Sedrick

@monilshah98 Hello, Did you find the answer? I have the same problem .

> You mention that feature selection was done using opensmile with initial feature set of 6373, and then feature selection was performed. > > What is the config file used...

> dia125_utt3.mp4 > I also met this issue, and I found the utterance in the show and cut the video by myself, u can check if it gonna work bro....

> > 哦哦那个博客确实访问不了了，我看看能不能找到备份。这些数字就是用论文里说的audio处理库 提取出所有features之后，我们需要的 feature对应的 indices > > 噢噢噢好的 另外那个数据加载的内存那块您能优化一下吗，大于32G真的太大了，实验室机器跑不起来 @HEBAULC 您好，我正在做端到端的多模态识别，我有几个问题想请教一下，请问您方不方便留个联系方式？