MobileCLIP: Fast Image-Text Models through Multi-Modal Reinforced Training
Paper
• 2311.17049 • Published
• 5
MobileCLIP Models + DataCompDR Data
MobileCLIP: Mobile-friendly image-text models with SOTA zero-shot capabilities. DataCompDR: Improved datasets for training image-text SOTA models.
-
-
apple/mobileclip_s0_timm
Image Classification • Updated • 83 • 12 -
apple/mobileclip_s1_timm
Image Classification • Updated • 41 • 3 -
apple/mobileclip_s2_timm
Image Classification • Updated • 121 • 6 -
apple/mobileclip_b_timm
Image Classification • Updated • 310 • 2
apple/mobileclip_b_lt_timm
Image Classification • Updated • 28 • 5Note ^ MobileCLIP checkpoints for the timm library (image-tower only)
-
apple/MobileCLIP-S1-OpenCLIP
Zero-Shot Image Classification • Updated • 2.44k • 10 -
apple/MobileCLIP-S2-OpenCLIP
Zero-Shot Image Classification • Updated • 59.3k • 18 -
apple/MobileCLIP-B-OpenCLIP
Zero-Shot Image Classification • Updated • 1.98k • 3
apple/MobileCLIP-B-LT-OpenCLIP
Zero-Shot Image Classification • Updated • 337 • 10Note ^ MobileCLIP checkpoints for the OpenCLIP library
-
apple/MobileCLIP-S0
Updated • 260 • 12 -
apple/MobileCLIP-S1
Updated • 54 • 11 -
apple/MobileCLIP-S2
Updated • 63 • 11 -
apple/MobileCLIP-B
Updated • 60 • 4
apple/MobileCLIP-B-LT
Updated • 10 • 10Note ^ MobileCLIP checkpoints, original format
-
apple/DataComp-12M
Viewer • Updated • 12.8M • 61 • 4 -
apple/DataCompDR-12M
Viewer • Updated • 12.8M • 2.09k • 34 -
apple/DataCompDR-12M-bf16
Updated • 389 • 4
apple/DataCompDR-1B
Viewer • Updated • 1.28B • 4.76k • 30Note ^ DataCompDR datasets
-
apple/MobileCLIP-S3
Updated • 14 • 1 -
apple/MobileCLIP-S4
Updated • 39 • 4
apple/MobileCLIP-L-14
Updated • 10Note ^ MobileCLIP2 checkpoints trained on DataCompDR