---
license: apache-2.0
base_model: Qwen/Qwen2.5-Coder-7B-Instruct
datasets:
  - b-mc2/sql-create-context
language:
  - en
library_name: transformers
pipeline_tag: text-generation
tags:
  - text-to-sql
  - text2sql
  - sql
  - qlora
  - peft
  - unsloth
  - qwen2.5
---

# Qwen2.5-Coder-7B Text-to-SQL — filtered-train variant (QLoRA)

Same recipe as [`junmingg/qwen2.5-coder-7b-text2sql`](https://huggingface.co/junmingg/qwen2.5-coder-7b-text2sql),
but the **80 training rows whose gold SQL is unparseable** (0.32% — WikiSQL label noise) were dropped before
fine-tuning (24,920 train rows). This is the **label-noise ablation** described in the main model card.

## Results (held-out 500 examples — identical references)

| Model | Exact match | Semantic equiv. | SQL validity |
|---|---|---|---|
| Base (zero-shot) | 3.8% | 67.0% | 100.0% |
| Main model (full 25k train) | 78.8% | 86.2% | 99.2% |
| **This model (filtered 24.9k train)** | 78.6% | 85.4% | **99.8%** |

![ablation](benchmark.png)

On the valid-reference subset (497 rows) this model reaches **100%** validity. Cleaning the noisy labels moved
**validity, not accuracy** — exact-match and semantic-equivalence are unchanged within run-to-run noise.

**Use the main model unless you specifically want this ablation.** Usage, training details, evaluation
methodology, and limitations are identical to the
[main model card](https://huggingface.co/junmingg/qwen2.5-coder-7b-text2sql). Code + harness:
https://github.com/junmingg/Unsloth-Qwen2.5-Coder-7b-Text-to-SQL-SFT

## License & attribution

Apache-2.0 (base model). Training data: `b-mc2/sql-create-context` (CC-BY-4.0).