Knowledge Graph — Coursera Notes › Academic disciplines › Computer Science / Information Technology › Cloud Computing › Snowflake
Snowflake Cortex
service · part of Snowflake
A set of AI functions within Snowflake for tasks like LLM completion, search, and sentiment analysis.
Inside Snowflake Cortex (10)
- Snowflake Cortex ML functions — Managed ML capabilities including FORECAST, ANOMALY_DETECTION, TOP_INSIGHTS, and CLASSIFICATION that typically require training on user data.
- Cortex Analyst — Snowflake service that converts natural-language business questions into SQL using a semantic model, enabling self-service analytics.
- Cortex Search service — A Snowflake service that provides search capabilities over data, queried via SQL or CLI.
- Cortex Fine-Tuning — Snowflake Cortex Fine-Tuning is a feature that allows users to fine-tune supported base models using domain-specific training data stored in Snowflake.
- Cortex LLM functions — Snowflake functions using pretrained language models for tasks like COMPLETE, SUMMARIZE, TRANSLATE, SENTIMENT, and EXTRACT_ANSWER.
- Cortex ML Classification — A higher-level managed function in Snowflake Cortex for classification tasks, offering less customization than Snowpark ML.
- Cortex ML functions — Managed functions for fast tasks such as forecasting, anomaly detection, classification, and metric contribution analysis.
- Cortex Search — Snowflake semantic search service for text stored in Snowflake tables, enabling natural language queries and supporting RAG architectures.
- SNOWFLAKE.CORTEX.COMPLETE — A function that accepts a message array for multi-turn conversations.
- SNOWFLAKE.CORTEX.SUMMARIZE — A Snowflake Cortex function that summarizes text, which can be chained with COMPLETE.
Connections
- Uses CLI
- Related to Cortex Search service
- Related to SNOWFLAKE.CORTEX.COMPLETE
- Related to SNOWFLAKE.CORTEX.SUMMARIZE
- Related to SNOWFLAKE.CORTEX.COMPLETE
- Related to mistral-7b
- Related to Cortex LLM functions
This is the text view of an interactive 3D knowledge graph — open this page with JavaScript enabled to explore it visually.