considerations of computational time rule out taking n-character chunks of code, counting the frequency of the visible ascii characters, 32 through 127, and training our neural net on the basis of these counts and target information about the code s language 考慮到排除取得n個(gè)字符代碼所需的計(jì)算時(shí)間,統(tǒng)計(jì)從ascii碼的32到127的范圍內(nèi)可視ascii碼字符出現(xiàn)的頻率,并在這個(gè)統(tǒng)計(jì)以及關(guān)于代碼語(yǔ)言的目標(biāo)信息的基礎(chǔ)上培訓(xùn)神經(jīng)網(wǎng)絡(luò)。