Evalution v0.0.8

What's Changed

• [CI] split tests into 2 groups, fix 255 limit by @CSY-ModelCloud in #116
• add maturin requirement by @CSY-ModelCloud in #117
• fix 4090 score by @CSY-ModelCloud in #118
• [CI] refactor ci. sync logic/codes from gptqmodel by @CSY-ModelCloud in #119
• [CI] don't clean whl for uv cache by @CSY-ModelCloud in #120
• [CI] use new docker image by @CSY-ModelCloud in #121
• [CI] all use cuda 130 docker image by @CSY-ModelCloud in #122
• increase timing, fix flaky failures by @CSY-ModelCloud in #124
• fix old transformers has a different func name by @CSY-ModelCloud in #125
• fix 2 positional arguments but 3 were given by @CSY-ModelCloud in #126
• Fix cb compat for Transformers 5.14+ by @Qubitium in #127

Full Changelog: v0.0.7...v0.0.8