Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -654,7 +654,10 @@ for doc in matched_docs:
|
|
| 654 |
|
| 655 |
### Example 7: Local Serving with Ollama
|
| 656 |
|
| 657 |
-
Run `code-embed-v2` locally with Ollama (native embed API or OpenAI-compatible endpoint)
|
|
|
|
|
|
|
|
|
|
| 658 |
|
| 659 |
```bash
|
| 660 |
# 1. Pull & run directly from Hugging Face
|
|
|
|
| 654 |
|
| 655 |
### Example 7: Local Serving with Ollama
|
| 656 |
|
| 657 |
+
Run `code-embed-v2` locally with Ollama (native embed API or OpenAI-compatible endpoint).
|
| 658 |
+
|
| 659 |
+
> [!NOTE]
|
| 660 |
+
> `code-embed-v2`'s native dimension is **1536-d** (active in SentenceTransformers, ONNX, FastEmbed, and TEI). Ollama/`llama.cpp` executes the 12-layer RoBERTa transformer graph up to layer-norm, outputting the **768-d Matryoshka representation**.
|
| 661 |
|
| 662 |
```bash
|
| 663 |
# 1. Pull & run directly from Hugging Face
|