Total runs:
1.3M
Run Growth:
-258.5K
Growth Rate:
-19.12%
Sponsored by
Tripo AI
apple / MobileCLIP-S2
Model's Last Updated: March 01 2025
Introduction of MobileCLIP-S2
Model Details of MobileCLIP-S2
MobileCLIP: Fast Image-Text Models through Multi-Modal Reinforced Training
MobileCLIP was introduced in MobileCLIP: Fast Image-Text Models through Multi-Modal Reinforced Training (CVPR 2024), by Pavan Kumar Anasosalu Vasu, Hadi Pouransari, Fartash Faghri, Raviteja Vemulapalli, Oncel Tuzel.
This repository contains the MobileCLIP-S2 checkpoint.
Highlights
-
Our smallest variant
MobileCLIP-S0obtains similar zero-shot performance as OpenAI 's ViT-B/16 model while being 4.8x faster and 2.8x smaller. -
MobileCLIP-S2obtains better avg zero-shot performance than SigLIP 's ViT-B/16 model while being 2.3x faster and 2.1x smaller, and trained with 3x less seen samples. -
MobileCLIP-B(LT) attains zero-shot ImageNet performance of 77.2% which is significantly better than recent works like DFN and SigLIP with similar architectures or even OpenAI's ViT-L/14@336 .
Checkpoints
| Model |
# Seen
Samples (B) |
# Params (M)
(img + txt) |
Latency (ms)
(img + txt) |
IN-1k Zero-Shot
Top-1 Acc. (%) |
Avg. Perf. (%)
on 38 datasets |
|---|---|---|---|---|---|
| MobileCLIP-S0 | 13 | 11.4 + 42.4 | 1.5 + 1.6 | 67.8 | 58.1 |
| MobileCLIP-S1 | 13 | 21.5 + 63.4 | 2.5 + 3.3 | 72.6 | 61.3 |
| MobileCLIP-S2 | 13 | 35.7 + 63.4 | 3.6 + 3.3 | 74.4 | 63.7 |
| MobileCLIP-B | 13 | 86.3 + 63.4 | 10.4 + 3.3 | 76.8 | 65.2 |
| MobileCLIP-B (LT) | 36 | 86.3 + 63.4 | 10.4 + 3.3 | 77.2 | 65.8 |
How to Use
First, download the desired checkpoint visiting one of the links in the table above, then click the
Files and versions
tab, and download the PyTorch checkpoint.
For programmatic downloading, if you have
huggingface_hub
installed, you can also run:
huggingface-cli download pcuenq/MobileCLIP-S2
Then, install
ml-mobileclip
by following the instructions in the repo. It uses an API similar to
open_clip
's
.
You can run inference with a code snippet like the following:
import torch
from PIL import Image
import mobileclip
model, _, preprocess = mobileclip.create_model_and_transforms('mobileclip_s2', pretrained='/path/to/mobileclip_s2.pt')
tokenizer = mobileclip.get_tokenizer('mobileclip_s2')
image = preprocess(Image.open("docs/fig_accuracy_latency.png").convert('RGB')).unsqueeze(0)
text = tokenizer(["a diagram", "a dog", "a cat"])
with torch.no_grad(), torch.cuda.amp.autocast():
image_features = model.encode_image(image)
text_features = model.encode_text(text)
image_features /= image_features.norm(dim=-1, keepdim=True)
text_features /= text_features.norm(dim=-1, keepdim=True)
text_probs = (100.0 * image_features @ text_features.T).softmax(dim=-1)
print("Label probs:", text_probs)
Runs of apple MobileCLIP-S2 on huggingface.co
89
Total runs
-8
24-hour runs
-11
3-day runs
-6
7-day runs
-4
30-day runs
More Information About MobileCLIP-S2 huggingface.co Model
Url of MobileCLIP-S2
MobileCLIP-S2 huggingface.co Url
Provider of MobileCLIP-S2 huggingface.co
Other API from apple
Updated:March 01 2025
Total runs:
373.7K
Run Growth:
-57.6K
Growth Rate:
-15.39%
Total runs:
327.5K
Run Growth:
-986.1K
Growth Rate:
-287.93%
Total runs:
65.5K
Run Growth:
-24.1K
Growth Rate:
-36.85%
Total runs:
33.7K
Run Growth:
22.9K
Growth Rate:
67.66%
Total runs:
21.6K
Run Growth:
-156.7K
Growth Rate:
-755.93%
Total runs:
13.0K
Run Growth:
2.0K
Growth Rate:
15.43%
Total runs:
10.4K
Run Growth:
-43.5K
Growth Rate:
-411.24%
Total runs:
9.0K
Run Growth:
-11.5K
Growth Rate:
-131.97%
Total runs:
7.0K
Run Growth:
2.8K
Growth Rate:
40.42%
Total runs:
6.5K
Run Growth:
2.8K
Growth Rate:
42.73%
Total runs:
4.7K
Run Growth:
3.1K
Growth Rate:
65.54%
Total runs:
4.0K
Run Growth:
2.4K
Growth Rate:
57.90%
Total runs:
3.3K
Run Growth:
1.0K
Growth Rate:
30.35%
Total runs:
3.2K
Run Growth:
2.6K
Growth Rate:
79.98%
Total runs:
2.0K
Run Growth:
155
Growth Rate:
8.19%
Total runs:
1.7K
Run Growth:
951
Growth Rate:
56.11%
Total runs:
1.5K
Run Growth:
480
Growth Rate:
32.54%
Total runs:
1.4K
Run Growth:
-58
Growth Rate:
-4.06%
Total runs:
1.1K
Run Growth:
-2
Growth Rate:
-0.18%
Total runs:
1.1K
Run Growth:
66
Growth Rate:
6.19%
Total runs:
919
Run Growth:
-17.9K
Growth Rate:
-1929.34%
Total runs:
838
Run Growth:
61
Growth Rate:
7.38%
Total runs:
828
Run Growth:
-115
Growth Rate:
-13.25%
Total runs:
733
Run Growth:
-125
Growth Rate:
-17.12%
Total runs:
537
Run Growth:
213
Growth Rate:
39.08%
Total runs:
455
Run Growth:
44
Growth Rate:
9.59%
Total runs:
455
Run Growth:
118
Growth Rate:
25.43%
Total runs:
438
Run Growth:
234
Growth Rate:
53.42%
Total runs:
399
Run Growth:
170
Growth Rate:
42.82%
Total runs:
364
Run Growth:
283
Growth Rate:
88.71%
Total runs:
343
Run Growth:
158
Growth Rate:
47.31%
Total runs:
332
Run Growth:
128
Growth Rate:
42.11%
Total runs:
283
Run Growth:
198
Growth Rate:
80.49%
Total runs:
262
Run Growth:
111
Growth Rate:
41.73%
Total runs:
261
Run Growth:
138
Growth Rate:
54.98%
Total runs:
253
Run Growth:
100
Growth Rate:
39.22%
Total runs:
244
Run Growth:
184
Growth Rate:
84.79%
Total runs:
233
Run Growth:
97
Growth Rate:
40.42%
Total runs:
230
Run Growth:
151
Growth Rate:
67.11%
Total runs:
215
Run Growth:
33
Growth Rate:
14.93%
Total runs:
142
Run Growth:
48
Growth Rate:
34.78%
Total runs:
138
Run Growth:
66
Growth Rate:
47.83%
Total runs:
121
Run Growth:
-151
Growth Rate:
-115.27%
Total runs:
107
Run Growth:
-197
Growth Rate:
-189.42%
Total runs:
96
Run Growth:
47
Growth Rate:
48.96%
Total runs:
79
Run Growth:
-393
Growth Rate:
-497.47%
Total runs:
74
Run Growth:
-438
Growth Rate:
-584.00%
Total runs:
66
Run Growth:
25
Growth Rate:
39.06%
Total runs:
61
Run Growth:
1
Growth Rate:
1.64%
Total runs:
57
Run Growth:
24
Growth Rate:
41.38%
Total runs:
55
Run Growth:
22
Growth Rate:
42.31%
Total runs:
50
Run Growth:
25
Growth Rate:
59.52%
Total runs:
46
Run Growth:
4
Growth Rate:
8.70%
Total runs:
43
Run Growth:
37
Growth Rate:
86.05%