Paper: NanoVDR: Distilling a 2B Vision-Language Retriever into a 70M Text-Only Encoder for Visual Document Retrieval | Blog | Project page

NanoVDR-Q-BERT-Qwen3VL2B-2048

⚠️ This model was renamed

It was published as nanovdr/NanoVDR-M. Old links and from_pretrained calls still resolve through a redirect, but please move to the new id.

- SentenceTransformer("nanovdr/NanoVDR-M")
+ SentenceTransformer("nanovdr/NanoVDR-Q-BERT-Qwen3VL2B-2048")

Why. NanoVDR started as one thing: a small text encoder that replaces the query side of a large vision-language retriever. It has since grown a second tower that replaces the document side, and a second teacher, so a name like -S no longer says enough. Two checkpoints only work together when they were distilled from the same teacher into the same width, and neither fact was recoverable from the old names.

The scheme is now NanoVDR-<Q|D>-<variant>-<teacher>-<width>[-ML], and the rule is simply that the teacher and the width have to match. This model is Q (query tower), BERT (backbone), distilled from Qwen3-VL-Embedding-2B into 2048 dimensions. -ML used to be spelled -Multi, which read as multi-vector when it meant multilingual.

BERT-base ablation variant. For production use, we recommend NanoVDR-Q-DistilBERT-Qwen3VL2B-2048-ML.

NanoVDR-Q-BERT-Qwen3VL2B-2048 is a 112M-parameter text-only query encoder for visual document retrieval, trained via asymmetric cross-modal distillation from Qwen3-VL-Embedding-2B. It uses BERT-base + a 2-layer MLP projector and achieves the highest v2 score among all NanoVDR variants.

Highlights

• Single-vector retrieval — queries and documents share the same 2048-dim embedding space as Qwen3-VL-Embedding-2B; retrieval is a plain dot product, FAISS-compatible, 4 KB per page (float16)
• Lightweight on storage — 454 MB model; doc index costs 64× less than ColPali's multi-vector patches
• Asymmetric setup — tiny 112M text encoder at query time; large VLM indexes documents offline once

Results

_{NDCG@5 (×100). Retention = Student / Teacher averaged across v1/v2/v3.}

Usage

Prerequisite: Documents must be indexed offline using Qwen3-VL-Embedding-2B (the teacher model). See the NanoVDR-Q-DistilBERT-Qwen3VL2B-2048-ML model page for a complete indexing guide.

from sentence_transformers import SentenceTransformer
import numpy as np

# doc_embeddings: (N, 2048) from teacher indexing (see prerequisite above)

model = SentenceTransformer("nanovdr/NanoVDR-Q-BERT-Qwen3VL2B-2048")
query_embeddings = model.encode(["What was the revenue growth in Q3?"])  # (1, 2048)

scores = query_embeddings @ doc_embeddings.T
top_k_indices = np.argsort(scores[0])[-5:][::-1]

Training Details

All NanoVDR Models

Want a system with no teacher at all?

Every model in this table targets Qwen3-VL-Embedding-2B at 2048 dimensions, so documents still have to be indexed by that 2B teacher. A second family targets the 8B teacher at 4096 dimensions and includes a document tower, so nothing multi-billion runs at indexing time either:

The two families are not interchangeable: 2048-d and 4096-d vectors live in different spaces and will not score against each other.

Citation

@article{nanovdr2026,
  title={NanoVDR: Distilling a 2B Vision-Language Retriever into a 70M Text-Only Encoder for Visual Document Retrieval},
  author={Liu, Zhuchenyang and Zhang, Yao and Xiao, Yu},
  journal={arXiv preprint arXiv:2603.12824},
  year={2026}
}

License

Apache 2.0