def expression2simptrad(self, expression): result = {} for charmode, glangcode in [("simp", "zh-CN"), ("trad", "zh-TW")]: # Query Google for the conversion, returned in the format: ["社會",[["noun","社會","社會","社會"]]] log.info("Doing conversion of %s into %s characters", expression, charmode) meanings = dictionaryonline.gTrans(expression, glangcode, False) if meanings is None or len(meanings) == 0: # No conversion, so give up and return the input expression result[charmode] = expression else: # Conversion is stored in the first 'meaning' result[charmode] = model.flatten(meanings[0]) return result
def generateincharactersystem(self, expression, charmode): log.info("Doing conversion of %s into %s characters", expression, charmode) # Query Google for the conversion, returned in the format: ["社會",[["noun","社會","社會","社會"]]] if charmode=="simp": glangcode="zh-CN" else: glangcode="zh-TW" meanings = dictionaryonline.gTrans(expression, glangcode, False) if meanings == None or len(meanings) == 0: # No conversion, so give up and return the input expression return expression else: # Conversion is stored in the first 'meaning' return model.flatten(meanings[0])
def generateincharactersystem(self, expression, charmode): log.info("Doing conversion of %s into %s characters", expression, charmode) # Query Google for the conversion, returned in the format: ["社會",[["noun","社會","社會","社會"]]] if charmode == "simp": glangcode = "zh-CN" else: glangcode = "zh-TW" meanings = dictionaryonline.gTrans(expression, glangcode, False) if meanings == None or len(meanings) == 0: # No conversion, so give up and return the input expression return expression else: # Conversion is stored in the first 'meaning' return model.flatten(meanings[0])
def preparetokens(config, tokens): if config.colorizedpinyingeneration: tokens = transformations.colorize(config.tonecolors, tokens) return model.flatten(tokens, tonify=config.shouldtonify)
def generatecoloredcharacters(self, expression): return model.flatten(transformations.colorize(self.config.tonecolors, transformations.tonesandhi(self.dictionary.tonedchars(expression))))
def expressiondictreading2color(self, expression, dictreading): return model.flatten(transformations.colorize(self.config.tonecolors, model.tonedcharactersfromreading(expression, dictreading)))
def generatecoloredcharacters(self, expression): return model.flatten( transformations.colorize( self.config.tonecolors, transformations.tonesandhi( self.dictionary.tonedchars(expression))))
y_true_cls = tf.argmax(y_true, dimension=1) j = model.multi_layer_cnn( inp=x, no_cnn_layer=10, channel=3, filter_size=3, filter_no_list=[32, 64, 128, 64, 128, 256, 512, 1024, 512, 256], stride_size=2, pool_size=2, pstride_size=2) for key, value in j.items(): globals()[key] = value flat_layer = model.flatten(globals()[key]) y = model.multi_layer_dnn(flat_layer, 4, [128, 256, 512, 256]) for key, value in y.items(): globals()[key] = value #final = globals()[key] final = model.final_layer( globals()[key], num_inputs=globals()[key].get_shape()[1:4].num_elements(), num_outputs=6, use_relu=False) ses = model.predict(final,