Add model card

README.md

@@ -0,0 +1,46 @@
+---
+language: en
+tags:
+  - embeddings
+  - sentence-transformers
+  - qwen3
+  - social-network
+  - ai-agents
+license: mit
+base_model: Qwen/Qwen3-Embedding-8B
+datasets:
+  - filter-with-espresso/moltbook-files
+---
+
+# Moltbook Embeddings
+
+Pre-computed text embeddings for the [moltbook-files](https://huggingface.co/datasets/filter-with-espresso/moltbook-files) dataset — a synthetic AI-agent social network with 232k posts across 3,628 communities.
+
+## Details
+
+| | |
+|---|---|
+| **Model** | [Qwen/Qwen3-Embedding-8B](https://huggingface.co/Qwen/Qwen3-Embedding-8B) |
+| **Vectors** | 219,252 |
+| **Precision** | float32 |
+| **Normalized** | Yes (L2) |
+| **Format** | NumPy `.npy` |
+| **Size** | ~3.6 GB |
+
+## Files
+
+- `embeddings.npy` — shape `(219252, D)`, one row per text
+- `embeddings_meta.json` — metadata (count + model name)
+
+## Usage
+
+```python
+import numpy as np
+
+embeddings = np.load("embeddings.npy")
+print(embeddings.shape)  # (219252, ...)
+```
+
+## How they were generated
+
+Texts were encoded with [sentence-transformers](https://sbert.net/) using `Qwen/Qwen3-Embedding-8B` in bfloat16, batch size 16, with L2 normalization, then stored as float32.