Huawei Technologies Co., Limited

Tencent Technology (shenzhen) Company Limited

Tencent Technology Company Limited

Wei Xue

Xing Jin

Yi Li

Lei Xiao

Yong Li

Qian Wang

Zhimao Guo

Bo Song Chen

Yongqiang Zou

Ke Zhang

Yang Wang

Hua Cai

Eryu Wang

Ke Yi

Wei Wang

Bin Xiao

Bo Chen

Tianxiang Wang

Yongjian Hu

Yanzhao Pang

Dapeng Liu

Xue Bai

Chunjian Bao

Jinming Huang

Chuanjiang Luo

Shubin Zhang

Peili Lv

Jiadao Wei

Yanwen Xiang

Weidong Shao

Xiangdong Deng

Xinyue Ma

Miao Xu

Jinjing Liu

Lili Zhao

Jinchao Gao

A packet measurement method includes obtaining, by a network device, reverse measurement indication information that is used to indicate a request to measure a reverse packet flow of a first packet flow, after receiving the first packet flow, adding, by the network device, the reverse measurement indication information to a packet in the first packet flow, and sending, by the network device, the packet that is in the first packet flow and that carries the reverse measurement indication information. By adding the reverse measuring indication information to the packet in the first packet flow, the method indicates a request to another network device on the packet forwarding path to measure the reverse packet flow of the first packet flow. Measurement on bidirectional packet flows is implemented through one measurement deployment.

Packet measurement method, device, and system

A port mode adaptation method includes a network device determining a status of a port on the network device. The method includes the network device setting a mode of the port to an Ethernet interface mode if the status of the port satisfies a first condition. In this way, when the network device identifies the mode of the port on the network device as the FlexE interface mode, and there is no link at the FlexE shim of the port, the network device may determine that a mode used by a port interconnected with the port is not a FlexE interface mode. In this case, the network device may adaptively set the mode of the port to the Ethernet interface mode, to ensure that the port on the network device can use a same mode as another port interconnected with the port.

Port mode adaptation method and apparatus

A data processing method in a data processing device is provided. First to-be-processed data input into a neural network are obtained. Iterative training is performed on the neural network for a first preset number of times by using first target data in the first to-be-processed data, to obtain a seed model of the neural network. First newly added data generated after an elapse of time corresponding to the first time window is obtained, and the first newly added data and the first to-be-processed data are combined into second to-be-processed data. Iterative training is performed on the seed model for a second preset number of times by using second target data in the second to-be-processed data, to obtain a first incremental model of the neural network. A first preset area overlaps between the second time window and the first time window. The first incremental model online is published.

Data processing method and apparatus for training incremental model of neural network

A method for transmitting data in flexible Ethernet (FlexE) and a device comprising transmitting data in FlexE whereby a first FlexE device receives, according to a first client calendar, data from a second FlexE device; the first FlexE device determines, based on an error data block in the data, that a first timeslot is damaged; the first FlexE device adds a timeslot damage notification to an overhead frame to be sent to the second FlexE device; the first FlexE device receives a second client calendar from the second FlexE device; and the first FlexE device receives, using the second client calendar, the data from the second FlexE device.

Method for transmitting data in flexible ethernet and device

A data processing method is described. A processor provides different training data sets to a plurality of graphics processing units (GPUs), respectively. The processor controls the plurality of GPUs to generate respective sets of modification parameters by performing respective training processes in a parallel manner using respectively managed versions of a predictive model according to the corresponding training data sets. The processor controls the plurality of GPUs to exchange, before completion of the respective training processes, at least a portion of the sets of modification parameters that have been generated by the plurality of GPUs. The processor also causes the plurality of GPUs to modify the respectively managed versions of the predictive model according to at least the portion of the sets of modification parameters exchanged among the plurality of GPUs.

Data processing method, apparatus, and electronic device

A method and device for processing promotion information and a system are provided. The method includes that: agreement information and exposure requirements of all promotion information within a preset period are acquired (); directional delivered targets are determined according to the agreement information and the exposure requirements, and the directional delivered targets are split into multiple non-intersected delivered target sets (); the promotion information is delivered to users corresponding to the corresponding delivered target sets according to the exposure requirements (); statistics about social propagation amounts of to the delivered promotion information is made in real time in a delivery process (); and exposure parameters are corrected according to the social propagation amounts (), so that delivery of the promotion information is regulated in real time. By the method, the effectiveness and accuracy of delivering the promotion information may be improved.

Method and device and system for processing promotion information

A parallel data processing method based on multiple graphic processing units (GPUs) is provided, including: creating, in a central processing unit (CPU), a plurality of worker threads for controlling a plurality of worker groups respectively, the worker groups including one or more GPUs; binding each worker thread to a corresponding GPU; loading a plurality of batches of training data from a nonvolatile memory to GPU video memories in the plurality of worker groups; and controlling the plurality of GPUs to perform data processing in parallel through the worker threads. The method can enhance efficiency of multi-GPU parallel data processing. In addition, a parallel data processing apparatus is further provided.

Data parallel processing method and apparatus based on multiple graphic processing units

A parallel data processing method based on multiple graphic processing units (GPUs) is provided, including: creating, in a central processing unit (CPU), a plurality of worker threads for controlling a plurality of worker groups respectively, the worker groups including a plurality of GPUs; binding each worker thread to a corresponding GPU; loading one batch of training data from a nonvolatile memory to a GPU video memory corresponding to one worker group; transmitting, between a plurality of GPUs corresponding to one worker group, data required by data processing performed by the GPUs through peer to peer; and controlling the plurality of GPUs to perform data processing in parallel through the worker threads.

Model parallel processing method and apparatus based on multiple graphic processing units

A pre-distortion coefficient corresponding to a transmit power currently set by a transmitting end is adjusted. A signal received quality report which is obtained according to the pre-distortion coefficient before adjustment is analyzed and fed back by a receiving end and a signal received quality report that is obtained according to the pre-distortion coefficient after adjustment and fed back by the receiving end. The analysis determines a variation situation of signal received quality of the receiving end. If the signal received quality of the receiving end becomes better, the pre-distortion coefficient after adjustment is used as the pre-distortion coefficient corresponding to the transmit power currently set by the transmitting end. The process then returns to the step of adjusting a pre-distortion coefficient corresponding to a transmit power currently set by a transmitting end.

Method and apparatus for adjusting pre-distortion coefficient

An apparatus for implementing digital baseband predistortion includes a transmission channel including a digital-to-analog converter, a modulator, an amplifier and a power amplifier, and further includes a feedback channel analog part including a diode detector, a filter and an analog-to-digital converter, and a feedback channel digital part including a predistorter, a mode obtaining unit, a predistortion coefficient generator and a feedback correcting unit. The diode detector is configured to obtain an envelope of an output signal of the power amplifier. Embodiments of the present invention further provide a method for implementing digital baseband predistortion by applying the foregoing single-chip. Due to a simple structure of the diode detector, not only the number of radio frequency devices on the feedback channel is reduced, implementation complexity of hardware is lowered, power consumption is reduced, but also a feedback channel analog part may be integrated onto the single-chip.

Method and apparatus for implementing digital baseband predistortion

Growing community of inventors

Shenzhen, China

Wei Xue