38 results about "Second quantization" patented technology

The invention provides a floating-point number quantization method and device, a floating-point number inverse quantization method and device, equipment and a storage medium, and belongs to the technical field of artificial intelligence. The method comprises: based on a compression rule, determining quantization information corresponding to an index of each floating-point number in floating-point numbers to be quantized, the quantization information corresponding to the index of each floating-point number comprising a first quantization value and / or a second quantization value, the compression rule comprises a rule that an index which indicates that quantization information with a low compression degree belongs to is greater than an index which indicates that quantization information with a high compression degree belongs to, and for each floating-point number in the floating-point numbers to be quantized, based on a first quantization value corresponding to the floating-point number, performing quantization processing on a mantissa of the floating-point number, and / or performing quantization processing on the mantissa of the floating-point number based on a second quantization value corresponding to the floating-point number. And quantizing the index of the floating-point number based on the second quantized value corresponding to the floating-point number. By adopting the method and the device, the required average error can be met during quantization.

A device for decoding a bitstream of encoded mesh data is configured to in response to determining that first transform parameters are overridden, infer a value of a second syntax element to be equal to a first value indicating that second transform parameters are present in the bitstream of encoded mesh data; in response to the value of the second syntax element being equal to the first value, determine the second transform parameters from a second syntax structure; in response to determining that the first quantization parameters are overridden, infer a value of a third syntax element to be equal to a first value indicating that second quantization parameters are present in the bitstream of encoded mesh data; and in response to the value of the third syntax element being equal to the first value, determine second quantization parameters from a third syntax structure.

The invention discloses a language task processing method and electronic equipment, and relates to the technical field of artificial intelligence. The method comprises the following steps: selecting a target weight parameter according to task calibration data and an activation function type of an original task processing model; according to each target weight parameter, the maximum weight parameter and the task calibration data, determining a fitting processing result of the quantization parameter; and determining a first quantization error function and a second quantization error function carrying a scaling factor according to a fitting processing result, and obtaining a numerical value of the scaling factor by minimizing a ratio of the first quantization error function and the second quantization error function. And performing scaling processing on each target weight parameter by using the scaling factor, and quantifying each scaling weight parameter based on the quantization precision parameter to obtain a task processing model for executing the to-be-processed language task. According to the method and the device, the problem that high accuracy of a language task processing result cannot be ensured on the basis of reducing resources used in a language task execution process in related technologies can be solved, and the language task processing precision is effectively improved.

An encoder, includes: memory; and circuitry coupled to the memory, in which in a first encoding process, the circuitry quantizes a current block using a first quantization matrix in both of a first case where the circuitry performs primary transform and the circuitry does not perform secondary transform and a second case where the circuitry performs both the primary transform and the secondary transform, in a second encoding process, the circuitry quantizes the current block using a second quantization matrix in a third case where the circuitry performs primary transform and the circuitry does not perform secondary transform, and in a third encoding process, the circuitry quantizes the current block without using a quantization matrix in both of a fourth case where both the primary transform and the secondary transform are skipped and a fifth case where both the primary transform and the secondary transform are performed.

Embodiments of the present application disclose a method and apparatus for determining a quantization parameter offset, a device, a storage medium and a program product. The method comprises: obtaining a first quantization parameter offset and a second quantization parameter offset of a coding unit, the first quantization parameter offset being calculated based on spatial information of the coding unit, and the second quantization parameter offset being calculated based on temporal information of the coding unit; determining representation information of the first quantization parameter offset, the representation information representing whether the first quantization parameter offset is used to improve reconstruction quality of the coding unit; determining a target quantization parameter offset of the coding unit according to the representation information of the first quantization parameter offset, the first quantization parameter offset and the second quantization parameter offset; and encoding the coding unit based on the target quantization parameter offset. The embodiments of the present application can effectively improve coding efficiency.

The invention relates to intra-tree geometric quantization of point clouds. An example apparatus includes one or more processors and a memory storing instructions that, when executed by the processors, cause the processors to receive points representing a point cloud in a three-dimensional space and generate a data structure representing the point cloud. Generating the data structure includes: encoding a location of each point in each dimension as a bit sequence according to a tree data structure; dividing each of the sequences into two or more portions according to a scaling depth; determining that a subset of the points is spatially isolated from the remainder of the points; quantizing each of the portions associated with the subset of the points according to a first quantization step size; quantizing each of the portions associated with the remaining portion of the points according to a second quantization step size; and including the quantized portion in the data structure.

The invention provides a quantization parameter storage method and device, a model reasoning method and device and a storage medium, and relates to the technical field of computers, in particular to the technical field of large models, the technical field of artificial intelligence and the technical field of model quantization. According to the specific implementation scheme, a calculation unit of a processor performs statistics based on reference data to obtain a statistical value of a first quantization parameter of a model; searching a target value of the first quantization parameter and a target value of a second quantization parameter of a model in a search space through the calculation unit based on the statistical value of the first quantization parameter; a target value of the first quantization parameter and a target value of the second quantization parameter are stored in a memory by the calculation unit. The quantization parameters are calculated and stored in advance, so that the quantization parameters which are calculated in advance can be read according to specific use requirements for quantization reasoning, occupation of a memory of a processor is reduced, calculation resources needing to be occupied in the reasoning process are reduced, and the reasoning speed and efficiency are improved.

The invention discloses a model quantification method and device, equipment, a storage medium and a computer program product, and relates to the field of model quantification, and the method comprises the steps: carrying out the training and quantification of a full-precision model disposed on edge equipment, and obtaining a first quantification model which comprises a first layer set; according to a sensitivity index and training cost corresponding to each model layer in the first layer set, a target training layer is selected from the first layer set, and the sensitivity index is used for representing the influence degree of the quantization operation on the model layer; performing quantitative perception training on the target training layer in the first quantitative model to obtain a second quantitative model, the second quantitative model including a second layer set; and performing bit width allocation on each model layer in the second layer set to obtain a target quantitative model. According to the invention, while the precision of the quantitative model is ensured, the cost required by quantitative training is reduced, so that the contradiction between insufficient quantitative precision after training and too high quantitative perception training cost in the prior art is solved.

The application discloses an AVS3 video encoding method, device and equipment and a computer storage medium, relates to the technical field of video processing, and obtains a target video to be encoded; a basic quantization parameter determined according to a set encoding mode is obtained; a first quantization scaling, a first quantization offset, a first quantization boundary and a first quantization step of the set encoding mode are obtained; a second quantization step of the AVS3 video encoding is obtained; conversion information between the second quantization step and the first quantization step is generated; the first quantization scaling, the first quantization offset and the first quantization boundary are respectively converted according to the conversion information, and a second quantization scaling, a second quantization offset and a second quantization boundary are obtained; the basic quantization parameter is adjusted based on the second quantization scaling, the second quantization offset and the second quantization boundary, and a target quantization parameter is obtained; and the target video is subjected to AVS3 video encoding based on the target quantization parameter, and an encoding result is obtained. The visual quality of the AVS3 video encoding is improved.

A device for decoding a bitstream of encoded mesh data is configured to in response to determining that first transform parameters are overridden, infer a value of a second syntax element to be equal to a first value indicating that second transform parameters are present in the bitstream of encoded mesh data; in response to the value of the second syntax element being equal to the first value, determine the second transform parameters from a second syntax structure; in response to determining that the first quantization parameters are overridden, infer a value of a third syntax element to be equal to a first value indicating that second quantization parameters are present in the bitstream of encoded mesh data; and in response to the value of the third syntax element being equal to the first value, determine second quantization parameters from a third syntax structure.

The invention provides a video processing method, receiving end equipment and transmitting end equipment, and relates to the technical field of video transmission. The video processing method comprises the following steps: in response to a condition that a first health degree score of a network at a current moment is less than or equal to a first threshold value, discarding at least part of video coding frames in a buffer area, and sending a request for obtaining an instant decoding refreshing frame to a sending end; in response to receiving an instant decoding refresh frame, decoding the instant decoding refresh frame, the instant decoding refresh frame being obtained by encoding a target region in the original image frame based on a first quantization parameter and encoding a background region in the original image frame based on a second quantization parameter, the first quantization parameter does not change along with the change of the first health degree score, and the second quantization parameter is negatively correlated with the first health degree score. According to the method and the device, the delay of the video picture can be reduced, and the information of the target area can be clearly reserved, so that the stability, accuracy and safety of remote operation can be improved.

The application discloses a multi-array geomagnetic underground pipeline detection system and method. The system comprises one or more geomagnetic sensor arrays, each array comprising a group of three-axis fluxgate sensors, an analog-to-digital converter (ADC), a geomagnetic background field cancellation signal generation circuit and an analog differential summation node. For a first time sequence segment, the ADC generates first digital geomagnetic time sequence data with a first quantization bit width; the cancellation signal generation circuit generates an analog cancellation signal according to prior background field data, and cancels a known strong magnetic interference component from subsequent analog signals at the analog differential summation node; then the ADC is dynamically adjusted to a second quantization bit width smaller than the first quantization bit width; and the equivalent magnetic moment vector data of a weak magnetic target body is determined by calculating a scaling factor corresponding to a least square error in combination with an amplitude reference benchmark in the first digital geomagnetic time sequence data.

The present disclosure relates to an LCEVC-based enhancement layer coding method, device, equipment and medium, applied to the technical field of video coding, which can improve the coding quality of video coding. The method comprises: in the case of using temporal prediction for a to-be-coded enhancement layer sequence, obtaining the code rate of the current frame base layer corresponding to the current frame enhancement layer and the average code rate of the C frame base layer before the current frame base layer, the to-be-coded enhancement layer sequence comprising a plurality of enhancement layers, and the current frame enhancement layer being any frame in the plurality of enhancement layers; in the case that the code rate of the current frame base layer is less than or equal to the average code rate, coding the current frame enhancement layer based on a first quantization parameter; in the case that the code rate of the current frame base layer is greater than the average code rate, updating the quantization parameter of the current frame enhancement layer from the first quantization parameter to a second quantization parameter, and coding the current frame enhancement layer based on the second quantization parameter, the second quantization parameter being less than the first quantization parameter.

The application provides a binarization network training method and device based on gradient quantization, which comprises the following steps: in a full-precision model full-quantization training process taking a sample image as a training sample and taking a class label of the sample image as a training label, quantizing weights and input activation values of the full-precision model to obtain first quantization data with a target bit width; in a case of outputting activation value gradients by performing back propagation on the first quantization data with the target bit width, quantizing the activation value gradients to obtain second quantization data with the target bit width; and updating the activation value gradients and weight gradients according to the first quantization data and the second quantization data to obtain a trained binarization network. The method realizes binarization gradient training, effectively reduces the training power consumption of the model and the requirements for memory and computing power, and further improves the training efficiency and inference performance of the classification model in a complex scene.

The invention relates to the technical field of computer systems, and provides a fine-grained quantization matrix multiplication device and method based on systolic arrays, and the device comprises a first systolic array which is used for receiving element streams of a first quantization factor matrix and a second quantization factor matrix, and carrying out the fusion calculation of quantization factors according to the element streams of the quantization factors, generating a quantization factor product; the second systolic array is used for receiving element streams of the first data matrix and the second data matrix, and performing multiplication and accumulation calculation according to the element streams of the matrixes; each first processing unit in the first systolic array is configured to send a quantization factor product obtained by calculation to a corresponding anchor processing unit in the second systolic array according to a mapping proportion; and the anchor processing unit is configured to receive the quantization factor product and serve as a propagation source point, the quantization factor product is propagated in the second systolic array in the row direction and the column direction within the quantization block size range corresponding to the quantization factor product, and the calculation efficiency is improved.

The application discloses a language task processing method and electronic equipment, and relates to the technical field of artificial intelligence. The method comprises the following steps: selecting target weight parameters according to task calibration data and an activation function type of an original task processing model; determining a fitting processing result of a quantization parameter according to each target weight parameter, a maximum weight parameter and the task calibration data; determining a first quantization error function and a second quantization error function carrying a scaling factor according to the fitting processing result, and obtaining a value of the scaling factor by minimizing the ratio of the two; performing scaling processing on each target weight parameter by using the scaling factor, and quantizing each scaling weight parameter based on a quantization precision parameter to obtain a task processing model used for executing a to-be-processed language task. The application can solve the problem that related technologies cannot guarantee high accuracy of a language task processing result on the basis of reducing resources used in a language task execution process, and effectively improve language task processing precision.

The present application provides an encoding device, a decoding device and a storage medium. The encoding device performs up-conversion and down-conversion on a first quantization matrix to generate a second quantization matrix, the first quantization matrix has a first number of rows and a first number of columns equal to the first number of rows to form a square matrix, the second quantization matrix has a second number of rows and a second number of columns different from the second number of rows to form a rectangular matrix; and, using the second quantization matrix to quantize the transform coefficients of the current block, wherein, in the up-conversion, the circuit generates the second quantization matrix by performing up-conversion in a first direction in such a way that one of the second number of rows and the second number of columns is greater than the first number of rows, the first direction is the horizontal direction; and, in the down-conversion, the circuit generates the second quantization matrix by performing down-conversion in a second direction in such a way that the other of the second number of rows and the second number of columns is less than the first number of rows, the second direction is different from the first direction, and the second direction is the vertical direction.

The invention provides a large language model quantization compression method based on limit quantization and multiple distillation, which comprises the following steps of: performing first quantization precision conversion: understanding a benchmark test set by using a natural language, taking a 32-bit BERT benchmark model as a teacher model, and performing quantization training to obtain an 8-bit intermediate model; a second quantization precision conversion step: understanding the benchmark test set by using a natural language, taking an 8-bit intermediate model as a teacher model, and performing quantization training to obtain a 1-bit target model; and configuring the target model in an edge server for natural language analysis processing of the resource-constrained edge environment. The method has the beneficial effects that the classic BERT model is taken as a target model, deep compression optimization is carried out on the BERT model, 32-bit storage bits are quantized and compressed to 1-bit storage bits, the size of the model is greatly reduced, meanwhile, a multiple distillation technology is used, and the practicability of the method in limited edge equipment is improved on the premise that the minimum loss of model performance is ensured.

To efficiently quantize rectangular blocks of various shapes.SOLUTION: The encoding device 100 includes a circuit 160 and a memory 162, wherein the circuit 160 generates a second quantization matrix for rectangular blocks by up-converting a first quantization matrix for square blocks only in a first direction and down-converting the first quantization matrix only in the first direction, quantizes a plurality of transform coefficients of a target block using the second quantization matrix, and encodes the plurality of quantized transform coefficients of the target block, and wherein the circuit 160 copies matrix elements of the first quantization matrix in the first direction in the up-conversion process; In the down-conversion process, the circuit 160 thins out the matrix elements of the first quantization matrix in the first direction and encodes only the information about the first quantization matrix out of the first quantization matrix and the second quantization matrix into a bit stream, where the first direction is the vertical direction.SELECTED DRAWING: Figure 31

A video decoding method according to the present invention includes: decoding a bitstream, to obtain P quantization coefficients of a current region; determining, according to the P quantization coefficients, parity of a quantization coefficient whose parity is hidden of a first quantization coefficient; determining a target context model corresponding to the first quantization coefficient, and decoding the first quantization coefficient which is encoded based on context using the target context model, to obtain a decoded first quantization coefficient; and determining the second quantization coefficient according to the parity of the quantization coefficient whose parity is hidden and the decoded first quantization coefficient.

The invention provides a method for determining a quantization strategy, a model quantization system and related equipment, and the method comprises the following steps: carrying out the free combination of a to-be-quantized object supported by a model and a data type which can be quantized, obtaining a plurality of different quantization strategies, determining the sensitivities of different quantization strategies based on a sensitivity algorithm, and carrying out the quantification of the to-be-quantized object. The sensitivity is used for indicating the influence of the quantization strategies on the precision of the to-be-quantized model, determining a first quantization strategy according to the sensitivity of each quantization strategy, and if the quantization evaluation of the first quantization strategy is qualified, using the first quantization strategy to quantize the model, so that the whole quantization process does not need quantization search; and if the quantization evaluation is unqualified, the precision of part of the to-be-quantized objects in the first quantization strategy is improved in a quantization rollback mode until the second quantization strategy which is evaluated to be qualified is obtained, so that compared with a traditional mode of performing quantization search based on large-scale parameters, the search space can be reduced, and the model quantization efficiency can be improved.

The application discloses a model quantization method and device, equipment, a storage medium and a program product, comprising: receiving quantization request information, the quantization request information comprising a first model and type information of a first user; determining first quantization configuration parameters required for quantizing the first model according to the type information of the first user, and sending a parameter configuration table corresponding to the first quantization configuration parameters; receiving second quantization configuration parameters selected from the parameter configuration table; and quantizing the first model according to the second quantization configuration parameters. By determining the parameters required for quantizing the first model according to the type information of the first user, and then quantizing the first model according to the quantization configuration parameters selected by the first user, different quantization configuration parameters can be provided for different users, so that the model quantized according to the quantization configuration parameters can better adapt to the specific needs of the users, and the model use efficiency is improved.

A first quantization value for encoding at least one frame of a content item may be determined based at least on a predetermined bitrate and a point in the content item associated with a scene change. A first duration associated with a first portion of the content item may be determined. The first portion of the content item may comprise the at least one frame and may be associated with the first quantization value. A second quantization value for encoding at least another frame of the content item may be determined based at least on the predetermined bitrate. A second duration associated with a second portion of the content item may be determined. The second portion of the content item may comprise the at least another frame and may be associated with the second quantization value.

The application relates to a speech coding method, a speech coding device, a computer device, a storage medium and a computer program product, and a speech decoding method, a speech decoding device, a computer device, a storage medium and a computer program product. The speech coding method comprises the following steps: performing signal sub-band decomposition based on an initial speech signal to be coded to obtain a first sub-band excitation signal and a second sub-band excitation signal; the frequency value corresponding to the second sub-band excitation signal is higher than the frequency value corresponding to the first sub-band excitation signal; performing quantization processing on the first sub-band excitation signal according to a first quantization precision to obtain a first excitation quantization signal, and performing quantization processing on the second sub-band excitation signal according to a second quantization precision to obtain a second excitation quantization signal; the first quantization precision is greater than the second quantization precision; and target coding data is obtained based on the first excitation quantization signal and the second excitation quantization signal. The method can improve coding efficiency.

The present disclosure belongs to the technical field of artificial intelligence. Provided are a floating-point number quantization method and apparatus, a floating-point number dequantization method and apparatus, and a device and a storage medium. The quantization method comprises: on the basis of a compression rule, determining quantization information corresponding to an exponent of each floating-point number among floating-point numbers to be quantized, wherein the quantization information corresponding to the exponent of each floating-point number comprises a first quantization value and / or a second quantization vale, and the compression rule comprises the rule whereby an exponent corresponding to quantization information which indicates a low compression level is greater than an exponent corresponding to quantization information which indicates a high compression level; and for each floating-point number among the floating-point numbers to be quantized, and on the basis of the first quantization value corresponding to the floating-point number, performing quantization processing on the mantissa of the floating-point number, and / or on the basis of the second quantization value corresponding to the floating-point number, performing quantization processing on the exponent of the floating-point number. By using the present disclosure, a required average error can be satisfied during quantization.

A method of processing video data includes: determining a quantization level of a coefficient for a current block from a plurality of quantization levels; determining an offset value based on the quantization level, where the offset value is a first offset value based on the quantization level being a first quantization level, or the offset value is a second, different offset value based on the quantization level being a second quantization level; determining a quantization parameter or an inverse quantization parameter for the coefficient based on the determined offset value; and performing one of quantization or inverse quantization on the coefficient based on the determined quantization parameter or the determined inverse quantization parameter as part of encoding or decoding the current block.

This invention relates to the field of signal modulation technology, proposing a noise-coupled Sturdy-MASH modulator and signal modulation method for converting analog input signals into digital output signals. The modulator comprises a first Delta-Sigma modulation module and a second Delta-Sigma modulation module cascaded sequentially. The first Delta-Sigma modulation module includes a quantizer Q1, and the second Delta-Sigma modulation module includes a quantizer Q2. The analog input signal is input to the first Delta-Sigma modulation module for modulation, and the quantizer Q1 generates a first quantization error and a first intermediate digital signal. The first quantization error is input to the second Delta-Sigma modulation module for modulation, and the quantizer Q2 generates a second quantization error and a second intermediate digital signal. The second quantization error is delayed by one unit before being fed to the input of the quantizer Q2 for subtraction, achieving noise coupling. This improves the first-order noise shaping capability without increasing the number of integrators, thereby enhancing the overall performance of the modulator while reducing circuit complexity.

The invention discloses a large model quantification method and device, computer equipment and a storage medium, which can be applied to various scenes such as financial investment management, automatic driving and face recognition, and the method comprises the following steps: obtaining a first large model; based on the type of the first large model, obtaining a first quantification method corresponding to the first large model; performing quantification processing on the first large model by using a first quantification method to obtain a second large model; and performing quantization processing on the second large model by using a second quantization method to obtain a target large model. According to the method, the first quantification method is utilized to quantify the first large model, the reasoning speed of the obtained second large model is improved by improving the quantification precision in the quantification process, and then the second quantification method is adopted to quantify the second large model, so that the occupation of a reasoning video memory is reduced, and the reasoning efficiency is improved. The reasoning speed of the obtained target large model is further improved, and the reasoning speed is improved while the small precision loss of the large model is kept.

A device for decoding a bitstream of encoded mesh data is configured to in response to determining that first transform parameters are overridden, infer a value of a second syntax element to be equal to a first value indicating that second transform parameters are present in the bitstream of encoded mesh data; in response to the value of the second syntax element being equal to the first value, determine the second transform parameters from a second syntax structure; in response to determining that the first quantization parameters are overridden, infer a value of a third syntax element to be equal to a first value indicating that second quantization parameters are present in the bitstream of encoded mesh data; and in response to the value of the third syntax element being equal to the first value, determine second quantization parameters from a third syntax structure.