Abstract: A method includes obtaining an image, a set of attribute labels, and a set of object labels and performing prompt tuning of a pre-trained vision-language model having first and second textual encoders and a vision encoder. The model is trained during prompt tuning to select one attribute label and one object label that match content in the image. Performing the prompt tuning includes, for each attribute label-object label pair, generating object textual features associated with the object label using the first textual encoder, generating attribute textual features associated with the attribute label using the second textual encoder, and generating image features associated with the image using the vision encoder. Intermediate outputs from initial layers of the textual encoders and the vision encoder are combined to generate layer-specific learnable prompt tokens that are appended to inputs of specified layers in the first and second textual encoders and the vision encoder.

Abstract: The present invention provides a polyurethane nanofiber waterproof moisture-permeable film, a preparation method thereof and application thereof. The polyurethane nanofiber waterproof moisture-permeable film is prepared from two kinds of isothiocyanate and two kinds of polyethylene glycol as raw materials, and the polyurethane nanofiber waterproof moisture-permeable film has good water pressure resistance, tensile property, and moisture permeability. In addition, the polyurethane nanofiber waterproof moisture-permeable film is prepared by adopting an electrostatic spinning method, is low in cost, simple to operate and high in production efficiency, and has a good application prospect in the fields of clothing, medical treatment, electronic appliances, and the like.

Abstract: A micro puree machine including a housing, a power shaft, a bowl assembly and a platform. The power shaft extends from the housing. The bowl assembly including at least one locking bowl element. The platform includes at least one complementary locking platform element that is configured to engage the at least one locking bowl element such that rotation of the bowl assembly relative to the platform is prevented at times the bowl assembly is positioned thereon. The platform is rotatable from a first position to a second position relative to the housing such that the platform raises the bowl assembly towards the power shaft during the rotation of the bowl assembly and platform. The raising of the bowl assembly facilitates connection between the power shaft and a blade assembly that is positioned in a lid assembly on the bowl assembly.

Abstract: A method of making a semiconductor device includes depositing a TiN layer over a substrate. The method further includes doping a first portion of the TiN layer using an oxygen-containing plasma treatment. The method further includes doping a second portion of the TiN layer using a nitrogen-containing plasma treatment, wherein the second portion of the TiN layer directly contacts the first portion of the TiN layer. The method further includes forming a first metal gate electrode over the first portion of the TiN layer. The method further includes forming a second metal gate electrode over the second portion of the TiN layer, wherein the first metal gate electrode has a different work function from the second metal gate electrode, and the second metal gate electrode directly contacts the first metal gate electrode.

Abstract: A micro puree machine including a housing, a power shaft, a bowl assembly and a platform. The power shaft extends from the housing. The bowl assembly including at least one locking bowl element. The platform includes at least one complementary locking platform element that is configured to engage the at least one locking bowl element such that rotation of the bowl assembly relative to the platform is prevented at times the bowl assembly is positioned thereon. The platform is rotatable from a first position to a second position relative to the housing such that the platform raises the bowl assembly towards the power shaft during the rotation of the bowl assembly and platform. The raising of the bowl assembly facilitates connection between the power shaft and a blade assembly that is positioned in a lid assembly on the bowl assembly.

Abstract: Embodiments of this application provide a packet transmission method, apparatus, and system. A first network device sends a probe packet to a second network device, the probe packet includes an identifier of the first network device, and the identifier of the first network device is used to indicate the second network device to add the identifier of the first network device to a destination address in a response packet of the probe packet; and receives the response packet from the second network device, where the destination address in the response packet is the identifier of the first network device; or receives an updated response packet from a third network device, where a destination address in the updated response packet is the identifier of the first network device. Both the third network device and the first network device are connected to a same user-side network device.

Abstract: A gun-type firing device using a caseless projectile is disclosed. The caseless projectile has a bullet head and an impact seat at two ends thereof and a groove adjacent to the impact seat. The gun-type firing device includes a gunstock, a gun bolt and a gun barrel. The gun bolt is located on one side of the gunstock. The gun bolt includes a cocking handle and an extractor. The cocking handle can perform a linear displacement movement along the extension direction of the gunstock. The extractor is located at one end of the cocking handle. The gun barrel is located on the gunstock and is adjacent to one end of the gun bolt. The gun barrel does not have rifling.

Abstract: A micro puree machine including a housing, a power shaft, a bowl assembly and a platform. The power shaft extends from the housing. The bowl assembly including at least one locking bowl element. The platform includes at least one complementary locking platform element that is configured to engage the at least one locking bowl element such that rotation of the bowl assembly relative to the platform is prevented at times the bowl assembly is positioned thereon. The platform is rotatable from a first position to a second position relative to the housing such that the platform raises the bowl assembly towards the power shaft during the rotation of the bowl assembly and platform. The raising of the bowl assembly facilitates connection between the power shaft and a blade assembly that is positioned in a lid assembly on the bowl assembly.

Abstract: A method includes obtaining, using a first electronic device, a weight matrix associated with a trained transformer model. The method also includes factorizing the weight matrix into a dictionary weight matrix and an intermediate matrix. The method further includes pruning the intermediate matrix to generate a sparse intermediate matrix. The method also includes fine-tuning the sparse intermediate matrix based on a training dataset to generate a fine-tuned sparse intermediate matrix. The method further includes determining an index matrix and a coefficient matrix based on the fine-tuned sparse intermediate matrix. In addition, the method includes deploying the dictionary weight matrix, the index matrix, and the coefficient matrix to a second electronic device without deploying the weight matrix to the second electronic device. A number of parameters in the dictionary weight matrix, the index matrix, and the coefficient matrix is smaller than a number of parameters in the weight matrix.

Abstract: A gas-supplying mouth mask includes: a mask body for leaning against a user's nose to development of a semi-open space between the mask body and the user's face, wherein the mask body comprising a breathable structure and an airtight structure; and a gas generation unit placed in the mask body for generation of functional gases which can be released to around the user's nose through the breathable structure. Based on the above structures, the gas-supplying mouth mask encircles the functional gases inside the semi-open space formed by the mask body and the user's face and causes a better utilization rate of functional gases.

Abstract: A bit index explicit replication (BIER) operations, administration, and maintenance (OAM) detection method includes a bit forwarding ingress router (BFIR) obtaining a detection request packet based on a first BIER OAM packet, and sending the detection request packet to at least one bit forwarding egress router BFER. The detection request packet includes a first packet and a first packet header. The first packet is a packet obtained by encapsulating the first BIER OAM packet. The first packet header includes a bit string, and the bit string indicates the at least one bit forwarding egress router BFER that is to be measured.

Abstract: The present invention relates to a micro puree machine that includes a blade assembly, a lid assembly, and a clip mechanism. The blade assembly has a central support hub, a blade, and an undercut. The lid assembly is configured to cooperate with a bowl assembly and includes a lid and a clip mechanism. The clip mechanism is movable between a first position and a second position and cooperates with the undercut at times the clip mechanism is the in the first position. The blade assembly is retained in the lid assembly at times the clip mechanism is in or near the first position, and the blade assembly is disengaged from the undercut and removable from the lid assembly at times the clip mechanism is in or near the second position.

Abstract: A method includes receiving one or more training corpora for training a machine learning model having a plurality of encoder blocks, where each encoder block includes an attention layer and a feedforward network. The method also includes using the one or more training corpora to train an attention dictionary shared across the plurality of encoder blocks. Training the attention dictionary may include training attention parameters of the attention layer in each of the plurality of encoder blocks, and the attention parameters for a given encoder block among the plurality of encoder blocks may be a weighted combination of columns from the attention dictionary shared across the plurality of encoder blocks.

Abstract: A method includes obtaining a parameter matrix associated with a linear layer of a first machine learning model and containing parameter values for parameters of the linear layer. The method also includes determining importance values corresponding to the parameter values. The method further includes generating factorized matrices such that a product of the importance values and factorized matrices contains approximated parameter values for the parameters of the linear layer. In addition, the method includes generating a second machine learning model representing a compressed version of the first machine learning model. The second machine learning model has first and second linear layers containing parameter values based on the importance values and the factorized matrices. The factorized matrices are generated based on weighted errors between the parameter values for the parameters of the linear layer and the approximated parameter values.

Abstract: A method for obtaining segment routing over Internet Protocol version 6 data plane (SRv6) tunnel information of Internet Protocol version 6 segment routing, including sending, by a first network device, a request packet to a second network device, where the request packet is used to request to detect reachability of an SRv6 tunnel and obtain SRv6 tunnel information of the second network device, and the second network device is a network device on the SRv6 tunnel, receiving, by the first network device, a response packet from the second network device, where the response packet includes the SRv6 tunnel information of the second network device, and obtaining, by the first network device, the SRv6 tunnel information of the second network device based on the response packet.

Abstract: The present invention relates to a micro puree machine that includes a blade assembly, a lid assembly, and a clip mechanism. The blade assembly has a central support hub, a blade, and an undercut. The lid assembly is configured to cooperate with a bowl assembly and includes a lid and a clip mechanism. The clip mechanism is movable between a first position and a second position and cooperates with the undercut at times the clip mechanism is the in the first position. The blade assembly is retained in the lid assembly at times the clip mechanism is in or near the first position, and the blade assembly is disengaged from the undercut and removable from the lid assembly at times the clip mechanism is in or near the second position.

Abstract: A gas permeable layer capable of supplying hydrogen includes a thin layer, encapsulating a hydrogen production formula. An outer side of the thin layer is airtight. An inner side is air-permeable. An inner side surface has a plurality of small holes. The thin layer can be a single layer or a composite layer. The hydrogen production formula does not dissipate. The hydrogen production formula absorbs moisture in the air or liquid water, thereby generating hydrogen. The hydrogen is released onto the skin and into the human body through the small holes. The hydrogen production formula includes metal peroxides, metal hydroxides, or metal hydrides and aluminum powder, or microsilica. The gas permeable layer can be used in sanitary products including eye masks, mouth masks, face masks, cosmetic facial masks, bras, pasties, nursing pads, sanitary napkins (towels), diapers, panty liners, wound dressing, woundplasts, bandage gauze, decubitus pads.

Abstract: A method of making a semiconductor device includes depositing a TiN layer over a substrate. The method further includes doping a first portion of the TiN layer using an oxygen-containing plasma treatment. The method further includes doping a second portion of the TiN layer using a nitrogen-containing plasma treatment, wherein the second portion of the TiN layer directly contacts the first portion of the TiN layer. The method further includes forming a first metal gate electrode over the first portion of the TiN layer. The method further includes forming a second metal gate electrode over the second portion of the TiN layer, wherein the first metal gate electrode has a different work function from the second metal gate electrode, and the second metal gate electrode directly contacts the first metal gate electrode.

Abstract: A micro puree machine including a housing, a power shaft, a bowl assembly and a platform. The power shaft extends from the housing. The bowl assembly including at least one locking bowl element. The platform includes at least one complementary locking platform element that is configured to engage the at least one locking bowl element such that rotation of the bowl assembly relative to the platform is prevented at times the bowl assembly is positioned thereon. The platform is rotatable from a first position to a second position relative to the housing such that the platform raises the bowl assembly towards the power shaft during the rotation of the bowl assembly and platform. The raising of the bowl assembly facilitates connection between the power shaft and a blade assembly that is positioned in a lid assembly on the bowl assembly.