56results about How to "Guaranteed Build Quality" patented technology

The invention provides a method and a device for generating a multiphase flow-containing solid rock three-dimensional model based on a diffusion model, and a medium, and can solve the problems that a multiphase pore and a rock skeleton structure cannot be generated at the same time and only a single-phase structure or a pure skeleton image can be reconstructed in the prior art. According to the method, the watershed algorithm is introduced to accurately segment the multiphase features, Gaussian noise is used as diffusion model input, and the model gradually learns' de-noising 'from the noise to obtain a real and complex multiphase rock image in training, so that a high-fidelity multiphase distributed rock internal image can be obtained only by inputting random noise in an inference stage.

The invention provides an image generation method and device, equipment, a medium and a program product. The method comprises the following steps: S101, carrying out autoregressive processing on a pre-filling token and a hidden space token with a plurality of resolutions through an autoregressive Transform model of an image generation network, and generating an input condition; wherein the hidden space token is obtained by encoding an original image by using an encoder; s102, carrying out up-sampling operation on the generated input condition; s103, inputting the up-sampled input condition into a diffusion model of the image generation network, performing diffusion denoising on the input condition through the diffusion model, and generating a hidden space token of the current resolution; s104, judging whether the current resolution reaches the target resolution or not; if yes, executing the step S105, and if not, returning to the step S101; and S105, restoring the hidden space token into an image through a decoder to obtain a target image.

The invention discloses a semi-fragile diffusion watermarking method and system based on an elliptic curve. The method comprises the following steps: constructing first-stage watermark information for tampering detection and second-stage watermark information for model traceability based on the elliptic curve; the method comprises the following steps of: performing first-stage watermark information embedding in a back stage of denoising through a diffusion process based on initial noise to obtain hidden space noise, introducing an extractor for extracting second-stage watermark information from an image generated by a fine-tuning decoder, and taking the second-stage watermark information as a label, embedding of two-stage watermark information is realized by pre-training an extractor and finely adjusting a decoder; the method comprises the following steps: extracting two-stage watermark information from a to-be-detected image by using a pre-trained extractor to realize model traceability, and detecting whether the to-be-detected image is tampered or not by comparing added noise estimated from hidden space representation inversion with the first-stage watermark information. Therefore, whether the image is maliciously tampered or not can be detected while the diffusion model corresponding to the image is traced.

The embodiment of the invention provides a three-dimensional model generation method and device, equipment, a medium and a program product. The method comprises the following steps: acquiring a first image and texture description information of a target object; based on a multi-view image generation model and the first image, generating a second image of the target object under each preset view and a normal graph corresponding to the second image; performing three-dimensional reconstruction based on the plurality of second images to obtain a first three-dimensional model of the target object; based on the normal graph, optimizing the first three-dimensional model to obtain an optimized second three-dimensional model; and performing texture mapping on the second three-dimensional model based on the first image and the texture description information to obtain a target three-dimensional model of the target object. Through the technical scheme of the embodiment of the invention, the generation convenience of the three-dimensional model can be improved, the generation cost of the three-dimensional model is reduced, and the generation quality of the three-dimensional model is ensured.

A vertical connection structure for fabricated prefabricated wallboards relates to the technical field of fabricated buildings and comprises a prefabricated concrete shear wall in up-down connection fit with a shear wall at a floor, a plurality of hollow holes are formed in the middle of the prefabricated concrete shear wall in a penetrating mode in the vertical direction, and a plurality of vertical steel bars are reserved in the middle of the shear wall at the floor. The hollow holes correspond to the vertical steel bars in a one-to-one mode, when the hollow holes and the vertical steel bars are connected and matched up and down, the vertical steel bars stretch into and are contained in the axis positions of the hollow holes, concrete at the positions of the hollow holes of the prefabricated concrete shear wall is poured through core grouting, and installation of the prefabricated concrete shear wall is completed. And the prefabricated concrete shear wall and the shear wall at the floor form a whole. According to the utility model, the vertical connection stress of the prefabricated wallboard is reliable, the standardization of the prefabricated wallboard is realized, the construction is convenient, and the construction cost is reduced.

The invention provides a large model cue word compression method based on dynamic budget allocation. The method aims at optimizing the cue word length of a large language model so as to improve the calculation efficiency of the model and reduce the calculation cost. Through cooperative work of the dynamic budget distributor and the vocabulary selector, the cue words can be compressed on the premise of ensuring semantic integrity. Firstly, a dynamic budget distributor dynamically distributes a vocabulary retention budget of each sentence based on the correlation between each sentence and a question; then, the vocabulary selector calculates the retention probability of each vocabulary through the strategy network, and key vocabularies are preferentially retained; and evaluating loyalty and similarity of compression prompts through a reward function module, and guiding a strategy gradient to update network parameters. According to the method, the performance of the large language model is remarkably improved, and the method is suitable for prompt optimization in scenes such as multi-round dialogues and long text analysis.

The invention provides an LLM model reasoning method based on a medical rule base and related equipment. The rule base comprises a plurality of inference rules used for assisting the LLM model in executing an inference task, and the inference rules are structured data composed of a plurality of fields; wherein the plurality of fields comprise a plurality of condition fields used for describing reasoning conditions, and target fields used for describing reasoning targets matched with the reasoning conditions; the condition field comprises an entity serving as a reasoning condition. The method comprises the steps of obtaining a target dialogue text input by a user, and extracting a text entity contained in the target dialogue text; searching at least one target inference rule related to the text entity from a plurality of inference rules contained in a rule base; and on the basis of the at least one target reasoning rule, constructing a cue word, inputting the cue word into an LLM model, and executing logical reasoning related to the reasoning task by the LLM model on the basis of the cue word to generate a response text corresponding to the target dialogue text.

The utility model belongs to the technical field of electrical wire pipe installation, and particularly relates to an electrical wire pipe connecting device for a deformation joint, which comprises the deformation joint between two floors, a cassette is arranged in each floor, the cassettes in the two floors are connected through a flexible wire pipe assembly, and the flexible wire pipe assembly comprises a first lock nut arranged on the cassette and a second lock nut arranged on the cassette. The inner wall of the first lock nut is connected with a flexible metal hose, and the first lock nut is connected with a copper core flexible wire through a hoop wire clamp. A second lock nut is arranged on the side wall, away from the deformation joint, of the hidden box, and a hidden wiring pipe is connected to the inner wall of the second lock nut and arranged in the floor. Through the arrangement of the flexible line pipe assembly, the deformation upper limit of a deformation joint using the device is improved, the application range is widened, the situation that an electrical line pipe is broken due to the fact that the electrical line pipe is short is reduced, and the building quality is guaranteed.

The embodiment of the invention provides a key phrase generation model training method and device and a key phrase generation method and device. The training method comprises the following steps: acquiring a training document and a plurality of key phrase true values corresponding to the training document; distributing a plurality of key phrase true values to a plurality of control codes based on an optimal transmission theory; inputting the training document into a generator to enable the generator to output a plurality of candidate key phrases; inputting the plurality of candidate key phrases into a selector to enable the selector to output a decision label which is used for indicating that the corresponding candidate key phrases are reserved or discarded; and constructing a first loss function corresponding to the generator and a second loss function corresponding to the selector, and adjusting and optimizing the generator and the selector according to the first loss function and the second loss function to obtain a target key phrase generation model. According to the technical scheme, the accuracy and recall rate of key phrase generation can be considered, and the key phrase generation quality is guaranteed.

The invention provides a large language model text traceability method based on virtual cue word embedding, and belongs to the technical field of artificial intelligence natural language processing and the technical field of digital content security, the method comprises the following steps: (1) constructing a chat input instruction text with a role format; (2) obtaining embedded vector input; (3) obtaining system cue word embedding representation; (4) obtaining a generated text with a user identity information watermark; and (5) recovering user identity information to realize traceability of the generated text. According to the method, on the premise that main body parameters of the large language model are not modified and model text generation quality and response efficiency are not affected, instant embedding of traceable identity information in the large language model generation process is achieved, meanwhile, robustness of embedded watermark information in a text transformation scene is guaranteed, and the method is suitable for large language model generation. Verifiable and reproducible identity identification information can be stably extracted from the generated text, and reliable technical support is provided for source tracing of the generated text.

The invention provides a design method of a river ecological garden landscape dike dam, which comprises the following steps of: 1, acquiring hydrological data of a river and surrounding environment information of the river, establishing a three-dimensional terrain model, and analyzing and determining flood frequency and flood peak flow parameters of the river; water conservancy projects, ecological protection and landscape design are organically combined, the river flood control function is guaranteed, meanwhile, a river ecological system is protected and restored to the maximum extent, the surrounding landscape quality is improved, and the requirements of people for the beautiful living environment are met; through accurate hydrological data collection and analysis, design parameters are optimized in combination with an advanced finite element analysis method and a genetic algorithm, so that the dike dam has higher stability and safety under adverse conditions of flood, earthquake and the like, and the disaster risk is reduced; in the construction process, strict quality detection and process supervision are conducted on all construction links, the construction quality of the dike dam is ensured, the service life of the dike dam is prolonged, and the later maintenance cost is reduced.

The invention discloses a 300-meter water depth jacket S420 high-strength steel welding construction method which is characterized in that preheating treatment is carried out in the processes of plate cutting and blanking, plate pre-bending, reel pipe closing, transition groove cutting and longitudinal seam welding, and during pre-bending, a multi-time slow pressing and pre-bending mode is adopted, so that mechanical indentations can be prevented from occurring in the pre-bending process; during pipe rolling, pipe rolling and opening closing are completed in a multi-rolling slow forming mode, and rolling'dead bending 'can be avoided; in the process of cutting and grinding the transition groove, the workpiece is synchronously heated, so that defects can be avoided in the grinding process, and the influence on subsequent welding is avoided; in the longitudinal seam welding process, the requirement for the preheating temperature is improved, only electric heating can be adopted during preheating, the welding process is strictly controlled, and after-heating hydrogen elimination treatment is conducted after welding, so that the hydrogen content is reduced, and cracks are avoided; the construction quality of the second sea foundation is better guaranteed, the potential safety hazard of offshore operation of the jacket is reduced, and it is guaranteed that the main structure meets the service requirement.

This invention discloses an attention acceleration method, apparatus, device, and medium based on a diffusion model. The method includes: acquiring the input feature set, current time step, and local window parameters corresponding to a target multi-head attention module in a target diffusion model; reshaping the shape of each input feature in the input feature set based on a preset tensor format to obtain a three-dimensional feature set; determining a feature recombination strategy corresponding to the three-dimensional feature set based on the current time step, and recombinating the three-dimensional feature set based on the feature recombination strategy to obtain a target feature result set corresponding to the target multi-head attention module; and performing local window attention calculation on the target feature result set based on the local window parameters and the feature recombination strategy to obtain the target attention result corresponding to the target multi-head attention module. Through the technical solution of this invention, efficient acceleration can be achieved while maintaining generation quality, improving the actual utilization rate of hardware.

The application discloses a behavior tree automatic generation method and device based on a large language model and electronic equipment, relates to the technical field of artificial intelligence and automatic control, and comprises the following steps: acquiring task description information, constructing a node variable context object, injecting a structured prompt word, and forming a resource constraint-aware generation context; according to a semantic vector of the task description, searching for high-quality cases that meet a quality threshold, screening dynamic few-sample examples through a selection algorithm, and embedding generation prompt words; calling a large language model deployed locally to generate a behavior tree, performing syntax checking and logic defect detection on the generation result through an evaluation operator, converting a detection failure result into structured feedback information, driving the model to perform minimum iteration refinement until the checking is passed or the maximum number of iterations is reached. The application solves the problems of low availability, poor stability and high deployment cost of the large language model in generating the behavior tree, and enables a lightweight large language model to achieve a generation success rate of about 90% on consumer-grade hardware.

The utility model discloses a high-strength thermal insulation concrete block, which comprises a brick body, and further comprises a reinforcing device arranged in the brick body, and a plurality of thermal insulation layers arranged in the brick body, the positioning devices are arranged on the two sides of the brick body; wherein the integral strength of the brick body can be improved by the reinforcing device, the positioning device can be used for positioning during building, heat insulation plates are mounted above and below the brick body, a cavity is processed in the brick body, and the reinforcing device comprises a plurality of first reinforcing strips which are mounted on the inner wall of the brick body. The utility model relates to the technical field of concrete blocks, in particular to a high-strength thermal-insulation concrete block, which can greatly improve the overall strength of a brick body through the matching of the brick body and a reinforcing device, further can greatly reduce the phenomenon that the brick body is damaged due to collision and the like during transportation, can ensure the subsequent use quality of the brick body, and can improve the production efficiency. And convenience is brought to the use of workers.

This invention provides a method for tracing the source of large language model text based on virtual prompt word embedding, belonging to the fields of artificial intelligence natural language processing technology and digital content security technology. The method includes: (1) constructing chat input command text with role format; (2) obtaining embedded vector input; (3) obtaining system prompt word embedding representation; (4) obtaining generated text with user identity information watermark; (5) restoring user identity information to achieve source tracing of generated text. This invention achieves real-time embedding of traceable identity information in the large language model generation process without modifying the main parameters of the large language model or affecting the model text generation quality and response efficiency. At the same time, it ensures the robustness of the embedded watermark information in text transformation scenarios, and can stably extract verifiable and reproducible identity information from the generated text, providing reliable technical support for tracing the source of generated text.

This application provides a railway communication transmission wireless cabinet and communication base station, including: a cabinet, transmission equipment, and wireless equipment. The transmission equipment and wireless equipment are located inside the cabinet, with the transmission equipment located on the side of the wireless equipment facing away from the bottom of the cabinet. The transmission equipment includes: a transmission power distribution unit and a transmission subframe, with the transmission power distribution unit located on the side of the transmission subframe facing away from the wireless equipment. The wireless equipment includes: a remote radio frequency unit and a baseband processing unit, with the remote radio frequency unit located on the side of the baseband processing unit facing the transmission equipment. Thus, by integrating the transmission equipment and wireless equipment into a single cabinet, the installation location of the equipment is further optimized, the space utilization within the cabinet is improved, the construction period is shortened, construction quality is guaranteed, and the floor space is reduced.

The invention provides an automatic welding method for low-temperature high-manganese steel, and belongs to the technical field of high-manganese steel welding. The automatic welding method comprises the following steps that (1) the low-temperature high-manganese steel to be in butt joint and a welding material are prepared; wherein the low-temperature high-manganese steel to be butted is not beveled, and the weld gap is 0-1mm; (2) assembling the low-temperature high-manganese steel to be butted on a platform, and fixing the low-temperature high-manganese steel to be butted by adopting a fixing clamp, so that a welding seam is 0-1mm; and (3) an automatic welding method is adopted for conducting single-pass welding on the low-temperature high manganese steel to be in butt joint, and single-face welding and double-face forming are achieved. Automatic welding of the low-temperature high-manganese steel is achieved through the single-face welding and double-face forming welding method, reverse buckling operation is reduced, the operation efficiency is improved, meanwhile, the welding quality of the low-temperature high-manganese steel is improved, and environmental pollution is reduced.

The embodiment of the invention provides a content generation method and related equipment, and the method comprises the steps: indicating the execution of a calling reasoning model through business indication information related to a content generation business, when de-noising processing is carried out on corresponding input content coding characteristics of the noisy data content under a time step tj in the T time steps under the constraint of the first data content, obtaining a cache type for cache multiplexing according to content analysis parameters indicated by a cache multiplexing strategy related to the time step tj; and if the cache type is an inter-step cache type, multiplexing a predicted noise feature corresponding to a time step ti having an inter-step multiplexing relationship with the time step tj, performing denoising processing on an input content coding feature corresponding to the time step tj, and performing content generation based on an output content coding feature corresponding to the time step tj obtained by the denoising processing, the second data content is obtained. Therefore, the model reasoning speed can be effectively improved, and waste of computing resources is avoided.

This invention relates to a dynamic retrieval enhancement generation method based on critical path entropy increase trends and document selection. Addressing the issues of retrieval response lag and false triggering caused by single-point threshold triggering in existing technologies, as well as noise interference and resource waste caused by direct splicing of Top-K retrieval documents, this invention creatively proposes a critical path entropy increase trend triggering mechanism. It optimizes single-point entropy monitoring of all tokens to critical path token entropy increase trend monitoring, achieving on-demand retrieval triggering by dynamically determining uncertainty increments, thereby suppressing false triggering caused by noise fluctuations in non-critical locations and alleviating retrieval response lag. Furthermore, this invention proposes a document selection model based on direct preference optimization of DPO, optimizing the coarse-grained injection of Top-K documents into a fine-grained filtering and injection mechanism. By filtering out irrelevant and redundant documents, it reduces the invalid occupation of the context window, thereby solving the problems of reduced generation quality and wasted computational resources caused by noise interference.

The utility model belongs to bridge floor construction technical field discloses a kind of support device suitable for steel concrete composite beam wing plate.The first support piece in support device is detachably arranged at the bottom of the second support piece, and the support assembly is low in dismounting difficulty and high in dismounting efficiency. The adjacent support assemblies in the same support structure are provided with elastic locks, and the first end of the elastic lock can abut against the inner wall surface of the first support piece when it is elastically deformed, and the second end can abut against the inner wall surface of the second support piece. A single elastic lock can lock the first support piece and the second support piece of the same support assembly, improving the connection strength of the first support piece and the second support piece. The adjacent support assemblies in a single support structure can be locked by the locking mechanism, and the adjacent support structures are connected to each other, ensuring the structural self-stability of the support device and improving the overall support effect of the support platform on the steel concrete composite beam wing plate, thereby ensuring the construction and construction quality of the steel concrete composite beam.

The present invention discloses a light, heat-insulating, waterproof and fire-proof composite wallboard and a production method thereof, which relates to the technical field of the building materials industry, and in particular to a composite building material board with functions such as lightness, heat-insulating, waterproof and fire-proof, which is prefabricated by an automated production line, and the production process technology of the board. The composite wallboard of the present invention includes an inner wallboard and an outer wallboard. The production process of the inner and outer wallboards includes a mold making section, an inner wallboard casting and molding section, an outer wallboard casting and molding section, a wallboard maintenance section, a robot handling and stacking section, and an automatic control and industrial computer management system. The entire production process of the present invention adopts industrial computer control and digital process management. The present invention has a series of advantages such as light weight, good heat-insulating effect, environmental protection, energy saving, easy construction, shortened construction period, low cost, etc., and solves a series of shortcomings of the existing wall such as easy cracking, large deformation, unstable quality, etc.

The utility model relates to strip board component foundation and upper structure integration arched house, be equipped with arched main part structure, arched main part structure is constituted by several strip board components, and each strip board component of constituting same arched main part structure is arranged in proper order along the arch, and the adjacent side of adjacent strip board components abuts each other, and the bottom of arched main part structure both sides is equipped with the foundation strip board component perpendicular to ground, and the strip board component is in cuboid shape, and the wedge-shaped gap between adjacent strip board components is filled with concrete mortar, and the foundation fixed groove for installing foundation strip board component is set up on the foundation or underground foundation structure, and the lower part of foundation strip board component is fixedly set in the corresponding foundation fixed groove, the house of the utility model simplifies structure, and the construction is convenient, and the construction cost is low, and the period is short.

The invention discloses a reinforced concrete laminated slab structure based on brackets and a manufacturing method thereof, and relates to the technical field of building construction. According to the technical scheme, the prefabricated plate supporting frame and the bracket structure frame are included, the prefabricated plate supporting frame is connected with the bracket structure frame, the prefabricated plate supporting frame comprises prefabricated plate stress ribs and prefabricated plate distribution ribs, the prefabricated plate stress ribs and the prefabricated plate distribution ribs are connected in a binding mode, and tie bars are installed on the prefabricated plate stress ribs; the two ends of the Z-shaped steel bars are parallel to the prefabricated plate stress bars, and the Z-shaped steel bars are fixed to the prefabricated plate stress bars through the haunched area stirrups and the haunched area distribution bars. The method has the advantages that the construction period is short, the cost is low, the quality is easy to guarantee, the strength is high, transportation and hoisting are convenient, typesetting is flexible, an inner frame supporting system does not need to be erected, turnover materials and civilized construction expenses are reduced, the requirement for saving energy is met, and the method is universally suitable for the field of building construction.

This invention discloses a modular construction method for prefabricated utility tunnel structures, comprising the following steps: S1, prefabricating several segments of reinforced concrete integrated utility tunnel columns and beams; S2, excavating a foundation pit to construct a foundation platform, forming several spaced-apart sockets on the platform; S3, hoisting the integrated utility tunnel column and beam structure as a whole, positioning and fixing the bottom of the integrated structure into the sockets; S4, repeating step S3 to complete the installation of the next integrated utility tunnel column and beam structure; S5, pouring concrete at the joints between adjacent integrated utility tunnel columns and beams. This invention can further reduce construction costs, shorten the construction period, and ensure construction progress while maintaining construction efficiency and quality, and can effectively control problems such as construction waste, site dampness, and dust pollution during construction.

This application relates to the field of building construction technology, and discloses a precast cap beam assembly system, control method, and control device, including: a mounting frame, an assembly base, a detection module, and a controller assembly. The mounting frame is detachably mounted on the bridge pier; the assembly base is movably connected to the mounting frame via a lateral adjustment mechanism and a longitudinal adjustment mechanism, and is used to support the precast cap beam; the detection module is mounted on the mounting frame and is used to detect the relative position between the connecting rod and the connecting hole; the controller assembly is connected to the detection module, the lateral adjustment mechanism, and the longitudinal adjustment mechanism. In this application, the assembly system can automatically calibrate and adjust the position of the precast cap beam, reducing construction difficulty, facilitating efficient assembly of the precast cap beam, improving construction efficiency, and ensuring the overall construction quality and safety performance of the bridge.

The invention relates to the technical field of vehicle fault diagnosis, and discloses a wheel bolt looseness detection method and device, a vehicle and a medium, and the method comprises the steps: collecting an inertial measurement unit signal, a wheel speed signal and a steering wheel hand torque signal of the vehicle; determining target detection data based on the inertial measurement unit signal, the wheel speed signal and the steering wheel hand torque signal; detecting whether a wheel bolt of the vehicle is loosened by using the target detection data to obtain a detection result; and controlling the vehicle to execute corresponding actions based on the detection result. According to the method, the existing sensor of the vehicle is directly used for collecting the signals, then the signals are fused with the multi-dimensional vehicle running state signals, feature extraction and combination processing are combined to form target detection data, the target detection data are used for wheel bolt looseness detection, full-process automation from signal collection, data processing, state detection to vehicle control action execution is achieved, and the detection efficiency is improved. And the detection accuracy and the driving safety are improved, so that the vehicle fault diagnosis requirements of low cost, high real-time performance and easy deployment are met.

The invention provides a dimension stone roadbed construction grading optimization method and device, and the method comprises the steps: building a dimension stone pile porosity statistical calculation model based on a fractal theory, determining a relation curve of fractal dimensions and grading of a dimension stone pile, and carrying out the indoor model compaction test under a plurality of grading corresponding to different fractal dimensions, the settlement rate meeting the compaction condition is determined, the change of theoretical porosity before and after compaction is determined in combination with a block stone pile porosity statistical calculation model, and a porosity correction coefficient is obtained; based on the porosity correction coefficient, setting a porosity range, under the condition that the fractal dimension is not greater than the critical fractal dimension, determining a first grading range of the block stone pile, and screening a grading range corresponding to a porosity index meeting the settlement rate index requirement in the first grading range as a target grading. According to the method, the contradiction between the compaction quality and the porosity requirement of the block stone roadbed construction is solved, and the block stone roadbed construction quality and the natural convection cold conduction efficiency are improved.

The present application relates to earthwork backfill engineering technical field, specifically relates to a kind of super large particle size wide gradation soil and stone mixture maximum dry density determination method, using large tonnage convex block roller compaction or dynamic compaction compaction, obtain the maximum dry density under different gradation conditions, and under different gradation conditions, different more than 20mm particle content, without technical treatment to the super-diameter particles in the original gradation of soil and stone mixture.In construction, with the measured gradation curve and more than 20mm particle content, compared with different gradation conditions, and under different gradation conditions, different more than 20mm particle content conditions, the larger of the two under the closer gradation condition and the closer gradation condition more than 20mm particle content is the maximum dry density.The maximum dry density determined by this method is closer to the actual construction condition, which can truly, objectively and accurately reflect the maximum dry density of soil and stone mixture, and ensure the construction quality of the project.

The present application relates to the cross field of artificial intelligence and computer architecture, in particular to an adaptive KV cache compression method, system and device for an audio-text multi-modal large model. The method comprises opening a full-precision hot table and a low-bit cold table in the video memory; writing the newly generated audio or text Token in the FP16 format into the hot table, and when the number of Tokens in the hot table reaches a preset threshold, gradually migrating the Tokens in the hot table to the cold table; identifying the modal type of the Token during the migration process, selecting a corresponding quantization grouping strategy according to the identified modal type, and performing INT2 compression storage to the cold table after online calculation of the quantization parameters of the Token to be migrated. Through the cooperation of the above steps, the present application realizes the improvement of the model inference throughput and the natural full-duplex human-computer interaction under the premise of ensuring the audio synthesis quality.