334 results about "First pass" patented technology

A system may, in a first pass: divide content to be summarized into a plurality of chunks and, for each chunk: execute a language model with the chunk and an instruction to summarize the chunk, generate, based on the executed language model, a summary of the chunk. In a subsequent pass, the system may: generate a plurality of groups of summaries, each group of summaries from among the plurality of groups of summaries comprising two or more summaries, each summary corresponding to a respective chunk, for each group of summaries from among the plurality of groups: execute a language model with the group of summaries and an instruction to summarize the group of summaries, generate, based on the executed language model on the group of summaries, a group summary. The system may iteratively repeat the subsequent pass for group summaries until a summary of the content is reached.

A method of operating a memory device includes applying a first read voltage to a first word line WLn corresponding to target memory cells, applying a first pass voltage to a second word line WLn−1 corresponding to memory cells adjacent to the target memory cells, and applying a second pass voltage to a third word line WLn+1 corresponding to memory cells adjacent to the target memory cells. The second pass voltage is higher than the first pass voltage.

A system may, in a first pass: divide content to be summarized into a plurality of chunks and, for each chunk: execute a language model with the chunk and an instruction to summarize the chunk, generate, based on the executed language model, a summary of the chunk. In a subsequent pass, the system may: generate a plurality of groups of summaries, each group of summaries from among the plurality of groups of summaries comprising two or more summaries, each summary corresponding to a respective chunk, for each group of summaries from among the plurality of groups: execute a language model with the group of summaries and an instruction to summarize the group of summaries, generate, based on the executed language model on the group of summaries, a group summary. The system may iteratively repeat the subsequent pass for group summaries until a summary of the content is reached.

A multiplexer includes first and second filters respectively with a first pass band and a second pass band on a higher frequency side than the first pass band. The first filter includes series-arm resonators and parallel-arm resonators, a resonance band width of at least one of the series-arm resonators is wider than the first pass band, one of the series-arm resonators is coupled closest to a common terminal among the series-arm resonators and the parallel-arm resonators. An anti-resonant frequency of the one of the series-arm resonators is lower than a high frequency end of the second pass band and is lowest among anti-resonant frequencies of the series-arm resonators.

Automatically searching for cryptocurrency artifacts to recover cryptocurrency assets. Digital data suspected to contain unknown cryptocurrency artifacts may be searched to detect candidate seed phrases, each comprising a permutation of words from a single predetermined seed phrase library consecutively listed in a sequence having a fixed length predetermined according to a cryptocurrency standard. A multi-pass seed phrase validation test may be executed for each candidate seed phrase to determine if it is a valid seed phrase by: executing a first pass comprising verifying a checksum of the candidate seed phrase, and upon verifying the first pass, executing a second pass comprising verifying in a blockchain if there has been activity using a cryptocurrency asset encrypted using a public key in a key pair recovered from the candidate seed phrase. Upon verifying the valid seed phrase, cryptocurrency assets may be recovered using a private key in the key pair.

The invention discloses a camouflage target detection method and device based on feature enhancement, and a readable medium, and the method comprises the steps: obtaining a to-be-detected camouflage target image, inputting the to-be-detected camouflage target image into a trained camouflage target detection model, obtaining multi-scale features through a multi-scale feature extraction module in an encoder, and carrying out the detection of the camouflage target image; wherein the second-level feature, the third-level feature and the fourth-level feature respectively pass through three feature enhancement modules and a context aggregation module to obtain a second aggregation feature, a third aggregation feature and a fourth aggregation feature; performing up-sampling on the second aggregation feature, performing element-by-element addition on the second aggregation feature and the first-level feature to obtain a first addition result, and processing the first addition result through a first convolution module to obtain a first mask; the first mask, the second aggregation feature, the third aggregation feature and the fourth aggregation feature are input into a feature decoupling module to obtain a second background suppression feature; and obtaining a corresponding prediction detection result based on the second background suppression feature, and solving the problem of high similarity of the foreground and the background.

Electronic components and methods. The electronic components include a first device die and a first dummy die located in the first device level and having at least one first pass through conductive layer through the first dummy die. Also included is a second device die and a second dummy die located in a second device level above the first device level, the second device die attached to the first device die and the second dummy die attached to the first dummy die and having at least one second pass through conductive layer electrically connected to the at least one first pass through conductive layer. Also included is a third device die located in a third device level above the second device die, the third device die attached to at least one of a top surface of the second device die or a top surface of the second dummy die.

The present invention discloses a sealing mechanism for a butterfly valve body, comprising: a valve body main body, and a sealing assembly is arranged inside the valve body main body. In the present invention, the right end of the valve body main body is the water inlet, and the left end is the water outlet. At this time, the annular butterfly plate is in a closed state. When the annular butterfly plate is driven by the valve stem to open, as the annular butterfly plate rotates, it pushes the second sealing pad and the first sealing pad towards the support plate, and at the same time pushes the fixing plate to move and compresses a plurality of first springs. When the medium flows, it first passes through the support plate, and under the action of the triangular buffer rod, the medium is buffered to reduce the impact force of the medium. Then it flows to the fixing plate, and under the action of the flow splitting plate, the medium is split into multiple streams, and the impact force of the medium is buffered again. With the cooperation of the triangular buffer rod and the flow splitting plate, the impact force of the medium is split and buffered, reducing the wear of the annular sealing ring caused by the impact force and improving the sealing performance.

The present invention belongs to the technical field of non-oriented silicon steel production and discloses a method for improving the cold-rolled edge quality of high-grade non-oriented silicon steel burr coils. The method comprises: cold rolling using a multi-stand continuous rolling mill or a single-stand reversible rolling mill for 5 to 7 rolling passes, controlling the flatness of the plate after the first, second, and final rolling passes to meet a specific formula; in addition, the rolling speed of the first pass, the emulsion flow rate of the first and second passes, and the original roughness of the mill work rolls can be further controlled. The present invention achieves the purpose of controlling rolled edge cracks and reducing edge waves by adopting a large edge wave rolling mode and adjusting the rolling speed, emulsion flow rate, and the original roughness of the work rolls, thereby achieving optimal cold-rolled edge quality of the burr coils and improving production efficiency and product yield.

A system may, in a first pass: divide content to be summarized into a plurality of chunks and, for each chunk: execute a language model with the chunk and an instruction to summarize the chunk, generate, based on the executed language model, a summary of the chunk. In a subsequent pass, the system may: generate a plurality of groups of summaries, each group of summaries from among the plurality of groups of summaries comprising two or more summaries, each summary corresponding to a respective chunk, for each group of summaries from among the plurality of groups: execute a language model with the group of summaries and an instruction to summarize the group of summaries, generate, based on the executed language model on the group of summaries, a group summary. The system may iteratively repeat the subsequent pass for group summaries until a summary of the content is reached.

The invention relates to an active inductor assembly for a power data system and a related method. An active inductor assembly for a data on power (DoP) system includes a first active inductor and a second active inductor. The first active inductor includes a first pass element and a first control circuit. The first control circuit is configured to control the first pass element such that (i) the first pass element simulates the first inductor in response to a change in voltage across the first pass element, and (ii) the first pass unit adjusts a voltage drop across the first pass element during steady-state operation. The second active inductor includes a second pass element and a second control circuit. The second control circuit is configured to control the second pass element such that (i) the second pass element simulates the second inductor in response to a change in voltage across the second pass element, and (ii) the second pass element adjusts a voltage drop across the second pass element during steady-state operation.

The invention discloses a hot rolling method of an isotropic SP700 titanium alloy wide thin plate. The method comprises the following steps: 1, coating the surface of an SP700 titanium alloy plate blank with an anti-oxidation coating; 2, rolling for the first heating number; 3, cutting treatment; 4, quenching treatment; 5, rolling for the second heating number; 6, cutting treatment; 7, rolling for the third heating number; 8, cutting treatment; 9, fourth fire rolling; and 10, finished product heat treatment. One-heating-number near-beta rolling is conducted on the SP700 titanium alloy plate blank, follow-up beta quenching treatment and alpha-beta phase region medium-temperature reversing rolling are combined, the wide fine-grain thin plate with the B texture as the main part is prepared, and therefore the longitudinal and transverse mechanical property difference is controlled within 15 MPa, wrapping ply rolling is adopted for being assisted by reasonable pass deformation distribution, and the thickness of the thin plate is reduced. The minimum thickness of the obtained wide-width thin plate can reach 0.5 mm, the maximum width of the wide-width thin plate can reach 1800 mm, in addition, material loss of a common process is reduced, and production benefits are improved.

Type definitions of user-defined types in application program code for which definitions are absent (“unknown types”) are inferred. A static analyzer implements two passes of a fixed-point type inference algorithm. Each pass encompasses a plurality of traversals of the application's control flow to build inferred definitions of unknown types until the inferred definitions are maximally built. To build an inferred definition, based on inferring a variable is an unknown type, the static analyzer infers member variables / functions of the unknown type based on contextual information associated with the variable. Type information of unknown types is propagated along control flow paths. After the first pass terminates, unknown types can be assigned known types based on matching of inferred definitions. Inferred definitions of remaining unknown types are incorporated into the application program code. A second pass of type inferencing and data flow analysis are then performed with the inferred definitions incorporated therein.

The embodiment of the invention discloses a programming method of a memory device, the memory device and a memory system. The programming method comprises the following steps: applying a programming voltage to a selected word line; applying a first pass voltage to a first word line adjacent to the selected word line; and applying a second pass voltage to a second word line adjacent to the selected word line; wherein the first passing voltage and the second passing voltage comprise at least two rising stages, and the rising moments of at least one rising stage in other rising stages after the first rising stage of the first passing voltage and the second passing voltage are different. The rising moment is the moment when the voltage value starts to increase on the basis of a certain voltage value.

The utility model provides a fan volute, an air supply structure and an air conditioner, and belongs to the technical field of air conditioners. The fan volute comprises a main body, the main body is of a cylindrical structure, and a first passing opening and a second passing opening are formed in the main body. Along the axis direction of the main body, the inner surface and the outer surface of the main body are respectively provided with an adhesive layer made of a flexible material. According to the fan volute, the wind resistance during operation of the fan can be reduced through the adhesion layer, noise is reduced, and the use experience of a user can be improved.

In some examples, a peripheral circuit is configured to: when performing a first read operation on memory cells coupled to a selected word line, apply a first pass voltage to a first word line, apply a second pass voltage to a second word line, and apply a third pass voltage to a third word line, wherein the first word line comprises at least one word line physically located above and below the selected word line respectively, both the second word line and the third word line comprise word lines physically located on a side of the first word line away from the selected word line, memory cells coupled to the second word line comprise programmed memory cells, memory cells coupled to the third word line comprise unprogrammed memory cells, and the first pass voltage, the second pass voltage and the third pass voltage are all different.

The present application relates to strip steel cold rolling technical field, including step one, determine horizontal roll gap;Step two, the installation and fixation of strip steel;Step three, determine the cross-sectional shape of the strip steel width direction;Step four, non-uniform extension rolling;Step five, uniform extension rolling.The single stand cold rolling method of the present application adopts non-uniform extension method to carry out cold rolling operation on strip steel through first pass and second pass, reduces the transverse thickness difference of cold rolled strip steel, can make subsequent third pass~fifth pass cold rolling more convenient, adopts uniform extension method to roll strip steel through third pass~fifth pass, can make the extension amount of each point in the width direction of strip steel as consistent as possible, so as to achieve the purpose of improving the shape, can reduce the transverse thickness difference of single stand cold rolling mill from 0.009mm~0.01mm to 0.007mm~0.008mm, through straightening when cold rolling the strip steel, can straighten the strip steel wave, is beneficial to the coiling and roll bite of strip steel.

This roll forming machine used for roll forming of a metal plate includes: a first roll group which includes at least one set of rolls forming a first pass through which a metal plate passes, and in which the inclination angle of the normal direction of a cross section of the first pass with respect to the horizontal direction is a first angle (θ1); and a second roll group which includes at least one set of rolls forming a second pass through which the metal plate passes and which is arranged upstream of the first roll group, and in which the inclination angle of the normal direction of a cross section of the second pass with respect to the horizontal direction is a second angle (θ2) larger than the first angle (θ1).

Systems and methods are provided for a transcription system with voice activity detection (VAD). The system includes a VAD module to receive incoming audio and generate an audio segment; and a speech decoder with a split predictor to perform, in a first pass, a decode operation to transcribe text from an audio segment into a message; wherein in the first pass, if the message is determined not to contain a split point based on a content-based analysis performed by the split predictor, the speech decoder forwards the message for display and if the message is determined based on the content-based analysis to contain the split point, the speech decoder performs in a second pass, a re-decode operation to transcribe text from the audio segment based on the split point wherein the split point is configured within an audio domain of the audio segment by the split predictor and forward the message for display.

Methods, apparatuses, and non-transitory computer-readable storage mediums are provided for video encoding. In one method, an encoder signals a Sequence Parameter Set (SPS) rice adaption enabled flag that indicates whether an extension of rice parameter derivation for binarization of abs_remainder and dec_abs_level is used, and a general constraint information (GCI) rice adaptation enabled flag, in GCI syntax to provide general control of the SPS rice adaptation enabled flag. The abs_remainder is coded with Golomb-rice code and bypass-coded bins in a second pass when remaining number of context-coded bins is greater than or equal to 4 while coding in a first pass, and the dec_abs_level is directly coded in the second pass using the Golomb-Rice code and the bypass-coded bins when remaining number of context-coded bins is less than 4 while coding in the first pass.

A radio-frequency module includes first filters, a second filter, a first switch, a second switch, and a first variable circuit element. The first filters have a first pass band that includes a first communication band. The second filter has a second pass band that includes a second communication band. The first switch selects the connection destination of each of a plurality of antenna terminals from among first terminals of the first filters and the second filter. The second switch selects the connection destination of a second terminal of the second filter from among a first power amplifier and a first low-noise amplifier. The first variable circuit element is connected between the second filter and the second switch and shifts an attenuation band on the first pass band side in the second filter.

The utility model relates to the technical field of glass bottle production, in particular to a guide mechanism for continuous feeding of glass bottles, and in the feeding process of the glass bottles, the glass bottles irregularly placed on a conveying belt move forwards along with the conveying belt. When entering an area formed by the fixed column, the adjusting column, the movable column and the guide belt, the glass bottle firstly passes through the adjusting area. The angle of the glass bottle is preliminarily adjusted by adjusting the positions of the adjusting column and the moving column relative to the fixed column and changing the angle of the adjusting area, so that the glass bottle enters the guide area, the width of the guide area is slightly larger than the diameter of the glass, and the glass bottle is in a vertical state under the action of the guide area; bottle bodies of the glass bottles in the vertical state are parallel to the conveying direction of the conveying belt. Through position adjustment of the adjusting column and the moving column, the feeding guide requirements of glass bottles with different diameters can be met, and the limitation that a traditional guide mechanism can only process glass bottles with fixed diameters is overcome. The method can be widely applied to the technical field of glass bottle production.

The invention belongs to the technical field of wire rolling, and particularly discloses a rod and wire rolling method for eliminating steel billet corner defects and a rolling mill stand. The method comprises the following steps that firstly, a heated steel billet is conveyed to a steel rotating device; then the steel billet enters a steel rotating roller way under the action of a guide and guard guide plate, the steel billet is rotated by 45 degrees through a steel rotating device, and the corner of the steel billet is aligned with a groove hole in a roller; then the steel billet is conveyed to a rolling mill through a conveying roller way and a V-shaped anti-torsion guide roller; and finally, the steel billet is rolled into a steel bar through seven passes of rough rolling, wherein large deformation is adopted in the first pass. In order to solve the problem of residual crack defects in billet corner rolling, a pass system is optimized, and the mode that upper and lower flat rollers and left and right vertical rollers are used for simultaneous pressing-down rolling in four directions is adopted, so that the problems that in the continuous casting process, due to the fact that the billet corner cooling speed is higher than that of other parts, shrinkage is uneven, and tensile stress is generated on the surface are solved; therefore, cracks are generated at the corners of the steel billets, and the product quality is influenced.

A method of manufacturing a part with a plurality of cut segments includes receiving a sheet of material with a machining apparatus and removing material during one or more first passes with the machining apparatus to form a plurality of segments in the sheet of material. The method also includes forming a tab by removing material during the one or more first passes with the machining apparatus for forming the segments, the tab connecting two segments within the sheet of material to each other and removing material with the machining apparatus to form joints on the segments.

The door body comprises a door main body and further comprises a first panel, the width of the first panel is consistent with that of the door main body, the first panel can be restrained at the edge of the top of the door main body in an up-and-down moving mode, and the first panel is arranged on the door main body when the cleaning machine is installed in a cabinet. The first panel moves upwards till the top edge of the first panel is aligned with the cabinet door body on the side, and a limiting structure capable of limiting the first panel is arranged on the back of the door body. The cleaning machine can be suitable for various cupboard sizes and dishwasher sizes, during installation, the lower edge of the cleaning machine can be aligned with the lower edge of a cupboard door body by adjusting the bottom feet of the cleaning machine, then the upper edge of the cleaning machine is consistent with the height of a cupboard door by adjusting the vertical position of the first panel, and then the position of the first panel is limited by adjusting the limiting structure; in this way, the height of the door body of the cleaning machine is matched with the height of the installation position of the cabinet, the cabinet does not need to be transformed, the whole cabinet is attractive, assembly gaps can be avoided, and the cabinet is kept clean and tidy.

The utility model belongs to the technical field of bar machining, and discloses a bar cutting and coiling production line. The bar cutting and coiling production line comprises a cutting device, a cleaning device, a feeding device and a coiling device which are sequentially arranged, according to the bar cutting and coiling production line, bars are cut into blank pieces through the cutting device, scraps on the surfaces of the blank pieces are cleaned through the cleaning device, then the blank pieces enter the feeding device, and the cutting device is used for cutting the bars into the blank pieces. Blanks are output in order through the feeding device, and finally the blanks output by the feeding device are loaded into trays through the tray loading device, so that automatic cutting, cleaning and tray loading are achieved, the process of manually cleaning and loading the blanks into the trays is omitted, the labor cost is greatly saved, and meanwhile the physical and psychological health of workers is guaranteed.

A multi-pass free forging punching method for deep blind holes in an engine cylinder block includes the following steps: jaw pressing, chamfering, pre-drawing, upsetting, wide anvil strong pressure drawing, rolling and drawing blanking, upsetting, step forming, pre-punching the blind hole in the first pass with a variable diameter punch, mechanically expanding the blind hole in the second pass, rolling forging and drawing, deep-drawing the blind hole in the third pass with a long punch, rolling forging and drawing, flattening the end face, and finishing the finished product. The blind hole forming process is controlled in multiple passes and stages, with the use of a variable diameter punch for pre-punching and mechanically expanding the blind hole, and a long punch for deep-drawing the blind hole. Furthermore, the forming effect is ensured by combining rolling forging and drawing and end face flattening processes. The multi-pass blind hole forming process can achieve staged shape control in each process, facilitate precise parameter control, suppress inner wall shrinkage, overcome the large inner wall shrinkage phenomenon of traditional punching methods, ensure fiber streamline integrity and dimensional reliability, and avoid the bottom folding phenomenon that is easily generated during traditional blind hole drawing. This method improves forging efficiency and provides more stable quality.

A recursive Analog Content Addressable Memory (ACAM) device includes a first comparison stage and a second comparison stage. The first comparison stage includes a first match line, a first base segment comparator connected between the first match line and a reference node, a first incremented segment comparator, and a first pass transistor. The first pass transistor and the first incremented segment comparator are connected in series between the first match line and the reference node. The second comparison stage includes a first inverter connected to the first pass transistor, a second match line connected to the first inverter, and a second base segment comparator connected between the second match line and the reference node. The device enables efficient comparison of multi-bit values.

The invention relates to a memory cell including a dual anti-fuse device, a memory structure, and a method of operation. A memory cell is disclosed that includes a dual anti-fuse device between a first pass-gate transistor and a second pass-gate transistor. A dual antifuse device includes first and second antifuses having a common terminal and also each having an additional terminal opposite the common terminal. The first pass-gate transistor is connected between the first bit line and an additional terminal for the first pass-gate transistor. The second pass gate transistor is connected between the second bit line and an additional terminal of the second pass gate transistor. A common terminal of the first and second antifuses and gates of the first and second pass-gate transistors are connected to a word line. A memory structure including an array of such memory cells and related methods of operation are also disclosed. Within the array, different word line and bit line bias conditions may be employed to reliably perform program or read operations for selected antifuses in selected cells.

This application belongs to the field of electrical steel manufacturing technology, specifically relating to an ultra-thin non-oriented silicon steel and its preparation method. The preparation method includes: obtaining a silicon steel slab, and performing five cold rolling passes on the silicon steel slab to obtain ultra-thin non-oriented silicon steel. During the five cold rolling passes, there are no external heating sources or heat preservation equipment. The silicon content of the silicon steel slab is 4.0-4.5 wt%. The five cold rolling passes are sequentially divided into a first pass, a second pass, a third pass, a fourth pass, and a fifth pass. The initial rolling temperature of the first pass is room temperature, and the final rolling temperature of the first pass is 60-65℃. This application achieves a self-heating temperature by coordinating and matching the pass reduction rate, rolling speed, and tension, thereby achieving the target thickness and ensuring the plate shape.