
+      ]
+    ],
+    "model": "pplx-embed-1-context-4B"
+  }'
+```
+
+</details>
+
+<details>
+<summary>Using Transformers</summary>
+
+```python
+from transformers import AutoModel
+
+# Standard embeddings
+model = AutoModel.from_pretrained(
+    "perplexityai/pplx-embed-1-4B",
+    trust_remote_code=True
+)
+
+texts = [
+    "What drives human curiosity?",
+    "Curiosity is the desire to learn and understand new things."
+]
+
+embeddings = model.encode(texts)  # Shape: (2, 2560)
+```
+
+</details>
+
+<details>
+<summary>Using SentenceTransformers</summary>
+
+```python
+from sentence_transformers import SentenceTransformer
+
+model = SentenceTransformer(
+    "perplexityai/pplx-embed-1-4B",
+    trust_remote_code=True
+)
+
+embeddings = model.encode(texts)
+```
+
+</details>
+
+<details>
+<summary>Contextualized Embeddings</summary>
+
+```python
+from transformers import AutoModel
+
+model = AutoModel.from_pretrained(
+    "perplexityai/pplx-embed-1-context-4B",
+    trust_remote_code=True
+)
+
+# Nested structure: each inner list is a document with sequential chunks
+doc_chunks = [
+    [
+        "Curiosity begins in childhood with endless questions.",
+        "As we grow, curiosity drives us to explore new ideas.",
+        "Scientific breakthroughs often start with a curious question."
+    ],
+    [
+        "The curiosity rover explores Mars searching for ancient life.",
+        "Each discovery sparks new questions about the universe."
+    ]
+]
+
+embeddings = model.encode(doc_chunks)  # Shape: (5, 2560)
+```
+
+</details>
+
+## Technical Details
+
+For comprehensive technical details and evaluation results, see our paper on arXiv.
+
+## Contact
+
+- Website: https://perplexity.ai
+- API Support: api-support@perplexity.ai