sea-clip-tiny-abl-wit

SEA-CLIP-Tiny — ablation: Only WIT

One row of the ablation table of SEA-CLIP-Tiny (ACCV 2026). Same architecture, pipeline and hyperparameters as the main model; the difference is the training mixture.

Usage

import open_clip

model, _, preprocess = open_clip.create_model_and_transforms('hf-hub:fassabilf/sea-clip-tiny-abl-wit')
tokenizer = open_clip.get_tokenizer('hf-hub:fassabilf/sea-clip-tiny-abl-wit')

Model

Results

Retrieval R@1 on the held-out splits of each training source, zero-shot ImageNet accuracy, and the paper's retrieval-only Avg@1 over XM3600, Flickr30k-200 and XTD-200 (%).

Reproduce

Training and evaluation code: https://github.com/fassabilf/sea-clip-tiny. The exact training configuration of this checkpoint is in params.txt in this repo.

Citation

@inproceedings{seacliptiny2026,
  title     = {SEA-CLIP-Tiny: Efficient Multilingual Text-Vision Embedding for Southeast Asian Languages},
  booktitle = {Asian Conference on Computer Vision (ACCV)},
  year      = {2026}
}