README.md

---
license: apache-2.0
language: [en]
tags: [embeddings, gguf, ggml, text-embeddings, xlm-r, crispembed, ollama]
pipeline_tag: feature-extraction
base_model: Snowflake/snowflake-arctic-embed-l-v2.0
---

# arctic-embed-l-v2 GGUF

GGUF format of [Snowflake/snowflake-arctic-embed-l-v2.0](https://huggingface.co/Snowflake/snowflake-arctic-embed-l-v2.0) for use with [CrispEmbed](https://github.com/CrispStrobe/CrispEmbed) and [Ollama](https://ollama.com).

## Files

| File | Quantization | Size |
|------|-------------|------|
| [arctic-embed-l-v2-q4_k.gguf](https://huggingface.co/cstr/arctic-embed-l-v2-GGUF/resolve/main/arctic-embed-l-v2-q4_k.gguf) | Q4_K | 0 MB |
| [arctic-embed-l-v2-q8_0.gguf](https://huggingface.co/cstr/arctic-embed-l-v2-GGUF/resolve/main/arctic-embed-l-v2-q8_0.gguf) | Q8_0 | 0 MB |
| [arctic-embed-l-v2.gguf](https://huggingface.co/cstr/arctic-embed-l-v2-GGUF/resolve/main/arctic-embed-l-v2.gguf) | F32 | 0 MB |

**Recommended:** Q8_0 for quality (cos vs HF: L2=1.0), Q4_K for size (L2=1.0).

## Quick Start

### CrispEmbed
```bash
./crispembed -m arctic-embed-l-v2 "Hello world"
./crispembed-server -m arctic-embed-l-v2 --port 8080
```

### Ollama (with [CrispStrobe fork](https://github.com/CrispStrobe/ollama/tree/feat/xlmr-embedding))
```bash
echo "FROM arctic-embed-l-v2-q8_0.gguf" > Modelfile
ollama create arctic-embed-l-v2 -f Modelfile
curl http://localhost:11434/api/embed -d '{"model":"arctic-embed-l-v2","input":["Hello world"]}'
```

### Python (CrispEmbed)
```python
from crispembed import CrispEmbed
model = CrispEmbed("arctic-embed-l-v2-q8_0.gguf")
vectors = model.encode(["Hello world", "Goodbye world"])
```

## Model Details

| Property | Value |
|----------|-------|
| Architecture | XLM-R |
| Parameters | 560M |
| Embedding Dimension | 1024 |
| Layers | 24 |
| Pooling | CLS |
| Tokenizer | SentencePiece |
| Language | en |
| Q8_0 vs HuggingFace | L2=1.0 |
| Q4_K vs HuggingFace | L2=1.0 |

## Server API

CrispEmbed server supports four API dialects:
- `POST /embed` -- native
- `POST /v1/embeddings` -- OpenAI-compatible
- `POST /api/embed` -- Ollama-compatible
- `POST /api/embeddings` -- Ollama legacy

## Credits

- Original model: [Snowflake/snowflake-arctic-embed-l-v2.0](https://huggingface.co/Snowflake/snowflake-arctic-embed-l-v2.0)
- Inference: [CrispEmbed](https://github.com/CrispStrobe/CrispEmbed) (MIT, ggml-based)

## Provenance and EU AI Act Art. 53 note

- **Upstream model:** [Snowflake/snowflake-arctic-embed-l-v2.0](https://huggingface.co/Snowflake/snowflake-arctic-embed-l-v2.0) — published by `Snowflake`.
- **Upstream licence:** `apache-2.0`. This repository redistributes under the same terms; it grants no rights the upstream licence does not.
- **What was done here:** format conversion and/or quantisation only (GGUF/GGML). No training, no fine-tuning, no merging, no distillation, no change to architecture, vocabulary or capability. Only the numeric representation of the upstream weights differs.
- **Training data:** documented — where it is documented at all — by the upstream provider; see the upstream model card. No training data was used, added or selected by this repository. No training-content summary was found on the upstream model card at the time of writing; that documentation gap is upstream's and is not filled here.
- **Provider status:** under Regulation (EU) 2024/1689 the upstream authors remain the provider of this model. Converting the serialisation format does not make this repository the provider of a new general-purpose AI model, and no such claim is made. Questions about training content, copyright policy or model capability belong upstream.