Retriv at BLP-2025 Task 2: Test-Driven Feedback-Guided Framework for Bangla-to-Python Code Generation

TL;DR

This paper presents a test-driven, feedback-guided framework for generating Python code from Bangla instructions using a fine-tuned LLM, achieving high accuracy in a shared task despite low-resource language challenges.

Contribution

It introduces a novel iterative, test-driven approach for Bangla-to-Python code generation with feedback, improving performance in low-resource language settings.

Findings

Achieved Pass@1 score of 0.934 in the shared task

Demonstrated effectiveness of test-driven iterative refinement

Highlighted challenges in Bangla instruction understanding

Abstract

Large Language Models (LLMs) have advanced the automated generation of code from natural language prompts. However, low-resource languages (LRLs) like Bangla remain underrepresented due to the limited availability of instruction-to-code datasets and evaluation benchmarks. To address this, the BLP Workshop at IJCNLP-AACL 2025 introduced a shared task on "Code Generation in Bangla". In this work, we propose a method that combines instruction prompting with a test-driven, feedback-guided iterative refinement process using a fine-tuned Qwen2.5-14B model. The model generates code from Bangla instructions, tests it against unit tests, and iteratively refines any failing outputs through three evaluation passes, using test feedback to guide each step. This approach helped our team "Retriv" to secure 2nd place in the shared task with a Pass@1 score of 0.934. The analysis highlights challenges in Bangla instruction understanding and Python code generation, emphasizing the need for targeted methods in LRLs. We made experimental scripts publicly available for the community.