jastorj/couchmind-v5.8.1_cold_start-cw-30K-16bit

TEXT GENERATIONPricing:Input $0.4 / Cached $0.08 / Output $0.8Concurrent Unit Cost:1Model Size:7.6BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Jul 27, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

jastorj/couchmind-v5.8.1_cold_start-cw-30K-16bit is a 7.6 billion parameter model fine-tuned from Snowflake/Arctic-Text2SQL-R1-7B, specifically optimized for Text-to-SQL generation. It was trained on the NL2SQL++ v5.8.1_cold_start dataset, incorporating code-with-thought reasoning. This model excels at converting natural language questions into syntactically valid Couchbase SQL++ queries, making it ideal for database interaction and data retrieval tasks.

Loading preview...

Model Overview

jastorj/couchmind-v5.8.1_cold_start-cw-30K-16bit is a specialized 7.6 billion parameter language model, fine-tuned from the Snowflake/Arctic-Text2SQL-R1-7B base model. Its primary function is Text-to-SQL generation, specifically targeting Couchbase SQL++ queries. The model was trained using the NL2SQL++ v5.8.1_cold_start dataset, which includes a unique code-with-thought reasoning component to enhance its ability to generate precise and well-reasoned SQL queries.

Key Capabilities

  • Natural Language to SQL++ Conversion: Translates natural language questions into syntactically correct Couchbase SQL++ queries.
  • Code-with-Thought Reasoning: Incorporates a reasoning process to construct detailed and accurate SQL queries, as demonstrated in its training examples.
  • Schema-Aware Query Generation: Designed to work with provided database schemas, ensuring queries adhere to specific bucket, scope, and collection names.
  • Optimized for Couchbase: Tailored for generating queries compatible with Couchbase's SQL++ dialect.

Training Details

The model was fine-tuned using LoRA (Low-Rank Adaptation) with Unsloth, and its weights are quantized to 16-bit. The training dataset comprised 1149 examples, focusing on complex Text-to-SQL scenarios. The training configuration emphasizes precision and adherence to SQL++ syntax, making it a robust tool for database interaction.

Good for

  • Automated SQL++ Query Generation: Ideal for applications requiring automatic conversion of user queries into Couchbase SQL++.
  • Database Interaction: Facilitating easier interaction with Couchbase databases for users without deep SQL++ expertise.
  • Data Retrieval Systems: Enhancing systems that need to dynamically generate queries based on natural language input.

Finetunes
1 models