Apple has published 19 collections on Hugging Face with 161 items in total, from FastVLM and MobileCLIP to Core ML packages for Stable Diffusion and Depth Anything. Each section below is one of Apple's collections, in the order of its upvotes.
Collections read from Hugging Face on 2026-11-02.
4 items · 129 upvotes · Updated 2025-08-25
8 items · 117 upvotes · Updated 2026-03-02
Efficient Vision Encoding for Vision Language Models
16 items · 84 upvotes · Updated 2026-03-02
A collection of AIMv2 vision encoders that supports a number of resolutions, native resolution, and a distilled checkpoint.
30 items · 64 upvotes · Updated 2026-04-23
MobileCLIP2: Mobile-friendly image-text models with SOTA zero-shot capabilities trained on DFNDR-2B
4 items · 55 upvotes · Updated 2025-08-25
5 items · 52 upvotes · Updated 2025-08-25
AIM: Autoregressive Image Models
7 items · 42 upvotes · Updated 2025-08-25
22 items · 40 upvotes · Updated 2026-03-02
MobileCLIP: Mobile-friendly image-text models with SOTA zero-shot capabilities. DataCompDR: Improved datasets for training image-text SOTA models.
8 items · 35 upvotes · Updated 2025-08-25
16 items · 31 upvotes · Updated 2025-08-25
6 items · 29 upvotes · Updated 2025-08-25
DCLM Models + Datasets
7 items · 21 upvotes · Updated 2025-08-25
CLIP Models trained using DFN-2B/DFN-5B datasets
2 items · 15 upvotes · Updated 2025-08-25
4 items · 14 upvotes · Updated 2025-08-25
Depth Pro: Sharp Monocular Metric Depth in Less Than a Second
2 items · 12 upvotes · Updated 2025-08-25
4 items · 11 upvotes · Updated 2025-08-25
8 items · 11 upvotes · Updated 2026-03-02
Benchmark for the design of efficient continual learning of image-text models over years.
3 items · 7 upvotes · Updated 2026-04-07
Simple Self-Distillation
5 items · 4 upvotes · Updated 2026-03-17
CLaRa models
These Spaces are static pages: the model runs in your own browser, on your device, and nothing is sent to a server to be processed. Most need a recent browser with WebGPU, and the first visit downloads the model, which can be a few hundred megabytes.
Every link goes to Hugging Face. We don't host these models or Spaces and we haven't tested them; the collections and the Spaces' type were read from Hugging Face's public API on the date above. Check each model's license and card before you use it in an app.