Modelarts second training notes

One 、 Speech classification —— Detection and differentiation of different voices

      1. Creating buckets — Import voice files into obs bucket                                                                                                   —— Speech classification — Create project — choice “ New dataset ”— Enter location ： choice train— Output location ：output— spot           blow       “ Create project ”— Synchronous data source — Select object , Add the corresponding label , Finish marking — So click after marking             blow “ Start training ”— Choose free — Submit                                                                                             2.—— Return to automatic learning — find “ Deploy online ”— Online services — Sound detection , look for “ more ”— start-up                               3.—— After the progress is full — stay “ Online services ” in , Click on the file name — Upload — forecast

Two 、 Text classification                                                                                                                                   1.AI Gallery— Asset markets — Data sets — Yes “ Text categorization data sets - Takeout review ”— download — Select bucket for storage （ It's better to build a new barrel ）— determine                                                                                                                         2. moderlarts— Data management — Data sets — Create a dataset — Text — Data set input location ：fenlei— Data set output location ： Create folder out— determine — establish                                                                                                  3. — Click on “ more ”— Import —OBS Directory location ：fenlei— choose “ Text and annotation are separated ”（ An important step ！）— determine — determine — After the pass , Click on “ Release ”（ Even though it is still importing ）                                                                     4.— Open another interface again — Text classification — Create project — Existing data sets — establish — Start training — next step — Submit         — After training , Deployment （ Don't switch to online services , Because we need to use it json）— Conduct “ forecast ”