118results about How to "Improve tuning efficiency" patented technology

The invention discloses a user session recommendation method based on a deep neural network. The method comprises steps as follows: collecting shopping website log data, reconstructing commodity webpages browsed by user sessions, and generating a training set D', a cross validation set V' and a test set T'; extending all the sessions so as to generate a training set D, a cross validation set V and a test set T; setting neutral network parameters and generating a deep neutral network model; training a neutral network and performing parameter adjustment with an automatic parameter adjusting framework based on the genetic algorithm; calculating the correct rate of the model on the test set T; calculating a path compression ratio. With the adoption of the method, the recommendation accuracy can be significantly improved, the problem about difficulty in neutral network parameter adjustment is solved, the algorithm for automatically adjusting network related configuration is provided on the basis of the genetic algorithm, and the network parameter adjustment efficiency is greatly improved. The method is applicable to online shopping websites to recommend commodities which users may buy to the users while the online shopping websites interact with the users.

The invention relates to an aim detecting structure, which comprises an aim picture on the basic plate, and a frame with known width on the periphery of said picture. The invention can be used in COG (Chip-On-Glass) technique, that with the aim mark of chip, after COG pressing, judging the deflection amount via the relative position of aim mark and the frame; therefore, the invention can directly calculate the deflection after compressing chip to adjust the machine after COG process.

The invention discloses an urban traffic signal optimization method and an urban traffic signal optimization system. The method comprises the following steps of judging whether a current urban fragmented region is reasonable or not according to the signal period duration, the spatial position and the traffic flow of an intersection on a trunk; judging whether an urban fragmented region control strategy is reasonable or not when the current urban fragmented region is judged to be reasonable; if the urban fragmented region control strategy is judged to be reasonable, determining whether an optimized space exists in the coordination of road sections or not; if no optimized space exists in the coordination of road sections, judging whether the intersection needs to be optimized or not; if theintersection needs to be optimized, carrying out the signal control, the traffic organization and the traffic order optimization on the intersection. In this way, the traffic signal optimization is carried out on the whole city level. The traffic jam can be relieved. Meanwhile, the signal optimization can be performed specifically. The optimization efficiency of the optimal technical personnel isimproved, and the ordered development of signal optimization work is guaranteed.

The present invention discloses a method of adaptively switching sound effect output channels and a terminal, and belongs to the terminal technology field. The method is applied to the sound effect processing of the terminal, the terminal comprises a sound effect output channel adaptive switching module, and the method of adaptively switching the sound effect output channels by the sound effect output channel adaptive switching module comprises the steps of removing a first sound effect output channel which is a current sound effect output channel; obtaining a to-be-switched second sound effect output channel; switching the sound effect output channel from the first sound effect output channel to the second sound effect output channel; playing the sound source data via the second sound effect output channel. According to the method of adaptively switching the sound effect output channels and the terminal of the present invention, the purpose sound effect output channel can be switched to real-timely and accurately, the sound effect real-time response processing is guaranteed, and the tone tuning efficiency and the auditory experience of the terminal are improved.

The invention provides a self-calibration control device of a warp knitting machine yarn guide bar and belongs to the technical field of mechanical-electrical integration. The problem that needling is required to be frequently calibrated by human eyes in the warp knitting machine yarn guide bar which is electronically and transversely driven is solved. The needling camera time of the yarn guide bar is determined by a coder (2) which is arranged on the main shaft of a warp knitting machine and is interconnected with a terminal control unit (1); relative position images of a warp knitting machine yarn guide needle and a slot needle positioned in the same plane at the needling time are acquired through a high-definition camera (3); after being processed by an image algorithm processing module (4), a calibration instruction of the yarn guide bar is given by a wireless data transmission transmitting module (5), and a wireless data transmission receiving module (6) which can be conveniently plugged into a warp knitting main control system receives the instruction and transmits the instruction to the main control system to calibrate the warp knitting machine yarn guide bar. Automatic calibration control on the yarn guide bar in two states, i.e. a halt state and a start state, can be realized without labor participation, and the control precision and the intelligence degree of a warp knitting machine numerical control system can be obviously improved.

The invention relates to a hyper-parameter threshold range determination method and device, a storage medium and electronic equipment. The method comprises the steps of obtaining a target hyper-parameter value used in each training process as a first target hyper-parameter value and obtaining an evaluation index value of a model obtained after each training as a first evaluation index value in a process of carrying out multiple training on the model based on a first training data set; and determining a threshold range of the target hyper-parameter according to the first target hyper-parametervalue and the first evaluation index value. Due to the fact that the determination of the threshold range of the target hyper-parameter refers to the corresponding evaluation index, the appropriate target hyper-parameter threshold range can be selected according to different requirements of users. Thus, hyper-parameter tuning can be carried out subsequently according to the target hyper-parameterthreshold range, and the range of repeated tuning attempts is narrowed, so that the efficiency of hyper-parameter tuning is improved, the labor cost is reduced, and the efficiency of artificial intelligence model training is improved.

A steering control method and apparatus of a motor-driven power steering system, ma include setting, by a controller, a virtual steering model including a reaction force apparatus provided between a steering wheel and a rack gear; inducing, by the controller, a state equation for the virtual steering model, the state equation representing momentum of the steering wheel, the reaction force apparatus, and the rack gear as state variables of the state equation; determining, by the controller, target steering torque acting in the reaction force apparatus through numerical integration of the state equation; and feedback-controlling, by the controller, a steering motor control amount to the motor-driven power steering system to bring a steering torque of the motor-driven power steering system into agreement with the target steering torque.

The invention provides a novel multifunctional mobile phone middle frame polishing-grinding machine. A Y-axis movable platform is provided with a water spraying device and five sets of clamping devices, the water spraying device comprises a draining groove, a water spraying pipe and a draining pipe which are mounted on a Y-axis movable platform, a U-shaped arm is provided with six sets of polishing wheel sets, two third motors and a rotating shaft, and the six sets of polishing wheel sets are evenly mounted on the U-shaped arm through mounting seats; and the polishing wheel sets comprise polishing wheels and protective covers arranged on the polishing wheels, the rotating shaft is provided with driving gears, each mounting seat is provided with a driven gear, the driving gears are connected with the driven gears through belts, the polishing wheels are mounted on the two sides of the driven gears correspondingly, and the third motors can drive the six sets of polishing wheel sets to dosynchronous polishing movement. The novel multifunctional mobile phone middle frame polishing-grinding machine has the advantages that dry polishing and waxing and polishing can be conducted, meanwhile, wet polishing can also be conducted, polished products are good in consistency and high in machine adjusting efficiency, the quality is ensured, the procedure is easy to control, the structure is simple, and wax adding is even.

The invention discloses an optical calibrator for the multi-barrel axis parallelism of artillery, and relates to the technical field of calibration for multi-barrel artillery. The optical calibrator comprises bullet-shaped axial lasers, a calibration assembly, a measurement assembly and a total reflection prism group, wherein the bullet-shaped axial lasers are respectively installed in chambers ofa reference barrel G0 and a measured barrel G1 of the artillery and emit laser beams which are overlapped with barrel axes of the artillery, the total reflection prism group and corresponding prismsin the calibration assembly and the measurement assembly to enable quadrangular optical paths to be coplanar, the bullet-shaped axial laser L2 emits a laser beam, the laser beam is reflected to a reticle of a focal plane of an angle measurement telescope w2 by a semi-transmitting and semi-reflective prism F to form a light spot, the non-parallelism deviation between the reference barrel and the measured barrel can be judged through comparing the overlap ratio of the light spot with a light spot, which is reflected to the angle measurement telescope w2 by an optical path, from the axial laser L1, the axis of the measured barrel is adjusted until the two light spots are overlapped, and thus the calibration for the parallelism of the two barrels is completed.

A steering control method and apparatus of a motor-driven power steering system, ma include setting, by a controller, a virtual steering model including a reaction force apparatus provided between a steering wheel and a rack gear; inducing, by the controller, a state equation for the virtual steering model, the state equation representing momentum of the steering wheel, the reaction force apparatus, and the rack gear as state variables of the state equation; determining, by the controller, target steering torque acting in the reaction force apparatus through numerical integration of the state equation; and feedback-controlling, by the controller, a steering motor control amount to the motor-driven power steering system to bring a steering torque of the motor-driven power steering system into agreement with the target steering torque.

The invention discloses an electric energy meter correction table and an electric energy meter correction method. The electric energy meter correction table comprises a PC (personal computer), an electric energy meter power supply circuit, a control unit, a surface plate, two guiders, connecting terminal seats and electric energy meter clamping devices, wherein the connecting terminal seats and the electric energy meter clamping devices are mounted on the surface plate. The connecting terminal seats are disposed on the surface plate and below meter positions of electric energy meters to be measured. Each connecting terminal seat comprises a plurality of power source terminal vibrating needles connected with the electric energy meter power supply circuit and a plurality of communication terminal vibrating needles electrically connected with a communication connector of the PC. Each electric energy meter clamping device comprises a press block arranged above the meter positions and driven by an air cylinder. The guiders are arranged on the surface plate and on two sides of the meter positions. The electric energy meter correction table is disposed on a work station of an electric energy meter production line and between an initial assembling work station and a packaging work station of the production line. The electric energy meter correction table is high in correction efficiency, capable of directly setting and correcting electric energy meters on and along with the production line, production efficiency of electric energy meter correction is improved and production cost is reduced.

The invention discloses a method for the computer aided design of a microwave filter. The method comprises the following steps of: 1) determining the size of a resonator according to the center frequency of the filter; 2) determining a group of standard delay amounts Gamma1, Gamma2, ..., GammaN according to the bandwidth and order N of the filter; 3) simulating a single-port network consisting ofan input port and a first-order resonator, and regulating the physical size of the first-order resonator to make the port delay characteristic of the network meets Gamma1; 4) adding a second-order resonator, and regulating the physical sizes of the first-order and second-order resonators to make the port delay of the network meets Gamma2; 5) by parity of reasoning, obtaining the physical sizes ofa half of resonators; and 6) mirroring the resonators with the regulated physical sizes to obtain the sizes of the other half of resonators to finish the design of the filter. In the method, the delay characteristic of the single-port filter is utilized, and the method has the advantages of small calculated amount and high efficiency as well as high center frequency and bandwidth setting accuracy, and is particularly suitable for the computer aided design (CAD) of the high-order microwave filter.

The invention provides a semi-automatic chamfering machine. The semi-automatic chamfering machine comprises a base frame, a bottom plate, locating plates, locking pliers, a work table, micrometers, a motor, a movable sliding plate and an air cylinder, wherein the bottom plate is arranged on the base frame; a sliding rail fixing plate is installed on the bottom plate, a second sliding rail is arranged on the sliding rail fixing plate, a plurality of second sliding blocks are arranged on the second sliding rail, the movable sliding plate is connected to the second sliding rail through the second sliding blocks, a first sliding rail is arranged on the movable sliding plate, a plurality of first sliding blocks are arranged on the first sliding rail, and the air cylinder is arranged on one side of the first sliding rail; the motor is arranged on the left side of the base frame and is connected to the movable sliding plate through a pull rod; the work table is arranged on the first sliding blocks, the front side of the work table is connected to a micrometer fixing frame, and the micrometer fixing frame is provided with the micrometers; and a PLC is arranged inside the base frame. Compared with the prior art, the semi-automatic chamfering machine has the beneficial effects that propulsive assistance of the air cylinder replaces manual thrust, and the micrometers limit a jig so that accuracy control can be realized.

The embodiment of the invention discloses a hyper-parameter determination method and device, computer equipment and a storage medium. According to the embodiment of the invention, the method comprisesthe steps: obtaining a hyper-parameter group of at least one to-be-trained model, wherein the to-be-trained model belongs to a target model type, the hyper-parameter group comprises hyper-parametersof at least one dimension, and the hyper-parameters of each dimension indicate model structure features of the to-be-trained model under each dimension; distributing the target hyper-parameter group to a target hyper-parameter evaluation node in the distributed cluster; through the target hyper-parameter evaluation node, adopting the target hyper-parameter set to perform model training on a targetto-be-trained model, wherein the target to-be-trained model is a to-be-trained model corresponding to the target hyper-parameter set; performing model evaluation on the trained model to obtain a hyper-parameter evaluation result of each hyper-parameter group; based on the hyper-parameter evaluation result, selecting an optimal hyper-parameter group of the target model type from the hyper-parameter groups. According to the scheme, the hyper-parameters can be efficiently determined by improving the hyper-parameter tuning efficiency.

The embodiment of the invention discloses a model training method and device, an electronic device and a readable storage medium. The method comprises the steps of obtaining the first training data and the second training data, training a plurality of basic models based on the first training data, determining the model parameters of each basic model, and determining a basic model used in a combined model and the corresponding combination coefficients of the used basic model through a greedy algorithm based on the second training data. According to the technical scheme, the base model used in the combined model and the corresponding combination coefficient of the used base model can be automatically determined, the parameter adjusting efficiency in the model training process can be improved, and the accuracy and the objectivity of the model are improved.

The invention provides a global occlusion self-adaptive pedestrian training / recognition method, system and device and a medium. The pedestrian training method comprises the steps of receiving a training data set; extracting an attribute feature map of each picture; fusing all attribute feature maps of same pedestrians to obtain fused features of a plurality of pedestrians; extracting local features from the plurality of attribute feature maps and extracting global features from the fusion features of the plurality of pedestrians; extracting attention of the local features according to the local features and the global features so as to calculate features after attention enhancement of the local features used for representing the local features and features after shielding of the picture shielding information of the global features used for representing the global features; and performing loss calculation on the features after the attention of the local features is enhanced and the features after the picture shielding information of the global features is shielded. According to the invention, the model error caused by manual operation deviation is reduced; the problem that common model evaluation indexes cannot guide engineering threshold setting is solved.

The invention provides an axial fine adjustment and tensioning shaft, and belongs to the technical field of testing tools. The axial fine adjustment and tensioning shaft comprises a tensioning assembly, an axial fine adjustment assembly and a motor transmission shaft. The tensioning assembly comprises a tensioning hand wheel, a tensioning mandrel, a guide taper sleeve, a telescopic taper sleeve, a loosening cylinder, a tensioning mandrel front supporting seat and a tensioning mandrel rear supporting seat. The axial fine adjustment assembly comprises an axial fine adjustment hand wheel and a fine adjustment mandrel. The tensioning assembly is connected with the axial fine adjustment assembly. The fine adjustment assembly is connected with the motor transmission shaft. The tensioning mandrel and the fine adjustment mandrel are coaxially sleeved together and can synchronously generate axial displacement, the fine adjustment mandrel is connected with the axial fine adjustment hand wheel, the tensioning hand wheel is connected with the tensioning mandrel, and the loosening cylinder is connected with a spring pull ring arranged on the tensioning mandrel in a sleeving manner. The outer portion of the tensioning mandrel is sleeved with the guide taper sleeve, a loosening movable plate is arranged on the outer portion of the loosening barrel and is connected with the spring pull ring, and the telescopic taper sleeve is connected with the loosening cylinder. One end of the fine adjustment mandrel penetrates into and is connected with the motor transmission shaft.

The invention discloses a camera parameter adjustment method and a device thereof, electronic equipment and a storage medium, and relates to the technical field of computer vision, and the method comprises the steps: detecting a first image collected by a reference camera, and obtaining a first detection result; detecting a second image collected by the camera to be subjected to parameter adjustment, obtaining a second detection result, and shooting the first image and the second image in the same preset scene; and according to the first detection result and the second detection result, adjusting a first parameter of the camera to be subjected to parameter adjustment when collecting the second image to obtain a second parameter of the camera to be subjected to parameter adjustment in the preset scene. According to the method and the device, the electronic equipment and the storage medium disclosed by the invention, automatic parameter adjustment is performed on the camera to be subjected to parameter adjustment based on the result of visual detection on the image collected by the reference camera, so that the visual detection result corresponding to the image collected by the camera to be subjected to parameter adjustment can reach the best without manual participation, Therefore, the parameter adjustment efficiency of the camera is improved.

A level bar with a hinting function comprises a bar body; at least one level bubble device is disposed on the bar body; at least one hollow part is disposed in the bar body; an induction device, a controller, and a sounding device; the controller is respectively connected with the induction device and the sounding device; the induction device comprises an independent sensor for detecting an angle of the bar body with the level line; the independent sensor does not contact the level bubble device; the controller receives the information of the angle of the bar body with the level line transmitted by the independent sensor; if the angle information is within a preset range, the sounding device is triggered to generate a sound. The invention not only can determine whether the angle of a surface to be detected with the level line is a preset angle by conventional level bubble centering, but also provides another set of independent determination tools, wherein the independent sensor is set to detect the angle of the bar body with the level line, and the controller determines whether the angle is within a preset range; the invention is convenient for use and high in precision.

The invention provides a galvanometer motor comprising a magnetic conductive machine shell, a coil assembly, a rotor assembly and a grating encoder, and one end of the magnetic conductive machine shell is provided with a unique complete installation face; the coil assembly comprises a wire and a framework, the framework is fixed in the magnetic conductive machine shell, and the wire is wound on the framework to form a coil; the rotor assembly penetrates through the framework in the height direction of the magnetic conductive machine shell and can swing relative to the framework. The grating encoder comprises a coded disc and a receiver, and the coded disc and the receiver are both located on the mounting surface. According to the technical scheme, the installation convenience of the grating encoder is improved.

The invention discloses an error correction method and system for arc corner machining equipment. The arc corner equipment comprises a milling cutter, and the error correction method of the arc cornermachining equipment comprises the steps: setting initial operation parameters according to a preset machining program, wherein the initial operation parameters comprise clamping parameters, size inspection standards, the cutting direction of the cutting cutter and the correction direction; acquiring predetermined detection data in product processing, wherein the predetermined detection data comprises the actual milling diameter of the milling cutter; analyzing the detection data, and calculating correction parameters according to a predetermined correction model; distributing the correction parameters to corresponding processing equipment; correcting the machining parameters of the corresponding machining equipment according to the correction parameters; according to the machining direction and the correction direction of the machining equipment and the correction parameters, judging the abnormality of the machining equipment, so the time and labor required for machine adjustment of the machining equipment are reduced, and the machine adjustment efficiency is improved.

The invention discloses a head-up display adjustment and calibration method, and a special apparatus for applying the method. The head-up display adjustment and calibration method comprises: the step 1: fixing a datum lens before a head-up display device which does not change a CRT kinescope, wherein the datum lens is a telescope with a reticle or a borescope and a test frame aligns to the datum lens when the head-up display device is energized and displays the test frame; the step 2: deenergizing the head-up display device, and dismounting and changing the CRT kinescope in the head-up display device; and the step 3: energizing the head-up display which changes the CRT kinescope, and adjusting the position of the CRT kinescope until that the datum lens aligns to the test frame. The head-up display adjustment and calibration method takes the telescope or the borescope as the external reference to dismount and change the CRT kinescope in the head-up display device, that is, before dismounting, the test frame displayed on the glass through the old CRT kinescope is taken as the datum; and after a new CRT kinescope is exchanged and the reticle of the datum lens is the datum target position so that the difficulty in precise adjustment and calibration of the character precision of the head-up display is reduced.

A machining equipment error correction method based on a single machine is applied to a machining equipment error correction system based on the single machine; the machining equipment error correction system based on the single machine comprises a computer, and the computer comprises a storage medium and a processor; the machining equipment error correction method based on the single machine comprises the steps that the processor sets initial operation parameters according to a preset machining program, wherein the initial operation parameters comprise clamping parameters and standard inspection parameters; the processor acquires predetermined detection data in product processing, analyzes the detection data, calculates correction parameters according to a predetermined correction model,distributes the correction parameters to corresponding processing equipment, corrects the machining parameters of the corresponding machining equipment according to the correction parameters, and stores the detection data and the correction parameters in the storage medium, so that the time and labor required for machine adjustment of the machining equipment are reduced, and the machine adjustmentefficiency is improved.

The invention provides a screen printing plate alignment method for an elliptical printing machine and an alignment system. The alignment system is arranged on the elliptical printing machine and involves screen printing plates, visual assemblies, a miniature industrial personal computer, a display, a position adjusting assembly, bedplates and a cross beam. The alignment system in the scheme adopts a modular design, is compact in structure, saves space, is small in size, and can be directly arranged on the basis of an original elliptical machine; due to the fact that the screen printing plate position alignment system operates together with the other bedplates, each elliptical printing machine only needs to be provided with one screen printing plate position alignment system (except the screen printing plates), and the manufacturing cost of the whole device is greatly reduced; and device data parameters are completely quantified, so that common employees can be qualified for operation through simple training, and requirements on operators are reduced.

The invention discloses a testing board for adjusting and calibrating the position of a probe, and the testing board includes at least one test point for testing the probe, wherein the test point includes a plurality of test areas having conductive functions, and all test areas are connected with resistors with different resistances. The plurality of test areas comprise the first test area and a plurality of second test areas, wherein the plurality of second test areas are disposed around the first test area to surround the first test area. The test point further includes a first insulation area and a plurality of second insulation areas. The first test area is spaced apart from the plurality of second test areas by the first insulation region, and each two of the plurality of second testareas are spaced apart by one second insulation region. The invention also relates to a testing method for adjusting and calibrating the position of the probe. The board and method are simple in implementation, and can effectively improve the adjustment and calibration efficiency of the probe.

The invention provides a piano rhythm calibration apparatus based on a chord vibration signal and an application method, which can increase the toning efficiency of a piano and improve the quality ofthe piano. The piano rhythm calibration apparatus comprises two calibration supports and a signal acquisition apparatus, wherein a ball screw is arranged between the two calibration supports, the ballscrew is provided with a nut pair I and a nut pair II, a movable supporting plate I is arranged outside the nut pair I, a movable supporting plate II is arranged outside the nut pair II, the movablesupporting plate I is provided with a lower pressing plate, the movable supporting plate II is provided with a traction plate, the lower pressing plate is connected with the traction plate by virtue of a traction wire, and the lower part of the lower pressing plate is provided with a lower pressing block. The piano rhythm calibration apparatus is provided with the lower pressing block, piano keysare pressed by the lower pressing block to vibrate chord, and the lower pressing block is returned after pressing the piano keys, so that the consistence of the pressing force at every time can be ensured.

The automatic calibration device in the positive branch confocal unstable cavity mirror includes: a He-Ne laser, a folded mirror with a small hole structure, a scraper mirror with a small hole, a convex mirror, and a detector that detects the output beam quality of the laser. It consists of a Terman wavefront sensor, a main controller and a concave mirror with a stepping motor actuator. The structure of the invention is simple and stable, and the processing technology is easy to realize. By adjusting the concave mirror of the main control mirror, it can be corrected inside the laser resonator Compared with the existing laser intracavity automatic alignment technology, the high-energy laser beam tilt, jitter and defocus error can simplify the installation and adjustment of the control system and the control algorithm is easy to implement because it captures the main factors that affect the beam quality.

The invention relates to a data processing method and device, computer equipment and a storage medium and relates to the technical field of computers, according to the embodiment of the invention, thesample data can be automatically acquired from the database; the parameter adjustment step and the model training step are executed, and the trained model can be used for executing the data processing step, so that on one hand, the automation of the data processing process is realized, the labor cost is reduced, the data processing efficiency is improved, personal errors are avoided, and the dataprocessing accuracy is improved; another aspect, exploring through an initial value, an exploring direction, a step length and the like of the hyper-parameter; determining and obtaining a target candidate value in an explored target range, and taking the target candidate value as an initial value to continue exploring; the method is advantaged in that the hyper-parameter exploration range is expanded, the process of searching the optimal hyper-parameter exploration direction is accelerated, a condition that the optimal hyper-parameter cannot be found because of improper exploration boundary setting is avoided, parameter adjustment accuracy and efficiency are improved, and integral data processing accuracy and efficiency are improved.