Update README.md

README.md

@@ -17,16 +17,51 @@ datasets:
 
 # **NaturalQuery-6.7B-v0.1**
 
+### NaturalQuery is a series of models with state-of-the-art performance on Text to SQL instructions.
+
 <img src="https://cdn-uploads.huggingface.co/production/uploads/648a374f00f7a3374ee64b99/hafdsfrFCqrVbATIzV_EN.png" width="600">
 
 **NaturalQuery** is a LLM that can translate natural language queries to SQL based on your schema. It is finetuned on 8k text to PostgreSQL Natural Language <> SQL pairs.
+NaturalQuery matches the state of the art models in text to sql for it's size and produces the best in the field for complex questions. 
+
+Here is a write up, small test done [here](https://chatdb.ai/post/naturalsql-vs-sqlcoder-for-text-to-sql).
+
+# Table of Contents
+1. [Benchmarks](#Benchmarks)
+    - [SQL-Eval on Novel Datasets](#SQL-Eval-on-Novel-Datasets)
+    - [SQL-Eval by Category](#SQL-Eval-by-Category)
+2. [Future Improvements](#Future-Improvements)
+3. [Usage](#Usage)
+    - [Installation](#Installation)
+    - [Loading the Model](#Loading-the-Model)
+    - [Generating Text](#Generating-Text)
+4. [SQL Generation Template](#SQL-Generation-Template)
+5. [Example SQL Output](#Example-SQL-Output)
+    - [Example Schemas](#Example-Schemas)
+    - [Example Queries](#Example-Queries)
+
+## **Benchmarks**
+
+## SQL-Eval on novel datasets not seen in training
+
+<img src="https://cdn-uploads.huggingface.co/production/uploads/648a374f00f7a3374ee64b99/omYzbu9ZO8hx7cnUtOZKg.png" width="600">
+
+<em>Big thanks to the [defog](https://huggingface.co/defog) team for open sourcing [sql-eval](https://github.com/defog-ai/sql-eval)</em>👏
+
+## **SQL-Eval by SQL Category**
+
+**NaturalQuery-6.7B-v0 matches or outperforms other industry leading models in multiple categories!**
+
+<img src="https://cdn-uploads.huggingface.co/production/uploads/648a374f00f7a3374ee64b99/G6lGyTbBUmEkKdmJ9r4jp.png" width="600">
+
+_The **date** category will be a strong focus in the next iteration of `v1`._
 
-**Future Improvements**:
+**Future Improvements in the next iteration**:
 
 - Much larger training set
 - More complex schemas, questions, and queries
+- Strong focus on Date Queries
 - Reward modeling via DPO
-- Benchmarking
 
 # **Usage**