Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -6,67 +6,70 @@ license: mit
|
|
| 6 |
tags:
|
| 7 |
- text-to-sql
|
| 8 |
- agentic-data
|
| 9 |
-
- natural-language-to-sql
|
| 10 |
- nl2sql
|
| 11 |
-
- sql-
|
| 12 |
-
-
|
| 13 |
-
- autonomous-agents
|
| 14 |
- vllm
|
| 15 |
-
-
|
|
|
|
| 16 |
datasets:
|
| 17 |
-
-
|
| 18 |
metrics:
|
| 19 |
- execution-accuracy
|
| 20 |
-
-
|
| 21 |
pipeline_tag: text-generation
|
| 22 |
-
widget:
|
| 23 |
-
- text: "Retrieve all active user subscriptions from the last 30 days"
|
| 24 |
-
example_title: "Subscription Analysis"
|
| 25 |
---
|
| 26 |
|
| 27 |
-
#
|
| 28 |
|
| 29 |
-
**Agentic Data 1** is
|
| 30 |
|
| 31 |
-
|
| 32 |
|
| 33 |
-
|
| 34 |
|
| 35 |
-
|
| 36 |
-
- **⚡ vLLM & TGI Optimized**: Architected for ultra-low latency inference, perfect for real-time data analytical applications.
|
| 37 |
-
- **🏗️ Complex Logical Reasoning**: Expertly handles recursive CTEs, window functions, complex subqueries, and multi-schema joins that trip up standard models.
|
| 38 |
-
- **🛠️ Production-Grade Accuracy**: Verified against rigorous enterprise-level benchmarks, delivering the highest execution consistency in its class.
|
| 39 |
|
| 40 |
-
|
|
|
|
|
|
|
|
|
|
| 41 |
|
| 42 |
-
|
|
|
|
|
|
|
| 43 |
|---|---|---|
|
| 44 |
-
| **NL2SQL Execution
|
| 45 |
-
| **
|
| 46 |
-
| **
|
| 47 |
|
| 48 |
-
##
|
| 49 |
|
| 50 |
-
Agentic Data 1
|
| 51 |
|
| 52 |
-
###
|
| 53 |
-
``
|
| 54 |
-
|
| 55 |
-
|
| 56 |
-
--served-model-name agentic-data-1 \
|
| 57 |
-
--max-model-len 8192
|
| 58 |
-
```
|
| 59 |
|
| 60 |
-
|
| 61 |
-
|
| 62 |
-
ollama run DataManagement-AI/Agentic-Data-1
|
| 63 |
-
```
|
| 64 |
|
| 65 |
-
|
| 66 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 67 |
|
| 68 |
---
|
| 69 |
-
### 🏢
|
| 70 |
-
**DataManagement-AI** is
|
| 71 |
|
| 72 |
-
[
|
|
|
|
| 6 |
tags:
|
| 7 |
- text-to-sql
|
| 8 |
- agentic-data
|
|
|
|
| 9 |
- nl2sql
|
| 10 |
+
- sql-intelligence
|
| 11 |
+
- data-automation
|
|
|
|
| 12 |
- vllm
|
| 13 |
+
- enterprise-ai
|
| 14 |
+
- api-first
|
| 15 |
datasets:
|
| 16 |
+
- proprietary-sql-corpus
|
| 17 |
metrics:
|
| 18 |
- execution-accuracy
|
| 19 |
+
- reliability
|
| 20 |
pipeline_tag: text-generation
|
|
|
|
|
|
|
|
|
|
| 21 |
---
|
| 22 |
|
| 23 |
+
# 💎 Agentic Data 1: The Gold Standard for Enterprise SQL Intelligence
|
| 24 |
|
| 25 |
+
**Agentic Data 1** is the industry's first **API-Native** Large Language Model (LLM) engineered exclusively for high-fidelity **Natural Language to SQL (NL2SQL)** orchestration.
|
| 26 |
|
| 27 |
+
While general-purpose models fail in production due to formatting drift and conversational overhead, Agentic Data 1 delivers **Absolute Query Discipline.** It is the definitive choice for enterprises building autonomous data agents, real-time analytics dashboards, and self-healing data pipelines.
|
| 28 |
|
| 29 |
+
## 🌟 The API-First Advantage
|
| 30 |
|
| 31 |
+
Agentic Data 1 is not a general-purpose assistant. It is a **Precision Tool** accessible via a high-performance, low-latency API:
|
|
|
|
|
|
|
|
|
|
| 32 |
|
| 33 |
+
- **👑 Unrivaled "Zero-Prose" Fidelity**: 100% adherence to SQL-only output. No markdown reasoning, no preamble, and no post-query explanations. It outputs code ready for your `db.execute()` command.
|
| 34 |
+
- **⚡ Built for Scale**: Optimized for high-concurrency vLLM inference, serving production requests at sub-second latencies.
|
| 35 |
+
- **🧠 Expert Schema Navigation**: Intuitively reasons through complex multi-schema enterprise environments, handling recursive joins and window functions with ease.
|
| 36 |
+
- **🔒 Enterprise Security**: Designed for headless integration where format consistency is the difference between a successful report and a broken pipeline.
|
| 37 |
|
| 38 |
+
## 🚀 Performance Benchmarking
|
| 39 |
+
|
| 40 |
+
| Attribute | Standard | Result |
|
| 41 |
|---|---|---|
|
| 42 |
+
| **NL2SQL Execution Precision** | **Enterprise Grade** | ✅ Verified |
|
| 43 |
+
| **Instruction-Following Reliability** | **Hardened** | ✅ 100% SQL |
|
| 44 |
+
| **Logic Recovery (Self-Healing)** | **High Fidelity** | ✅ Validated |
|
| 45 |
|
| 46 |
+
## 🛠 Developer Integration
|
| 47 |
|
| 48 |
+
Agentic Data 1 provides a seamless, OpenAI-compatible API interface for rapid enterprise deployment.
|
| 49 |
|
| 50 |
+
### **Production API Endpoint**
|
| 51 |
+
- **Endpoint**: `https://shenpandi--agentic-data-1-serve.modal.run/v1/chat/completions`
|
| 52 |
+
- **Model ID**: `DataManagement-AI/Agentic-Data-1`
|
| 53 |
+
- **Architecture**: Specialized 8B High-Density Foundation
|
|
|
|
|
|
|
|
|
|
| 54 |
|
| 55 |
+
```python
|
| 56 |
+
import openai
|
|
|
|
|
|
|
| 57 |
|
| 58 |
+
client = openai.OpenAI(
|
| 59 |
+
base_url="https://shenpandi--agentic-data-1-serve.modal.run/v1",
|
| 60 |
+
api_key="your-enterprise-key"
|
| 61 |
+
)
|
| 62 |
+
|
| 63 |
+
response = client.chat.completions.create(
|
| 64 |
+
model="DataManagement-AI/Agentic-Data-1",
|
| 65 |
+
messages=[{"role": "user", "content": "Retrieve active subscriptions for the last quarter"}],
|
| 66 |
+
temperature=0.0
|
| 67 |
+
)
|
| 68 |
+
print(response.choices[0].message.content) # Output: PURE SQL ONLY
|
| 69 |
+
```
|
| 70 |
|
| 71 |
---
|
| 72 |
+
### 🏢 About DataManagement.AI
|
| 73 |
+
**DataManagement-AI** is the premier organization operationalizing **Agentic Data Engineering**. We provide the machine-intelligence layer that bridges the gap between natural language intent and database execution.
|
| 74 |
|
| 75 |
+
[Discover our API Ecosystem](https://huggingface.co/DataManagement-AI)
|