Instructions to use carsonpoole/binary-siglip-vision with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use carsonpoole/binary-siglip-vision with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("image-feature-extraction", model="carsonpoole/binary-siglip-vision")# Load model directly from transformers import AutoTokenizer, AutoModel tokenizer = AutoTokenizer.from_pretrained("carsonpoole/binary-siglip-vision") model = AutoModel.from_pretrained("carsonpoole/binary-siglip-vision", device_map="auto") - Notebooks
- Google Colab
- Kaggle
File size: 476 Bytes
e49a81c ddcbab5 e49a81c ddcbab5 e49a81c ddcbab5 e49a81c ddcbab5 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 | ---
library_name: transformers
license: apache-2.0
language:
- en
---
Description
This is a fine tuned google/siglip-so400m-patch14-384 for the purpose of quantizing the embeddings to binary. It's only using the first 1024 embeddings, so if you use all 1152 of them your results will be less than desirable.
I updated the model today (April 30th) and evals are much better than before, but I'm continuing training so perf should only get better from here.
Evals
Coming soon |