Skip to content

๐Ÿ“ฐ All Posts

  • 2 min read

version 0.5

We are thrilled to share that EmbedAnything version 0.5 is out now and comprise of insane development like support for ModernBert and ReRanker models. Along with Ingestion pipeline support for DocX, and HTML letโ€™s get in details.

The best of all have been support for late-interaction model, both ColPali and ColBERT on onnx.

  1. ModernBert Support: Well it made quite a splash, and we were obliged to add it, in the fastest inference engine, embedanything. In addition to being faster and more accurate, ModernBERT also increases context length to 8k tokens (compared to just 512 for most encoders), and is the first encoder-only model that includes a large amount of code in its training data.
  2. ColPali- Onnx : ย Running the ColPali model directly on a local machine might not always be feasible. To address this, we developed aย quantized version of ColPali. Find it on our hugging face, link here. You could also run it both on Candle and on ONNX.
  3. ColBERT: ColBERT is aย fastย andย accurateย retrieval model, enabling scalable BERT-based search over large text collections in tens of milliseconds.
  4. ReRankers: EmbedAnything recently contributed for the support of reranking models to Candle so as to add it in our own library. It can support any kind of reranking models. Precision meets performance! Use reranking models to refine your retrieval results for even greater accuracy.
  5. Jina V3: Also contributed to V3 models, for Jina can seamlessly integrate any V3 model.

  6. ๐——๐—ข๐—–๐—ซ ๐—ฃ๐—ฟ๐—ผ๐—ฐ๐—ฒ๐˜€๐˜€๐—ถ๐—ป๐—ด

    Effortlessly extract text from .docx files and convert it into embeddings. Simplify your document workflows like never before!

  7. ๐—›๐—ง๐— ๐—Ÿ ๐—ฃ๐—ฟ๐—ผ๐—ฐ๐—ฒ๐˜€๐˜€๐—ถ๐—ป๐—ด:

Parsing and embedding HTML documents just got easier!

โœ… Extract rich metadata with embeddings
โœ… Handle code blocks separately for better context

Supercharge your documentation retrieval with these advanced capabilities.

  • 8 min read

Optimize VLM Tokens with EmbedAnything x ColPali

ColPali, a late-interaction vision model, leverages this power to enable text searches within images. This means you can pinpoint the exact pages in a PDF containing relevant text, even if the text exists only as part of an image. For example, suppose you have hundreds of pages in a PDF and even hundreds of PDFs. In that case, ColPali can identify the specific pages matching a queryโ€”an impressive feat for streamlining information retrieval. This system is widely come to be known as Vision RAG.

  • 4 min read

The path ahead of EmbedAnything

In March, we set out to build a local file search app. We aimed to create a tool that would make file searching faster, more innovative, and more efficient. However, we quickly hit a roadblock: no high-performance backend fit our needs.

  • 7 min read

About Embed Anything

EmbedAnything is an open-source Rust/Python framework that lets you generate vector embeddings for any data (text, images, audio) with minimal code. It's blazing fast, memory-efficient, and can handle massive datasets through vector streaming - meaning you can process 10GB+ files without running out of RAM. Whether you're building a search engine or recommendation system, you can start with pip install embed-anything and a few lines of Python.

  • 5 min read

Vector Streaming

Introducing vector streaming in EmbedAnything, a feature designed to optimize large-scale document embedding. By enabling asynchronous chunking and embedding using Rustโ€™s concurrency, it reduces memory usage and speeds up the process. We also show how to integrate it with the Weaviate Vector Database for seamless image embedding and search.