---
license: apache-2.0
language:
- en
- fr
base_model:
- Qwen/Qwen3.5-0.8B
pipeline_tag: image-text-to-text
library_name: transformers
tags:
- reranker
- cross-encoder
- multimodal
- text-ranking
- document-reranking
- vidore
- beir
- sentence-transformers
datasets:
- vidore/colpali_train_set
- lightonai/embeddings-fine-tuning
---

[](https://lighton.ai)
[](https://www.linkedin.com/company/lighton/)
[](https://x.com/LightOnIO)
📝 [Blog post](https://huggingface.co/blog/lightonai/lighton-rerank)
LightOn-rerank-PW-0.8B
Unified Text + Visual Document Reranker by LightOn
PW-0.8B |
LW-0.8B |
PW-2B |
LW-2B |
PW-4B |
LW-4B
---
## About the LightOn-rerank family
Production retrieval pipelines usually need two rerankers: one for text passages and one for visual documents (PDF pages, slides, scans). **LightOn-rerank** models are *unified* cross-encoder rerankers: a single model scores both text passages and document page images against a query, on top of any first-stage retriever (BM25, dense embeddings, or ColPali-family late-interaction models).
The models are built on Qwen3.5 backbone (hybrid linear + full attention) and jointly fine-tuned on text and visual reranking data with mixed-modality batches (LoRA, merged into the released weights). Training data is English-only; French performance transfers zero-shot from the multilingual backbone.
The family comes in two scoring flavours × three sizes (0.8B / 2B / 4B):
- **PW (pointwise)**: each candidate is scored independently. The model judges whether the document answers the query, and the score is `logit("Yes") − logit("No")`. One forward pass per candidate and no generation.
- **LW (listwise)**: generative listwise ranking, where 4 candidates are placed in a single prompt and the model generates a permutation (`[2] > [4] > [1] > [3]`). Larger candidate pools are ranked with a sliding window (window 4, stride 2, bottom-to-top). Cross-document attention makes LW markedly stronger on hard visual reranking, and unlike pointwise scoring it keeps improving with backbone size.
**LightOn-rerank-PW-0.8B** is the smallest pointwise member of the family, released primarily as a size/recipe reference point: it shows that independent Yes/No scoring degrades sharply below 2B parameters. If you want a 0.8B reranker, use [LightOn-rerank-LW-0.8B](https://huggingface.co/lightonai/LightOn-rerank-LW-0.8B) instead (+10.0 nDCG@10 on ViDoRe V3 with the same backbone).
## Results
**ViDoRe V3** (visual document reranking, 8 domains × EN/FR queries), overall nDCG@10, ColQwen2.5-v0.2 first stage, retrieve 100 / rerank 100. All models, including baselines, were re-evaluated under this same two-stage protocol, so numbers are mutually comparable but not comparable to vendor-reported end-to-end results.
| Model | Params | Scoring | ViDoRe V3 overall nDCG@10 |
|---|---|---|---|
| [LightOn-rerank-LW-4B](https://huggingface.co/lightonai/LightOn-rerank-LW-4B) | 4.5B | listwise | 64.69 |
| *Qwen3-VL-Reranker-8B* | *8B* | *pointwise (pooling)* | *64.23* |
| [LightOn-rerank-LW-2B](https://huggingface.co/lightonai/LightOn-rerank-LW-2B) | 2.2B | listwise | 62.66 |
| [LightOn-rerank-PW-2B](https://huggingface.co/lightonai/LightOn-rerank-PW-2B) | 2.2B | pointwise | 59.87 |
| [LightOn-rerank-PW-4B](https://huggingface.co/lightonai/LightOn-rerank-PW-4B) | 4.5B | pointwise | 59.80 |
| *jina-reranker-m0* | *2.4B* | *pointwise* | *59.40* |
| *Qwen3-VL-Reranker-2B* | *2B* | *pointwise (pooling)* | *59.18* |
| [LightOn-rerank-LW-0.8B](https://huggingface.co/lightonai/LightOn-rerank-LW-0.8B) | 0.85B | listwise | 58.25 |
| *MonoQwen2-VL-v0.1* | *2B* | *pointwise* | *57.76* |
| *First-stage only (ColQwen2.5, no rerank)* | — | — | *55.60* |
| **LightOn-rerank-PW-0.8B (this model)** | 0.85B | pointwise | **48.20** |
Note that this model lands **below the first-stage retriever**: at 0.8B, independent pointwise scoring makes the ranking worse than not reranking at all on several splits (see limitations).
### ViDoRe V3 detail (nDCG@10, ColQwen2.5 first stage, rerank-100)
| Domain | EN | FR |
|---|---|---|
| finance_en | 57.40 | 31.08 |
| finance_fr | 32.28 | 30.27 |
| computer_science | 73.96 | 50.86 |
| hr | 56.41 | 26.12 |
| energy | 64.90 | 63.97 |
| industrial | 54.88 | 25.99 |
| pharmaceuticals | 64.59 | 46.40 |
| physics | 47.84 | 44.22 |
| **mean** | **56.53** | **39.86** |
Overall nDCG@10: **48.20** (EN 56.53 / FR 39.86).
### BEIR results (text reranking)
nDCG@10, BM25 first stage, retrieve 100 / rerank 100. ⚠️ marks datasets in the text training mix (NQ, MSMARCO); the decontaminated mean excludes them.
| Dataset | nDCG@10 |
|---|---|
| fever | 79.92 |
| scifact | 75.23 |
| trec-covid | 82.11 |
| hotpotqa | 67.45 |
| nq ⚠️ | 54.58 |
| dbpedia | 39.87 |
| arguana | 38.69 |
| fiqa | 34.75 |
| msmarco ⚠️ | 36.07 |
| nfcorpus | 34.38 |
| touche-2020 | 37.04 |
| climate-fever | 22.22 |
| scidocs | 16.10 |
| **Mean (13)** | **47.57** |
| **Decontaminated mean (11, excl. ⚠️)** | **47.98** |
On text, pointwise holds up at 0.8B (47.98 decontaminated mean, slightly ahead of LW-0.8B at 47.30): the collapse described above is specific to visual reranking, where independent scoring of hard candidates fails at this size.
## Model Details
- **Model type:** multimodal cross-encoder reranker (**pointwise**: each candidate is scored independently as `logit("Yes") − logit("No")`)
- **Base model:** [Qwen/Qwen3.5-0.8B](https://huggingface.co/Qwen/Qwen3.5-0.8B) (Qwen3.5 hybrid linear + full attention VLM)
- **Parameters:** ≈0.85B (bfloat16, 1.7 GB)
- **Inputs:** query (text) + candidate document(s): text passage **or** page image
- **Fine-tuning:** joint text+vision LoRA (r=32, α=32, rsLoRA, merged into the released weights), mixed-modality batches (2 text + 2 vision groups per micro-batch), vision loss weight 1.3, lr 1e-4, 1 epoch (465 steps)
- **Data:** same 213k groups as the listwise models — 107k text groups (NQ, TriviaQA, MS MARCO; each a `[pos, neg_0, neg_1, neg_2]` 4-list with hard negatives mined via the NV-Retriever approach with GTE-ModernBERT) + 106k vision groups (ColPali train set with negatives mined by Nomic). For pointwise training each group is flattened into (query, document, Yes/No) triples.
- **Languages:** English (training), French (zero-shot transfer)
- **Requirements:** `transformers >= 5.4.0` (`qwen3_5` architecture); `sentence-transformers >= 5.4.0` for the CrossEncoder usage
## Usage: pointwise reranking
Each candidate is scored independently as `logit("Yes") − logit("No")` at the first generated position, so you can sort candidates by descending score. The model was trained with a fixed system prompt and user template: use them verbatim for best results.
### Using Sentence Transformers
Install Sentence Transformers (`>=5.4.0`):
```bash
pip install "sentence-transformers[image]"
```
The trained system prompt and user templates are baked into the bundled `reranker` chat template, so query-document pairs are formatted correctly out of the box:
```python
from sentence_transformers import CrossEncoder
model = CrossEncoder("lightonai/LightOn-rerank-PW-0.8B")
query = "What is late interaction in neural information retrieval?"
documents = [
"ColBERT computes token-level query-document interactions at search time...",
"The Eiffel Tower is located on the Champ de Mars in Paris.",
]
pairs = [(query, doc) for doc in documents]
scores = model.predict(pairs)
print(scores)
# [-4.125 -7.1875]
rankings = model.rank(query, documents)
print(rankings)
# [{'corpus_id': 0, 'score': -4.125}, {'corpus_id': 1, 'score': -7.1875}]
```
To rerank **page images**, pass a `PIL.Image` (or an image URL or file path string) as the document. Text and image candidates can be mixed in the same call:
```python
from PIL import Image
pairs = [
(query, Image.open("page_1.png")),
(query, "https://example.com/page_2.png"),
(query, "A text passage candidate for the same query."),
]
scores = model.predict(pairs)
```
Scores are raw `logit("Yes") − logit("No")` differences. You can map them to 0...1 probabilities with `model.predict(pairs, activation_fn=torch.nn.Sigmoid())`.
### Using Transformers
```python
import torch
from transformers import AutoModelForImageTextToText, AutoProcessor
model_id = "lightonai/LightOn-rerank-PW-0.8B"
model = AutoModelForImageTextToText.from_pretrained(
model_id,
dtype=torch.bfloat16,
attn_implementation="flash_attention_2", # optional, remove if flash-attn is not installed
device_map="cuda",
).eval()
processor = AutoProcessor.from_pretrained(model_id)
processor.tokenizer.padding_side = "left" # scores are read at the last position
YES_TOKEN_ID = 9175 # "Yes"
NO_TOKEN_ID = 2665 # "No"
SYSTEM_PROMPT = "Judge whether the document is relevant to the query. Answer Yes or No."
USER_TEMPLATE = (
"Given a query, determine if the document is relevant. "
"The query is: {query}\n\nDocument: {doc}"
)
query = "What is late interaction in neural information retrieval?"
documents = [
"ColBERT computes token-level query-document interactions at search time...",
"The Eiffel Tower is located on the Champ de Mars in Paris.",
]
texts = [
processor.apply_chat_template(
[
{"role": "system", "content": SYSTEM_PROMPT},
{"role": "user", "content": USER_TEMPLATE.format(query=query, doc=doc)},
],
tokenize=False,
add_generation_prompt=True,
)
for doc in documents
]
inputs = processor(
text=texts, return_tensors="pt", padding=True, truncation=True, max_length=2048
).to(model.device)
with torch.inference_mode():
logits = model(**inputs).logits[:, -1]
scores = (logits[:, YES_TOKEN_ID] - logits[:, NO_TOKEN_ID]).tolist()
# scores: [-4.125, -7.1875]
ranked = sorted(zip(scores, documents), reverse=True)
```
To score a **page image** instead of a text passage, replace the user message with:
```python
VISION_TEMPLATE = "Given a query, determine if the document image is relevant. The query is: {query}"
{"role": "user", "content": [
{"type": "image", "image": page_image}, # PIL.Image
{"type": "text", "text": VISION_TEMPLATE.format(query=query)},
]}
```
and pass `images=[page_image, ...]` to the processor call (keep the same system prompt).
### Serving with vLLM
```bash
vllm serve lightonai/LightOn-rerank-PW-0.8B --trust-remote-code --max-model-len 16384
```
```python
resp = client.chat.completions.create(
model="lightonai/LightOn-rerank-PW-0.8B",
messages=messages, # same system + user messages as above
max_tokens=1,
logprobs=True,
top_logprobs=20,
temperature=0.0,
extra_body={"chat_template_kwargs": {"enable_thinking": False}},
)
top = resp.choices[0].logprobs.content[0].top_logprobs
lp = {t.token: t.logprob for t in top}
score = lp.get("Yes", -100.0) - lp.get("No", -100.0)
```
Full-page document images can exceed 8k tokens, so keep `--max-model-len` at 16384 or higher when reranking page images.
## Notes & limitations
- At 0.8B, pointwise Yes/No scoring cannot discriminate hard visual candidates: −10.0 nDCG@10 vs the listwise sibling on the same backbone, with French hit hardest. Prefer [LightOn-rerank-LW-0.8B](https://huggingface.co/lightonai/LightOn-rerank-LW-0.8B) at this size.
- On ViDoRe V3 this model lands below the first-stage retriever (48.20 vs 55.60 for ColQwen2.5 with no reranking): on several splits you would get better rankings, faster, by not reranking at all. It is released as a cautionary reference point, not a recommended operating point.
- Training data is English-only. French works zero-shot (the backbone is multilingual) but is slightly behind English on average.
- BEIR contamination flag: NQ and MSMARCO are part of the text training data; headline text figures use decontaminated means that exclude them.
## The LightOn-rerank family
| Model | Backbone | Scoring | ViDoRe V3 overall nDCG@10 |
|---|---|---|---|
| [LightOn-rerank-PW-0.8B](https://huggingface.co/lightonai/LightOn-rerank-PW-0.8B) | Qwen3.5-0.8B | pointwise | 48.20 |
| [LightOn-rerank-LW-0.8B](https://huggingface.co/lightonai/LightOn-rerank-LW-0.8B) | Qwen3.5-0.8B | listwise | 58.25 |
| [LightOn-rerank-PW-2B](https://huggingface.co/lightonai/LightOn-rerank-PW-2B) | Qwen3.5-2B | pointwise | 59.87 |
| [LightOn-rerank-LW-2B](https://huggingface.co/lightonai/LightOn-rerank-LW-2B) | Qwen3.5-2B | listwise | 62.66 |
| [LightOn-rerank-PW-4B](https://huggingface.co/lightonai/LightOn-rerank-PW-4B) | Qwen3.5-4B | pointwise | 59.80 |
| [LightOn-rerank-LW-4B](https://huggingface.co/lightonai/LightOn-rerank-LW-4B) | Qwen3.5-4B | listwise | **64.69** |
Rule of thumb: **LW** models are stronger at every size (and the gap grows with size); **PW** models are cheaper to serve and score candidates independently. For the best quality pick [LW-4B](https://huggingface.co/lightonai/LightOn-rerank-LW-4B); for the best quality/cost trade-off pick [LW-2B](https://huggingface.co/lightonai/LightOn-rerank-LW-2B); for maximum throughput on text-heavy workloads pick a PW model.
## Citation
```
@misc{ananya2026lightonrerank,
title={One Adapter, Both Modalities: Field Notes from Building and Serving a Multimodal Reranker},
author={Ananya, Ishrat Jahan and Chatelain, Amelie},
year={2026},
howpublished={\url{https://huggingface.co/blog/lightonai/lighton-rerank}},
}
```