Beijing Xiaomi Pinecone Electronics Co., Ltd.

Zhejiang Furun Dyeing and Printing Co., Ltd.

Yifeng Wang

Guodu Cai

Xin Wang

Peng Gao

Shuo Yang

Lihan Li

Guozhu Fu

Xuan Sun

A training method includes: acquiring voice data of a user on a plurality of devices, and extracting acoustic features based on the voice data to construct a training data set; inputting the training data set into a neural network model, training the neural network model, and acquiring a loss function of each training of the neural network model; and determining that the training of the neural network model is completed when the loss function converges, wherein the loss function is a weighted sum of a classification loss function, a reconstruction loss function, a difference loss function, and a similarity loss function.

Method for training a voiceprint extraction model and method for voiceprint recognition, and device and medium thereof

A method for manufacturing fabric with flat screen and digital printed patterns includes: singeing, pretreatment, flat screen scraping sizing, drying, pattern printing, humidification, steaming treatment, washing, and non-ironing and soft setting. The pretreatment is a biological enzyme pretreatment process including: cloth feeding at 75° C., dipping, scouring and bleaching in a working fluid, heating to 90° C. for treatment of 30 minutes, draining, hot water washing and cold water washing. A shortened process flow can be realized, and the positioned sizing that uses a flat screen to scrape a sizing agent to a pattern position to be digital spray-printed on instead of traditional full-width sizing for sizing and printing, which greatly reduces the amount of sizing agent used, lowers production cost and reduces waste water pollution, and solves problem that the fabric is mottled caused by the application of sizing agent in areas without printed pattern due to full-width sizing.

Method for manufacturing fabric with high-efficiency and high-precision flat screen and digital printed patterns with energy saving and emission-reduction effect

A method for detecting an audio signal, the method comprises: obtaining a speech segment and a non-speech segment of an audio signal to be detected, extracting a first audio feature of the speech segment and a second audio feature of the non-speech segment, detecting the first audio feature using a predetermined speech segment detection model to obtain a first detection score, detecting the second audio feature using a predetermined non-speech segment detection model to obtain a second detection score, and determining whether the audio signal belongs to a target audio based on the first detection score and the second detection score.

Method and device for detecting audio signal, and storage medium

Growing community of inventors

Beijing, China

Yifeng Wang