Skip to content

Code translation project for Java and Python across Avatar and CodeNet datasets using multiple state-of-the-art LLMs (DeepSeek R1, DeepSeek V3, Llama-4 Maverick, GPT-4o and Qwen 2.5). Supports both direct and algorithm-based translation techniques for robust program translation.

Notifications You must be signed in to change notification settings

sdipto7/llm-code-translation

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

87 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Models

[deepseek-r1, deepseek-v3, llama-4-maverick, gpt-4o, qwen-2.5]

Datasets

[avatar, codenet]

Source Languages

[python, java]

Target Languages

[java, python]

Translation Types

[direct, algorithm]

Command Syntax

Translation Command

Translate from source language to target language using a specific translation type, dataset, and model.

bash scripts/translate.sh ${model} ${dataset} ${source_lang} ${target_lang} ${translation_type}

Testing Command

Test the translated code for a specific translation type, dataset, and model.

bash scripts/test.sh ${model} ${dataset} ${source_lang} ${target_lang} ${translation_type}

Note:

Substitute ${model} with a supported model (e.g., gpt-4o), ${dataset} with a dataset name (e.g., avatar), ${source_lang} with a source language (e.g., python), ${target_lang} with a target language (e.g., java), and ${translation_type} with the desired translation type (e.g., direct).

About

Code translation project for Java and Python across Avatar and CodeNet datasets using multiple state-of-the-art LLMs (DeepSeek R1, DeepSeek V3, Llama-4 Maverick, GPT-4o and Qwen 2.5). Supports both direct and algorithm-based translation techniques for robust program translation.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published