SLIM-SQL-TOOL

slim-sql-tool is a 4_K_M quantized GGUF version of slim-sql-1b-v0, providing a small, fast inference implementation, optimized for multi-model concurrent deployment.

slim-sql is part of the SLIM ("Structured Language Instruction Model") series, providing a set of small, specialized decoder-based LLMs, fine-tuned for function-calling.

Note: slim-sql is designed for small, fast, local prototyping and to be effective for 'one-table' lookups - it was not trained or optimized for complex joins and other sophisticated SQL queries.

To pull the model via API:

from huggingface_hub import snapshot_download           
snapshot_download("llmware/slim-sql-tool", local_dir="/path/on/your/machine/", local_dir_use_symlinks=False)  

Load in your favorite GGUF inference engine, or try with llmware as follows:

from llmware.models import ModelCatalog  

# this one line will download the model and run a series of tests
# includes two sample table schema - go to llmware github repo for end-to-end example  
ModelCatalog().tool_test_run("slim-sql-tool", verbose=True)  

Slim models can also be orchestrated as part of multi-model, multi-step LLMfx calls:

from llmware.agents import LLMfx

llm_fx = LLMfx()
llm_fx.load_tool("sql")
response = llm_fx.sql(query, table_schema)  

Note: please review config.json in the repository for prompt wrapping information, details on the model, and full test set.

Note: two sample 'hello world' csv tables are included - this is fabricated data - any similarity with real people is coincidental.

Model Card Contact

Darren Oberst & llmware team

Any questions? Join us on Discord