tutorial for llama2_ssd教程的疑问

[Merrybeans]

高通官方发布的tutorial_for_llama2_ssd_auto教程里，training步骤对ssd模块的训练loss下降应该是怎样的？目前我按教程做，loss从7.12降到2.78好像就不能再下降了。

而且，在生成LLaMA with SSD model libraries的时候，会出现这样的报错，完全看不懂，谁知道这是什么意思吗？：

ERROR - qnn-model-lib-generator: stderr : jni/kvcache_1_of_4.cpp:37:51: warning: mixture of designated and non-designated initializers in the same initializer list is a C99 extension [-Wc99-designator] .hybridCoo= {.numSpecifiedElements= 0, .numSparseDimensions= 0}}, ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ jni/QnnWrapperUtils.hpp:77:17: note: expanded from macro 'VALIDATE' retStatus = value; \ ^~~~~ jni/kvcache_1_of_4.cpp:36:51: note: first non-designated initializer is here