ColPali
ILLUIN TechnologyColPali applies the ColBERT late-interaction recipe to page screenshots: a vision-language model embeds each image patch of a rendered document page, a query is embedded token by token, and MaxSim scores the two multi-vector sets. That removes the OCR-and-chunk pipeline from document retrieval entirely. The line has grown past the original PaliGemma-based ColPali into ColQwen2, ColQwen2.5, the sub-1B ColSmol checkpoints and ColQwen-Omni, all released from the vidore Hub org with the shared colpali-engine training and inference code.
Published by ILLUIN Technology - code at github.com/illuin-tech/colpali, weights and the ViDoRe benchmark under the `vidore` Hugging Face org - not by a company called ViDoRe.
Openness
3 high confidence- weights
- open(45 ungated checkpoints on the vidore org, covering ColPali, ColQwen2/2.5, ColSmol and ColQwen-Omni)
- data
- open(vidore/colpali_train_set, the 118,195-pair training set, published ungated on the Hub)
- code
- open(colpali-engine[train] ships scripts/train/train_colbert.py and the per-model training configs used for the released checkpoints)
- license
- Gemma-License(non-OSI, resolved most-restrictive across the distributed SKUs - the PaliGemma-backed colpali-v1.2/v1.3-hf and -merged weights carry the Gemma terms)+MIT(OSI, the ColPali adapters and, with Apache-2.0 Qwen2-VL backbones, the ColQwen SKUs)
ColPali publishes its training set and training pipeline in full, which is as open as data and code get in this category. But some checkpoints, the PaliGemma-backed ColPali releases, merge in Gemma-licensed weights, and a family's license reading follows its most restrictive distributed checkpoint; the ColQwen line alone, released under MIT and Apache-2.0, would score higher.
- https://huggingface.co/api/datasets/vidore/colpali_train_set recorded 2026-09-11
`gated: false`, `private: false`, 118,195 train examples and a 500-row test split.
- https://huggingface.co/vidore/colpali-v1.3/raw/main/README.md recorded 2026-09-11
'ColPali's vision language backbone model (PaliGemma) is under `gemma` license ... The adapters attached to the model are under MIT license'; datasets: vidore/colpali_train_set; 127,460 query-page pairs described.
- https://huggingface.co/vidore/colqwen2-v1.0/raw/main/README.md recorded 2026-09-11
'ColQwen2's vision language backbone model (Qwen2-VL) is under `apache2.0` license. The adapters attached to the model are under MIT license'; `license: apache-2.0` front matter.
- https://raw.githubusercontent.com/illuin-tech/colpali/main/README.md recorded 2026-09-11
'This repository contains the code used for training and running visual document retrievers'; a Training section installing `colpali-engine[train]` and launching scripts/train/train_colbert.py against scripts/configs/pali/*.yaml; a licence column marking colpali-v1.x Gemma and the ColQwen/ColSmol SKUs Apache 2.0.
Adoption
3 high confidenceAdoption is measured as Hugging Face downloads summed across the ColPali family's retriever checkpoints on the vidore account, in every format published: ColPali, ColQwen2 and ColQwen2.5, ColSmol and ColQwen-Omni. The base initializations the adapters load are excluded so that one load is not counted twice, as are the paper's single-vector bi-encoder baselines and its ColIdefics ablation.
- https://huggingface.co/api/models?author=vidore&limit=100&sort=downloads&direction=-1 recorded 2026-10-17
45 vidore model repos with `downloads` (trailing 30 days) summing to 503,434; the 23 declared retriever checkpoints account for 503,356, led by colSmol-500M at 234,853 and colqwen2.5-v0.2 at 107,230. The 14 *-base initializations draw 65 and the four bi-encoder baselines 13.
Capability
4 high confidenceColPali's family, especially ColQwen2.5, sits near the top of the ViDoRe leaderboard it defined, a few points behind the current leader. It carries the same MaxSim late-interaction mechanism as ColBERT, extended from text retrieval into the document-image setting.
- https://arxiv.org/html/2510.01149v1 recorded 2026-09-11
Table 3, ViDoRe Leaderboard, nDCG@5: ColPali 2.92B 81.6 / 56.8 / 69.2; ColQwen2.5 3.75B 89.5 / 61.5 / 75.5; ColModernVBERT 0.25B 81.2 / 56.0 / 68.6; NemoRetriever-3B 91.0 / 66.3 / 78.7.
- https://raw.githubusercontent.com/illuin-tech/colpali/main/README.md recorded 2026-09-11
publisher's own ViDoRe column: colpali 81.3, colpali-v1.3 84.8, colqwen2-v1.0 89.3, colqwen2.5-v0.2 89.4, colSmol-500M 82.3.
Verified 2026-09-11