jastorj/couchmind-v5.7.6.1_qwen_stage_2-cw-12K-16bit

TEXT GENERATIONConcurrent Unit Cost:1Model Size:7.6BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Jun 18, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

jastorj/couchmind-v5.7.6.1_qwen_stage_2-cw-12K-16bit is a 7.6 billion parameter Qwen2.5-Coder-7B-Instruct model fine-tuned by jastorj. It specializes in Text-to-SQL generation, specifically for Couchbase SQL++ queries, leveraging the NL2SQL++ v5.7.6.1_qwen_stage_2 dataset with code-with-thought reasoning. This model is optimized for accurately translating natural language questions into syntactically valid SQL++ queries, making it highly effective for database interaction tasks.

Loading preview...

Model Overview

This model, jastorj/couchmind-v5.7.6.1_qwen_stage_2-cw-12K-16bit, is a specialized fine-tune of the Qwen/Qwen2.5-Coder-7B-Instruct base model. Developed by jastorj, it focuses on the critical task of Text-to-SQL generation, specifically tailored for Couchbase SQL++ queries.

Key Capabilities

  • Couchbase SQL++ Generation: Translates natural language questions into syntactically correct Couchbase SQL++ queries.
  • Schema-Aware Querying: Utilizes provided database schema (bucket, scope, collection names, and field properties) to generate precise queries.
  • Code-with-Thought Reasoning: Fine-tuned on the NL2SQL++ v5.7.6.1_qwen_stage_2 dataset, which incorporates code-with-thought reasoning to enhance query generation logic.
  • Optimized for Accuracy: Emphasizes generating queries that select only explicitly requested columns and adhere strictly to provided schema details.

Training Details

The model was fine-tuned using LoRA (Low-Rank Adaptation) with Unsloth, processing 11,314 training examples. It utilizes 16-bit merged weights and has a maximum sequence length of 12,000 tokens, making it suitable for complex query generation scenarios.

Good For

  • Developers and data analysts working with Couchbase databases who need to automate or assist in SQL++ query generation from natural language.
  • Applications requiring robust and accurate Text-to-SQL capabilities for structured data interaction.
  • Environments where precise query generation based on detailed schema information is paramount.

Finetunes
1 models