8results about How to "Improve finishing efficiency" patented technology

This utility model belongs to the field of packaging equipment technology, specifically a bottle cap sorting machine, including a feeding conveyor belt and a storage box. The storage box is located below the feeding conveyor belt, and the bottom of the feeding conveyor belt is inserted into the inner cavity of the storage box. The surface of the feeding conveyor belt is provided with lifting inclined plates, the upper surface of which is inclined, and the lifting inclined plates are evenly distributed on the surface of the feeding conveyor belt. A screening protrusion is provided in the middle of the feeding conveyor belt. A mounting frame is welded to the side wall of the feeding conveyor belt, and a deflecting frame is welded to the mounting frame. The deflecting frame of this novel machine guides the falling bottle caps to one end of the storage box, preventing the falling bottle caps from landing on the feeding conveyor belt, thereby avoiding knocking the bottle caps off the lifting inclined plates and improving the sorting efficiency of bottle caps. After the bottle caps exceed the screening protrusion, a fall prevention plate limits the bottle caps, preventing them from falling off the lifting inclined plates during lifting or discharge, thus reducing the failure rate of the sorting machine.

The invention belongs to the technical field of large-scale deep learning model training, and particularly relates to a video memory scheduling system based on dynamic sequence length assembly line parallel training. The system comprises a dynamic recalculation module, a video memory arrangement and prefetching module and a mapping reconstruction module, the system collects the length and load information of each micro-batch before each round of iteration, establishes an optimization model with the goal of minimizing the iteration time, dynamically generates a re-calculation plan, and realizes inter-stage video memory load balancing; by introducing a collating stream and an asynchronous scheduler outside a main computing stream, overlapping execution of video memory collating and computing prefetching is realized, and the influence of fragment collating on training performance is reduced; a dynamic physical block mapping strategy is designed based on a CUDA virtual memory management mechanism, the physical block granularity is adaptively adjusted according to the idle state of the video memory, and the API calling and context switching overhead is reduced. Experimental results show that the video memory utilization rate and the system stability are remarkably improved, and an efficient video memory scheduling scheme is provided for large model training.

This utility model relates to the field of electrical power and discloses a distribution cabinet with cable management function. It includes a cabinet body, a door hinged to the outer wall of the cabinet body, a fixed plate fixedly connected to the inner wall of the cabinet body, a slider slidably connected to the inner wall of the fixed plate, an insert block elastically connected to the top inner wall of the slider via a movable spring, an opening and closing mechanism provided on the inner wall of the slider, a dustproof mesh detachably installed on the inner wall of the door, a fixed frame hinged to the outer wall of the door, and a disassembly assembly provided on the inner wall of the door. The opening and closing mechanism includes a fixing ring. In this utility model, through the linkage design of the insert block, movable spring, hinge rod, and fixing ring, operators can quickly adjust the slider position simply by pressing the insert block, achieving flexible positioning and fixing of cables. This simplifies the cable management process and allows for precise adjustment according to actual needs, effectively improving cable management efficiency.

A kind of heat treatment process of mining link ring, including step 1: according to preliminary treatment process, the blank of mining link ring is pre-treated;Wherein, the preliminary treatment process includes: the blank of mining link ring is kept in 900±5 DEG C car-bottom furnace for 3h, after furnace, using 15-35 DEG C tap water quenching, tempering is kept in 500 DEG C car-bottom furnace for 1h, then heated to 700 DEG C and kept for 8-10h, after furnace, air cooling to room temperature;Step 2: using controllable atmosphere furnace heating, K oil is used as quenching medium, according to quenching process, the mining link ring obtained after finishing is quenched;Step 3: the mining link ring after quenching is tempered in turn according to first tempering process and second tempering process;Step 4: the mining link ring after two times of tempering is shot blasting according to shot blasting process.The scheme has no corrosion to parts and equipment, and can improve the mechanical properties of link ring.

The application provides a kind of commercial vehicle frame performance prediction multi-modal data fusion representation method and system, applied to vehicle frame data modeling technical field, the method first carries out automatic mining to multi-source heterogeneous engineering data, extracts structure topological parameter, section configuration parameter, material attribute parameter, connection parameter, P-S-N curve data and field response data, and forms target domain data set and independent test set;Then establish the full parameterization finite element model of frame, combined with the virtual load sample generated by statistical derivation of measured load spectrum, extract two-dimensional load damage feature matrix, and construct source domain data set;Finally, the geometric topological feature, parameter feature, connection feature and load feature are mapped to a unified three-dimensional voxel space, generating a three-dimensional multi-channel engineering data tensor.The application can improve the efficiency of commercial vehicle frame sample construction, multi-modal data utilization efficiency and the generalization ability of performance prediction model.

The present disclosure provides a code arrangement method, device, equipment, medium and program product, comprising: periodically collecting the code of each code set under each task to obtain the first target specification of the task code update of each code set; performing associated call analysis on the interfaces and implementations in different code sets to obtain an associated call relationship; determining an associated call implementation point code set collection and an associated call initiation point code set collection based on the associated call relationship, and obtaining a second target specification of the associated call implementation point code set collection and a third target specification of the associated call initiation point code set collection; performing dependency iteration mixing on the first target specification according to the second target specification and the third target specification to obtain a dependency iteration target specification; and processing the dependency iteration target specification and the dependency iteration target specification in the last period based on a specification modification model to determine an updated code set. The development efficiency and the code arrangement efficiency can be improved.

This invention discloses a method and system for intelligent generation and processing of meeting content based on a multimodal large model, comprising: collecting raw meeting data and performing standardized preprocessing; inputting the multimodal large model, extracting multimodal features and semantically aligning them; performing cross-modal hash encoding to generate binary codes and establish an index library; hash-retrieval of relevant segments to construct a directed graph of meeting content; optimizing the path using Monte Carlo tree search based on the directed graph structure; generating structured meeting content and outputting minutes, summaries, and action items. This invention achieves efficient extraction, accurate retrieval, and intelligent structured generation of meeting content by integrating a multimodal large model, cross-modal hash encoding, and Monte Carlo tree search.

The utility model provides a kind of bamboo shoot processing is used to kill green device.The bamboo shoot processing is used to kill green device including bottom plate, the top of the bottom plate is fixedly installed for the steaming box of bamboo shoot kill green, the inner wall of the steaming box is fixedly installed with electric heating pipe;Bracket is arranged in the upper portion of the steaming box, two mouth shape boards are fixedly installed on the bracket, two the mesh frame for placing bamboo shoot is arranged on the mouth shape board, the top of two mesh frames is provided with baffle;Mobile mechanism for exchanging the position of two mesh frames is assembled on the bottom plate.The bamboo shoot processing is used to kill green device provided by the utility model solves the technical problem that the efficiency of kill green is affected due to long time consumption for manually fishing out bamboo shoot in the process of existing bamboo shoot pre-cooking kill green.