danielbrdz
commented
3 months ago Model: https://huggingface.co/braindao/iq-code-evmind-v3.2-qwen2-1.5b-instruct-average
Script: https://github.com/EveripediaNetwork/iq-code-evmind/tree/master/Interesting%20Models
This model is small, its purpose is to help analyze the Solidity benchmark.
In the first early version of the benchmark, it scored 63 points and its base version 58, concluding that fine tuning helped improve its code solidity.
In the second early version of the benchmark, it scored 50 points.
Take the benchmark points with care, the benchmark is still under development and results may vary.
Fine tuning of the Qwen2 1.5b model to optimize its performance and Solidity code generation capability.