关于ppocr的动态配置问题

[duskkkrua]

config.EnableTensorRtEngine(1 << 30, 1, 20, precision, false, false);
if (!Utility::PathExists("./trt_det_shape.txt")) {
config.CollectShapeRangeInfo("./trt_det_shape.txt");
} else {
config.EnableTunedTensorRtDynamicShape("./trt_det_shape.txt", true);
}
同样在这一段卡顿太久，想知道能否有代码能运行获取它所影响的权重名称

W0122 10:58:00.951747 32272 helper.h:127] Try increasing the workspace size to 4194304 bytes to get better performance.
W0122 10:58:01.727627 32272 helper.h:127] TensorRT encountered issues when converting weights between types and that could affect accuracy.
W0122 10:58:01.727627 32272 helper.h:127] If this is not the desired behavior, please modify the weights or retrain with regularization to adjust the magnitude of the weights.
W0122 10:58:01.728623 32272 helper.h:127] Check verbose logs for the list of affected weights.
W0122 10:58:01.728623 32272 helper.h:127] - 17 weights are affected by this issue: Detected subnormal FP16 values.

Completed 100/100 predictions（这行是我自己写的）

=====================================
运行成功，但是不知道是哪17个。
我在调试控制台窗口中，没有找到类似trt input [lstm_1.tmp_0] dynamic shape info not set, please check and retry.这样的报错信息。

[duskkkrua]

另外还想请问一下，我启用了EnableSaveOptimModel，想不用每次启动都进行优化编译。
生成了_optimized.pdiparams和_optimized.pdmodel，但它仍在编译。