Visual Document Retrieval
ColPali
Safetensors
sentence-transformers
English
vidore
vidore-experimental
multi-vector
Instructions to use vidore/colpali-v1.3 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- ColPali
How to use vidore/colpali-v1.3 with ColPali:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- sentence-transformers
How to use vidore/colpali-v1.3 with sentence-transformers:
from sentence_transformers import SentenceTransformer model = SentenceTransformer("vidore/colpali-v1.3") sentences = [ "The weather is lovely today.", "It's so sunny outside!", "He drove to the stadium." ] embeddings = model.encode(sentences) similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] - Notebooks
- Google Colab
- Kaggle
how to use colpali for image retrieval?
#2
by susht - opened
i want to use query to search images, can i use coinse score? i find the shape of one text is (60, 128), and the shape of one image is (1030, 128), how to get cosine score?
You shouldn't use the cosine score: ColPali was built to work with the MaxSim operator ("late interaction") as it generates multi-vector embeddings. Use the ColPaliProcessor.score with your two embeddings to get your similarity score 👋🏼
tonywu71 changed discussion status to closed