---
license: apache-2.0
base_model: EMD123/otzaria-qwen3-embedding
library_name: llama.cpp
pipeline_tag: feature-extraction
tags:
- gguf
- llama.cpp
- embeddings
- hebrew
- aramaic
- torah
- jewish-texts
- otzaria
- qwen3
---

# Otzaria Qwen3 Embedding GGUF

GGUF export of the merged Otzaria Qwen3 embedding model.

## Source merged model

`EMD123/otzaria-qwen3-embedding`

## llama.cpp version used

`b9933`

## Files

- `otzaria-qwen3-embedding-0.6b-merged-F16.gguf` — 1.12 GB
- `otzaria-qwen3-embedding-0.6b-merged-Q4_K_M.gguf` — 0.37 GB

## Recommendation

For embedding quality, start with:

- `F16` for maximum quality.
- `Q8_0` for smaller size with usually limited quality loss.

Use `Q4_K_M` only after checking retrieval metrics.

## Important limitation

This is a retrieval model. It retrieves relevant texts; it does not provide halachic rulings by itself.