Blog

Andrey Vasnetsov, Co-founder and CTO at Qdrant has shared about vector search and applications with Learn NLP Academy. 

He covered the following topics:

• Qdrant search engine and Quaterion similarity learning framework;
• Similarity learning to multimodal settings;
• Elastic search embeddings vs vector search engines;
• Support for multiple embeddings;
• Fundraising and VC discussions;
• Vision for vector search evolution;
• Finetuning for out of domain.

We’ve released the new version of Qdrant engine - v.0.9.0. It features the dynamic cluster scaling capabilities. Now Qdrant is more flexible with cluster deployment, allowing to move shards between nodes and remove nodes from the cluster.

v.0.9.0 also has various improvements, such as removing temporary snapshot files during the complete snapshot, disabling default mmap threshold, and more.

You can read the detailed release noted by this link https://github.com/qdrant/qdrant/releases/tag/v0.9.0

We keep improving Qdrant and working on frequently requested functionality for the next release. Stay tuned!

We’re happy to share the result of the work we’ve been into during the last months - Quaterion. It is a framework for fine-tuning similarity learning models that streamlines the training process to make it significantly faster and cost-efficient.

To develop Quaterion, we utilized PyTorch Lightning, leveraging a high-performing AI research approach to constructing training loops for ML models.

This framework empowers vector search solutions, such as semantic search, anomaly detection, and others, by advanced coaching mechanism, specially designed head layers for pre-trained models, and high flexibility in terms of customization according to large-scale training pipelines and other features.

Here you can read why similarity learning is preferable to the traditional machine learning approach and how Quaterion can help benefit https://quaterion.qdrant.tech/getting_started/why_quaterion.html#why-quaterion   

A quick start with Quaterion:https://quaterion.qdrant.tech/getting_started/quick_start.html

And try it and give us a star on GitHub :) https://github.com/qdrant/quaterion

The new version of our engine - v0.8.0, went live. Let’s go through the new features it has:

• On-disk payload storage allows storing more with less RAM usage.
• Distributed deployment support is available. And we continue improving it, so stay tuned for new updates.
• The payload can be indexed in the process without rebuilding the segment.
• Advanced filtering support now includes filtering by similarity score.

Also, it has a faster payload index, better error reporting, HNSW Speed improvements, and many more. Check out the change log for more details https://github.com/qdrant/qdrant/releases/tag/v0.8.0.

We’ve released the new version of Qdrant neural search engine. 

Let’s see what’s new in update 0.7.0.

• 0.7 engine now supports JSON as a payload. 
• It redeems a lost API. Alias API in gRPC is available.
• Provides new filtering conditions: refactoring, bool, IsEmpty, and ValuesCount filters are available. 
• It has a lot of improvements regarding geo payload indexing, HNSW performance, and many more.

Read detailed release notes on GitHub.

Stay tuned for new updates.
If you have any questions or need support, join our Discord community.