File size: 3,311 Bytes
b2e54d9 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 | ---
base_model: Qwen/Qwen3-Reranker-8B
library_name: gguf
license: apache-2.0
pipeline_tag: text-ranking
tags:
- reranker
- gguf
- llama.cpp
- qwen3
- text-ranking
---
# Qwen3-Reranker-8B — GGUF (llama.cpp)
Working GGUF of [Qwen/Qwen3-Reranker-8B](https://huggingface.co/Qwen/Qwen3-Reranker-8B) for [llama.cpp](https://github.com/ggml-org/llama.cpp). Converted 2025-03-09 with the official `convert_hf_to_gguf.py`.
> **Other sizes:** [0.6B](https://huggingface.co/Voodisss/Qwen3-Reranker-0.6B-GGUF-llama_cpp) · [4B](https://huggingface.co/Voodisss/Qwen3-Reranker-4B-GGUF-llama_cpp) · [8B (this)](https://huggingface.co/Voodisss/Qwen3-Reranker-8B-GGUF-llama_cpp)
## Available files
| File | Quant | Size | Description |
| --------------------------- | ----- | -------- | -------------------------------------------------- |
| `Qwen3-Reranker-8B-F16.gguf` | F16 | 14.10 GB | Full precision, no quality loss |
| `Qwen3-Reranker-8B-Q8_0.gguf` | Q8_0 | 7.49 GB | 8-bit quantized, half the size |
## Does it work?
Yes. Most community GGUFs of Qwen3-Reranker produce garbage scores (`4.5e-23`) because they're missing reranker-specific tensors. See [llama.cpp #16407](https://github.com/ggml-org/llama.cpp/issues/16407). This one works:
```
Doc 0 (relevant): relevance_score = 0.99XX
Doc 1 (irrelevant): relevance_score = 0.00XX
```
## Quick start
```bash
llama-server -m Qwen3-Reranker-8B-f16.gguf --reranking --pooling rank --embedding --port 8081
```
```bash
curl http://localhost:8081/v1/rerank \
-H "Content-Type: application/json" \
-d '{
"query": "employment termination notice period",
"documents": [
"The Labour Code requires 30 calendar days written notice.",
"Corporate tax rates for small enterprises."
]
}'
```
Use **`/v1/rerank`**, not `/v1/embeddings`. The embeddings endpoint returns zeros for reranker models.
## What's different about this GGUF?
The official `convert_hf_to_gguf.py` detects Qwen3-Reranker and does things naive converters skip:
- Extracts `cls.output.weight` (the yes/no classifier) from `lm_head`
- Sets `pooling_type = RANK` metadata
- Bakes in the rerank chat template
- Sets `classifier.output_labels = ["yes", "no"]`
Without these, llama-server has nothing to compute scores from.
## Known broken GGUFs
- [DevQuasar/Qwen.Qwen3-Reranker-4B-GGUF](https://huggingface.co/DevQuasar/Qwen.Qwen3-Reranker-4B-GGUF) — broken.
## models.ini example
```ini
[Qwen3-Reranker-8B-f16]
model = /path/to/Qwen3-Reranker-8B-f16.gguf
reranking = true
pooling = rank
embedding = true
ctx-size = 32768
```
For a full multi-model setup guide (embedding + reranking + chat on one server), see the **[llama-server Qwen3 guide](https://gist.github.com/VooDisss/42bce4eb5c76d3c325633886c5e348ee)**.
## Convert it yourself
```bash
pip install huggingface_hub gguf torch safetensors sentencepiece
python -c "from huggingface_hub import snapshot_download; snapshot_download('Qwen/Qwen3-Reranker-8B', local_dir='Qwen3-Reranker-8B-src')"
python convert_hf_to_gguf.py --outtype f16 --outfile Qwen3-Reranker-8B-f16.gguf Qwen3-Reranker-8B-src/
```
## License
Apache 2.0 — same as the [original model](https://huggingface.co/Qwen/Qwen3-Reranker-8B). |