apple/MobileCLIP2-L-14

Updated Oct 9, 2025 • 34 • 4

Note Timm ViT-L/14 architecture trained on DFNDR-2B (dataset of MobileCLIP2)

apple/MobileCLIP-S3

Updated 3 days ago • 51 • 4

Note New architecture introduced in MobileCLIP2 paper but pretrained on DataCompDR (dataset of MobileCLIP v1)

apple/MobileCLIP-S4

Updated 3 days ago • 55 • 7

Note New architecture introduced in MobileCLIP2 paper but pretrained on DataCompDR (dataset of MobileCLIP v1)

apple/MobileCLIP-L-14

Updated Oct 9, 2025 • 25 • 1

Note Timm ViT-L/14 architecture pretrained on DataCompDR (dataset of MobileCLIP v1)

timm/MobileCLIP2-S0-OpenCLIP

Zero-Shot Image Classification • Updated Sep 11, 2025 • 3.5k • 1

Note 👇Timm checkpoints

apple/mobileclip2_coca_dfn2b_s13b_mscoco38k_s12m_context77

Updated Oct 9, 2025 • 18 • 1

Note 👇MobileCLIP2 CoCa models for synthetic caption generation used to train MobileCLIP2 models

apple/mobileclip2_coca_dfn2b_s13b_docci_s12m_context256

Updated Oct 9, 2025 • 16 • 1

Note 👇MobileCLIP2 CoCa models (context length=256). Higher chance of generating repeated output.

apple/mobileclip2_coca_dfn2b_s13b_context77

Updated Oct 9, 2025 • 21 • 1

Note MobileCLIP2 CoCa base model. It can be used for fine-tuning new CoCa models on high quality datasets.