使用开源的qwen等模型，在实际项目中生成的用例基本的编译都不能通过

Codium-ai / cover-agent

QodoAI Cover-Agent: An AI-Powered Tool for Automated Test Generation and Code Coverage Enhancement! 💻🤖🧪🐞

https://qodo.ai/

GNU Affero General Public License v3.0

4.38k stars 331 forks source link

Closed liuzyong closed 3 months ago

EmbeddedDevops1 commented 3 months ago

@liuzyong Based on Google translate I read the title as:

Using open source qwen and other models, the use cases generated in actual projects cannot pass basic compilation.

Can you please provide more details on the issue?

redvelvets commented 1 month ago

老哥解决了吗，碰到了的同样的问题，接入cover-agent接入业务代码后，许多生成的测试用例都无法通过（Skipping a generated test that failed），