Thank you very much for your contribution so that we can use this amazing repo, but I had some problems while training my own dataset, and the results didn't look very good. All parameter settings are default. Could you give me some advi...
我正在微调最近发布的IDEA-CCNL/Wenzhong2.0-GPT2-3.5B-chinese模型。 使用的脚本是wenzhong_qa,基于我们的业务场景进行了调整。 由于机器配置的限制,我们想结合DeepSpeed的ZeRO-3进行训练,但似乎并没有对模型参数进行切分。 机器配置如下,8块1080Ti: Every 1.0s: nvid
{if(params.isParamValid("cylinder_axis_point_1") || params.isParamValid("cylinder_axis_point_2")) mooseError("The 'cylinder_axis_point_1' and 'cylinder_axis_point_2' cannot be specified with axisymmetric models. The y-axis is used as the cylindrical axis of symmetry."); p1 = Point(...