12results about How to "Improve decoding speed" patented technology

Disclosed is a decoder. At least one computing unit of the decoder is configured to: acquire soft bit information to be decoded of a variable node n in a basic matrix; determine whether the number of decoding iterations i is less than a decoding iteration threshold; if yes, determine whether the number of decoding layers k is smaller than the number of rows a of the basic matrix; if it is determined that k is smaller than a, determine the soft bit information of the variable node n in a kth layer according to the soft bit information of the variable node n in a k-1th layer, and determine variable node information in a k+1th layer according to the soft bit information of the variable node n in the k-1th layer; determine check node information in the k+1th layer according to the variable node information in the k+1th layer; update k to k+1, re-determine whether k is smaller than a; if it is determined that k is greater than or equal to a, update i to i+1, and re-determine whether i is smaller than the decoding iteration threshold until i equals to the decoding iteration threshold. Also disclosed are a decoding method and a computer storage medium.

The invention provides a word line decoding circuit and a static random access memory. Comprising a coding latch module, a first decoding module, a plurality of second decoding modules and a first decoding latch module, the coding latch module receives an address signal and a first clock signal, and codes the address signal to obtain a first group of coded signals; the first decoding module receives the second clock signal and the corresponding signal segment, and decodes the received signal segment according to the second clock signal to obtain a first decoded signal; the second decoding module receives the corresponding signal segment and decodes the received signal segment to obtain a second decoding signal; the first decoding latch module is connected with the first decoding module and the plurality of second decoding modules, and opens the selected word line according to the second decoding signal; the data length of the first address signal representing the row address in the address signals represents the number of the enabled word lines. According to the scheme, the adjustability of the number of word lines can be improved.

The present disclosure provides a video processing method and device and related equipment, and relates to the technical field of video processing. The method comprises: obtaining an encoded data stream of a video, the video comprising a plurality of video frames, each video frame comprising a plurality of slices; for each slice, parsing the encoded data stream of the video to obtain slice data stream of the slice and header information of the slice data stream, triggering generation of an entropy decoding task of the slice, and putting the entropy decoding task of the slice into a task queue, the entropy decoding task being an entropy decoding task using CABAC; selecting a to-be-processed task from the task queue in a multi-thread parallel manner; and performing parallel processing on the to-be-processed task according to the slice data stream corresponding to the to-be-processed task and the header information of the slice data stream corresponding to the to-be-processed task. The method realizes parallel processing of CABAC entropy decoding, and improves the decoding speed of a decoder.

The application discloses a digital television program playing method, an electronic device, a computer readable storage medium, a computer program product, a chip and a digital television terminal, wherein the playing method comprises the following steps: when recording a target program, filtering out a transport stream packet to which an intra-coded frame of the target program belongs, and encrypting the transport stream packet to which the intra-coded frame belongs to obtain an encrypted packet; when playing back the target program, obtaining a recording file of the target program, wherein the recording file comprises the encrypted packet; in a first speed playback mode, decrypting and decoding the encrypted packet in the recording file to obtain an audio and video original code stream of the intra-coded frame; and playing the audio and video original code stream of the intra-coded frame at a first speed, wherein the first speed is higher than a standard speed. The method of the application can avoid large jumps of displayed pictures in a high speed playback mode, and improve display quality.

The application provides a data decoding method, and relates to the technical field of data decoding, and the method comprises the following steps: receiving serial port data, determining the byte length of each data segment in the serial port data based on a preset serial port protocol; storing instruction bits, address bits and data bits in the serial port data in instruction registers, address registers and data registers respectively; obtaining operation instructions, operation addresses and operation data based on the instruction registers, the address registers and the data registers, and performing operations corresponding to the operation instructions, the operation addresses and the operation data to complete decoding of the serial port data. The application can realize decoding of the serial port data through only registers, avoids the problem that the decoding system structure is excessively redundant and complex in the prior art serial port data decoding technology, and improves the decoding speed of the serial port data.

The invention discloses a video processing method and a related device, and the method comprises the steps: obtaining the attribute information of each to-be-decoded frame in a to-be-processed video code stream, enabling the attribute information to be used for indicating the frame type of each to-be-decoded frame, determining the to-be-decoded frame of which the frame type is a non-reference frame from the to-be-processed video code stream based on the attribute information, and removing the to-be-decoded frame. And obtaining a recombined code stream. Frame extraction processing is carried out from the recombined code stream based on the preset frame number to obtain a plurality of sampling frames, the extracted sampling frames are more likely to be key frames, the sampling number of depended frames to be decoded can be reduced, the total number of the extracted sampling frames is further reduced, and when decoding processing is carried out on the plurality of sampling frames to obtain a decoded video, the decoding efficiency of the decoded video is improved. The decoding speed can be greatly improved, and the video frame cutting efficiency is improved. In a word, the total number of the extracted sampling frames can be reduced, the decoding speed is improved, the video frame cutting efficiency is improved, the decoding process of the sampling frames is not affected, and smooth video frame cutting is guaranteed.

This disclosure provides embodiments of a video encoding / decoding and display method, apparatus, electronic device, and computer-readable medium. One specific implementation of the method includes: splitting each video frame to be encoded to obtain a set of video frame blocks; determining a set of video frame block encoding channels that meet interrupt triggering conditions; performing key video frame block peak shifting processing on each video frame block to be encoded to obtain a set of key video frame encoding blocks and a set of channel delay encoding time differences; performing video encoding on each video frame block to be encoded to obtain a set of video frame block channel bitstreams; determining the decoding and splicing threads for the video frame block channel bitstream set to obtain a set of decoding and splicing threads; performing video frame block decoding and splicing on the video frame block channel bitstream set to obtain a decoded video; and performing video frame display processing on the decoded video. This implementation can improve the balance of the bitstream after video encoding, improve the stability and real-time performance of transmitted video, and reduce video transmission latency.

A kind of manchester decoding method based on timing offset algorithm, including filtering module, signal waveform conversion module and manchester decoding module, the timing offset algorithm is implemented in manchester decoding module;After input signal passes through filtering module and removes interference signal, signal waveform is converted into baseband signal waveform by signal waveform conversion module, then is sent into manchester decoding module, microprocessor in manchester decoding module starts reading signal by external interrupt, while reading signal, real-time baseband signal pulse width is calculated using timing offset algorithm, according to the calculation result, the timer initial value of the collected baseband signal is revalued, the dynamic adjustment of microprocessor to baseband signal acquisition frequency is realized, finally, according to manchester encoding format, the collected baseband signal is decoded to obtain the final decoding data.The present application improves decoding speed, accuracy and environmental adaptability.

Embodiments of the present application provide a decoding method based on an attention mechanism and a text generation method of a large language model. The decoding method based on the attention mechanism is applied to a first decoding layer of a model based on the attention mechanism. The model includes a plurality of decoding layers in cascade. The first decoding layer is any decoding layer except the first decoding layer among the plurality of decoding layers. The decoding method based on the attention mechanism includes: performing linear transformation on an input vector of the first decoding layer to obtain a first value vector; obtaining an attention matrix of a second decoding layer in the large language model, the second decoding layer being before the first decoding layer; and performing weighted summation on the first value vector according to the attention matrix of the second decoding layer to obtain an attention output of the first decoding layer. The method decodes by obtaining the attention matrix of the second decoding layer, saves the calculation process of obtaining the attention matrix based on the input vector of the first decoding layer, and improves the decoding speed based on the attention mechanism.

A kind of architecture design method of polar code decoder based on custom instruction set, according to custom instruction control the hardware architecture of decoder, utilize instruction to control decoding process, reduce the error probability caused by circuit state in decoding process, so that the work of polar code decoder is more stable;FSC decoding algorithm is used to classify the nodes of SSC decoding tree into multiple different types of sub-nodes, and according to different sub-nodes, corresponding fast decoding is completed, the problems of large decoding delay, low throughput and not easy to realize hardware circuit are solved, the decoding speed and throughput of the decoder are improved, the decoding efficiency of polar code is improved, so that the work of polar code decoder is more stable.

The application discloses a kind of layer mixing control file processing method, device, equipment and storage medium, the method includes: obtaining initial first picture set, first picture set includes the alpha value of multiple frames initial control picture for controlling target animation is respectively stored in different color channels of first control picture, to obtain second picture set.This scheme can effectively reduce the total size and quantity of control picture, reduce reading, decoding and texture generation pressure in animation playing process, so as to reduce the burden of processor / memory, improve animation fluency, reduce power consumption.And can be flexibly selected according to actual demand lossless compression mode or lossy compression mode, to balance effect and performance.

The invention is suitable for the technical field of large language models, and provides a key value cache compression method and electronic equipment. The key value cache compression method comprises the following steps: determining a behavior mode corresponding to each attention head in a trained multi-mode large language model in a decoding stage; according to the behavior mode corresponding to each attention head, determining a target cache budget corresponding to each attention head; and for each attention head, performing key value cache compression processing on the attention head according to the target cache budget and the behavior mode corresponding to the attention head. Through the key value cache compression provided by the invention, different key value cache budget determination methods can be adopted for attention heads of different behavior modes, and different key value cache compression processing methods can also be adopted for the attention heads of different behavior modes, so that the effect of performing key value cache compression on a multi-modal large language model is improved, and the user experience is improved. And thus, the performance of the multi-modal large language model is improved.