CLIMP Mamba2

--- license: apache-2.0 tags: - clip - mamba - mamba2 - vision-language - contrastive-learning - vmamba datasets: - conceptual_12m pipeline_tag: zero-shot-image-classification --- # CLIMP-Mamba2 **C**ontrastive **L**anguage-**I**mage **M**amba **P**retraining (CLIMP) using Mamba2-1.3B as text encoder. ## Model Description | Component | Details | |-----------|---------| | Vision Encoder | VMamba-Base (128-256-512-1024 dims, depths [2,2,15,2]) | | Text Encoder | Mamba2-1.3B (AntonV/mamba2-1.3b-hf) | | Projection Dim | 768 | | Training Data | CC12M | | Image Resolution | 224x224 | | Loss | Symmetric InfoNCE (learned temperature) | ## Usage ```python from models import load_climp from data.utils import transform_image model = load_climp("mamba2") transform = transform_image(224) ``` See the [demo repository](https://github.com/NimrodShabtay/CLIMP) for evaluation code. ## Paper [CLIMP: Contrastive Language-Image Mamba Pretraining](https://arxiv.org/abs/2601.06891) ```bibtex @article{climp2026, title={CLIMP: Contrastive Language-Image Mamba Pretraining}, author={Shabtay, Nimrod and Zimerman, Itamar and Schwartz, Eli and Giryes, Raja}, journal={arXiv preprint arXiv:2601.06891}, year={2026} } ```

Source context: 6 downloads · 0 likes · Pipeline zero-shot-image-classification · Repo NimrodShabtay1986/CLIMP-Mamba2