Errors in Fuzz Drivers Generation?

[HushShushHush]

Hello,

I modify the config.rs and incoder.py to invoke a self-build LLM Incoder-1B. In the process of generating fuzz drivers, however, the Docker container will always receive an error message: ERROR [prompt_fuzz::program] Detected corrupt state of your last execution, please remove your library output in the output directory and restart, even though I have removed the corresponding output before interaction.

Are there any scripts that need to be further modified?

Thanks.

[Yunlongs]

@HushShushHush Hi! Thanks for your try. The support for various LLMs is in the beta version. If there is no special requirement, I strongly recommend you use the model of ChatGPT or GPT4 that is currently sufficiently tested. Supporting other LLMs requires a great of effort and I am still working out for it.

[Yunlongs]

I added new features in v0.0.1 to support more LLM services. The requirement is that you deploy your models using the standard OPENAI API interfaces. See in usage.