Manage LanceDB vector tables, dataset imports, and materialized views.
npx skills add https://github.com/nebius/nebius-physical-ai --skill lancedb-nebius
Please help me install this Agent Skill. Skill: lancedb Source: https://github.com/nebius/nebius-physical-ai/tree/main/skills/tools/lancedb Command: npx skills add https://github.com/nebius/nebius-physical-ai --skill lancedb-nebius
This skill addresses the complexity of managing large-scale vector storage and perception datasets, ensuring consistency across CLI, SDK, and service-level operations.
Use the lancedb skill to create a new table for your perception data and import the BDD100K dataset.
High-intent search queries and answers about installing and using this skill.
You can manage vector storage for perception datasets by creating and maintaining LanceDB tables for frame-aligned records. This skill handles table creation, dataset ingestion, and materialized view maintenance to ensure consistency across CLI and SDK operations.
To slice BDD100K data for failure-mode analysis, you can create materialized views to isolate specific perception scenarios. For example, this skill enables the creation of views like nighttime_person_train to filter and analyze targeted failure cases.
Yes, LanceDB vector-search works with LeRobot and BDD100K datasets. The skill streamlines dataset integration by importing these formats with automated UDF backfills to populate missing data fields automatically.
Yes, generating embeddings with LanceDB requires integration with S3-backed storage for data persistence and GPU-accelerated UDFs for embedding generation. These components support the perception data workflows.
The best way to maintain materialized views for frame-aligned vector records is using LanceDB table operations. This skill manages materialized view maintenance, allowing you to isolate specific perception scenarios like nighttime conditions for targeted analysis.
BermudaLocalsAutomate GPU-accelerated multimodal data curation with deduplication and filtering.
zerwizStores embeddings and metadata in a vector database for semantic search and retrieval.
nebiusProvide domain-specific technical context for robotics and GPU-accelerated AI pipelines on Nebius infrastructure.
gugug168Optimize ClickHouse schemas, queries, and ingestion for analytical workloads.
