Downloads · 30 days
0
deepghs/siglip_onnx
siglip_onnx is a zero-shot classification model from deepghs. Use it when you need labels you did not train the model on. It is set up for dghs-realutils.
Downloads · 30 days
0
Access
Public
Updated Jan 29, 2025
Repo size
13.5 GB
Likes
3
Public
Click a slice to open those files.
.onnx13.5 GB · 100%
From the Hugging Face model README
ONNX exported version of SigLIP models.
8 models exported in total.
| Name | Image (Params/FLOPS) | Image Size | Image Width (Enc/Emb) | Text (Params/FLOPS) | Text Width (Enc/Emb) | Created At |
|---|---|---|---|---|---|---|
| google/siglip-so400m-patch14-384 | 422.1M / 300.4G | 384 | 1152 / 1152 | 412.8M / 52.7G | 1152 / 1152 | 2024-01-08 |
| google/siglip-base-patch16-256-multilingual | 90.4M / 21.9G | 256 | 768 / 768 | 85.6M / 10.9G | 768 / 768 | 2024-01-08 |
| google/siglip-large-patch16-384 | 311.5M / 174.5G | 384 | 1024 / 1024 | 303.4M / 38.7G | 1024 / 1024 | 2024-01-08 |
| google/siglip-large-patch16-256 | 311.5M / 77.6G | 256 | 1024 / 1024 | 303.4M / 38.7G | 1024 / 1024 | 2024-01-08 |
| google/siglip-base-patch16-512 | 90.4M / 87.7G | 512 | 768 / 768 | 85.6M / 10.9G | 768 / 768 | 2024-01-08 |
| google/siglip-base-patch16-384 | 90.4M / 49.3G | 384 | 768 / 768 | 85.6M / 10.9G | 768 / 768 | 2024-01-08 |
| google/siglip-base-patch16-256 | 90.4M / 21.9G | 256 | 768 / 768 | 85.6M / 10.9G | 768 / 768 | 2024-01-08 |
| google/siglip-base-patch16-224 | 90.4M / 16.8G | 224 | 768 / 768 | 85.6M / 10.9G | 768 / 768 | 2023-10-01 |