NiChr0/ schemalytics-gold-agent

11 Downloads Updated 5 months ago

Qwen3.5-4B fine-tuned to generate dbt Gold aggregation models (agg_*) from a Schemalytics Silver layer plan.

ollama run NiChr0/schemalytics-gold-agent

curl http://localhost:11434/api/chat \
  -d '{
    "model": "NiChr0/schemalytics-gold-agent",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='NiChr0/schemalytics-gold-agent',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'NiChr0/schemalytics-gold-agent',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)

Models

Name

1 model

Size / Usage

Context

Input

schemalytics-gold-agent:latest

2.7GB · 256K context window · Text · 5 months ago

schemalytics-gold-agent:latest

2.7GB

256K

Text

Readme

`nichr0/schemalytics-gold-agent`

”`markdown

Schemalytics Gold Agent (Agent 4b)

Fine-tuned Qwen3.5-4B for dbt Gold layer plan generation.

Part of the Schemalytics pipeline — a tool that turns a PostgreSQL schema into a full dbt project using a multi-agent LLM pipeline.

What it does

Given the Silver layer plan (dim/fct models with measures and grains), this model generates Gold layer aggregation models:

agg_<grain>_<name> models for business-level reporting
Selects meaningful aggregation grains (daily, by customer, by product, etc.)
References Silver facts and dimensions

Usage

Used automatically by schemalytics when SCHEMALYTICS_LLM_PROVIDER=ollama:

”`bash ollama pull nichr0/schemalytics-gold-agent SCHEMALYTICS_LLM_PROVIDER=ollama schemalytics generate -c postgresql://… Training Base model: Qwen3.5-4B Method: QLoRA (r=8, α=16) via Unsloth Dataset: ~500 real-world database schemas (Spider benchmark + enterprise schemas) Task: Given Silver plan → output structured Gold aggregation plan (JSON) Parameters Temperature: 0 (deterministic) Context: 8,192 tokens Max output: 2,048 tokens

### `nichr0/schemalytics-gold-agent`

```markdown
# Schemalytics Gold Agent (Agent 4b)

Fine-tuned [Qwen3.5-4B](https://ollama.com/library/qwen3) for **dbt Gold layer plan generation**.

Part of the [Schemalytics](https://github.com/nichr0/schemalytics) pipeline — a tool that turns a
PostgreSQL schema into a full dbt project using a multi-agent LLM pipeline.

## What it does

Given the Silver layer plan (dim/fct models with measures and grains), this model generates Gold
layer aggregation models:

- `agg_<grain>_<name>` models for business-level reporting
- Selects meaningful aggregation grains (daily, by customer, by product, etc.)
- References Silver facts and dimensions

## Usage

Used automatically by `schemalytics` when `SCHEMALYTICS_LLM_PROVIDER=ollama`:

```bash
ollama pull nichr0/schemalytics-gold-agent
SCHEMALYTICS_LLM_PROVIDER=ollama schemalytics generate -c postgresql://...
Training
Base model: Qwen3.5-4B
Method: QLoRA (r=8, α=16) via Unsloth
Dataset: ~500 real-world database schemas (Spider benchmark + enterprise schemas)
Task: Given Silver plan → output structured Gold aggregation plan (JSON)
Parameters
Temperature: 0 (deterministic)
Context: 8,192 tokens
Max output: 2,048 tokens

Paste, drop or click to upload images (.png, .jpeg, .jpg, .svg, .gif)