incedo/codellama-ft-7b-v1.0
The incedo/codellama-ft-7b-v1.0 is a 7 billion parameter CodeLlama-based causal language model developed by incedo. This model is fine-tuned on over 400 test scripts written in Java using Cucumber and Selenium frameworks. It specializes in generating code related to test automation, making it highly effective for developers working with these specific technologies.
Loading preview...
Model Overview
The incedo/codellama-ft-7b-v1.0 is a specialized 7 billion parameter language model built upon the codellama/CodeLlama-7b-hf base. It has been meticulously fine-tuned by incedo using a proprietary dataset comprising over 400 test scripts. These scripts are specifically crafted in Java, leveraging the popular Cucumber and Selenium frameworks.
Key Capabilities
- Specialized Code Generation: Excels at generating code snippets and logic for test automation scenarios.
- Java, Cucumber, and Selenium Focus: Optimized for tasks involving these specific programming languages and testing frameworks.
- Fine-tuned Performance: Benefits from targeted training on real-world test scripts, enhancing its relevance for QA and development workflows.
Training Details
The model was trained for 25 epochs with a batch size of 2, utilizing paged_adamw_32bit optimizer and a learning rate of 2e-4. LoRA configuration included lora_alpha=16, lora_dropout=0.1, and r=64.
Ideal Use Cases
This model is particularly well-suited for developers and QA engineers who need assistance with:
- Automating test script generation in Java.
- Developing or extending test cases using Cucumber.
- Interacting with web elements via Selenium in a Java context.