File size: 499 Bytes
12c421a
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
---
license: apache-2.0
base_model: Qwen/Qwen2.5-Coder-3B-Instruct
tags: [text-to-sql, self-play, grpo, sql-zero]
---
# sql-zero-3b-challenger-iter1

Challenger 3B iter1 (HRPO-lite por nivel, sem rep-penalty). Difficulty 0.028->0.11; gen1 58k pares.

Parte do projeto **SQL-Zero** (self-play zero-anotacao para Text-to-SQL, adaptacao do Dr. Zero).
Protocolo de avaliacao: greedy@1, schema completo (DDL+3 rows), execution accuracy.
Treinado com verl (GRPO); reward de execucao 3-tier (1.0/0.1/0.0).