Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training a machine learning model to perform a machine learning task. In one aspect, a method performed by one or more computer is described. The method includes: obtaining data defining a compute budget that characterizes an amount of computing resources allocated for training a machine learning model to perform a machine learning task; processing the data defining the compute budget using an allocation mapping, in accordance with a set of allocation mapping parameters, to generate an allocation tuple defining: (i) a target model size for the machine learning model, and (ii) a target amount of training data for training the machine learning model; instantiating the machine learning model, where the machine learning model has the target model size; and obtaining the target amount of training data for training the machine learning model.

Some embodiments of the invention provide a method for processing requests for performing operations on resources in a software defined datacenter (SDDC). The resources are software-defined (SD) resources in some embodiments. The method initially receives a request to perform an operation with respect to a first resource in the SDDC. The method identifies a policy that matches (i.e., is applicable to) the received request for the first resource by comparing a set of attributes of the request with sets of attributes of a set of policies that place constraints on operations specified for resources. In some embodiments, several sets of attributes for several policies can be expressed for resources at different hierarchal resource levels of the SDDC. The method rejects the received request when the identified policy specifies that the requested operation violates a constraint on operations specified for the first resource.

In one embodiment, an apparatus comprises a memory and a processor. The memory is to store visual data associated with a visual representation captured by one or more sensors. The processor is to: obtain the visual data associated with the visual representation captured by the one or more sensors, wherein the visual data comprises uncompressed visual data or compressed visual data; process the visual data using a convolutional neural network (CNN), wherein the CNN comprises a plurality of layers, wherein the plurality of layers comprises a plurality of filters, and wherein the plurality of filters comprises one or more pixel-domain filters to perform processing associated with uncompressed data and one or more compressed-domain filters to perform processing associated with compressed data; and classify the visual data based on an output of the CNN.

In one embodiment, an apparatus comprises a storage device and a processor. The storage device may store a plurality of compressed images comprising one or more compressed master images and one or more compressed slave images. The processor may: identify an uncompressed image; access context information associated with the uncompressed image and the one or more compressed master images; determine, based on the context information, whether the uncompressed image is associated with a corresponding master image; upon a determination that the uncompressed image is associated with the corresponding master image, compress the uncompressed image into a corresponding compressed image with reference to the corresponding master image; upon a determination that the uncompressed image is not associated with the corresponding master image, compress the uncompressed image into the corresponding compressed image without reference to the one or more compressed master images; and store the corresponding compressed image on the storage device.

Some embodiments of the invention provide a method for processing requests for performing operations on resources in a software defined datacenter (SDDC). The resources are software-defined (SD) resources in some embodiments. The method initially receives a request to perform an operation with respect to a first resource in the SDDC. The method identifies a policy that matches (i.e., is applicable to) the received request for the first resource by comparing a set of attributes of the request with sets of attributes of a set of policies that place constraints on operations specified for resources. In some embodiments, several sets of attributes for several policies can be expressed for resources at different hierarchal resource levels of the SDDC. The method rejects the received request when the identified policy specifies that the requested operation violates a constraint on operations specified for the first resource.

Systems, apparatuses, and methods for controlling bandwidth through shared transaction limits are described. An apparatus includes at least a plurality of agents, a plurality of transaction-limit (T-Limit) nodes, a T-Limit manager, and one or more endpoints. The T-Limit manager creates a plurality of credits for the plurality of agents to send transactions to a given endpoint. Then, the T-Limit manager partitions the credits into N+1 portions for N agents, wherein the extra N+1 portion is a shared pool for use by agents when they run out of their private credits. The T-Limit manager assigns a separate private portion of the N portions to the N agents for use by only the corresponding agent. When an agent runs out of private credits, the agent's T-Limit node sends a request to the T-Limit manager for credits from the shared pool.

A computer implemented method comprises a server processing work requests of a work requester. The work requester can communicate to the server a processing dependency of one work request on a second work request. The server can associate the dependency with the work requests and/or a queue of work requests. The dependency include a condition to be met in association with processing the work requests, and the condition can include an action for the server to take in association with processing a work request. A computing system can comprise a work requester, a server, and a set of dependency-aware queues for processing a set of work requests. A queue and/or work requests on the queues can be associated with a processing dependency and the server can process work requests enqueued to the queues in an order based on the dependencies. A work requester/server interface can comprise a dependency framework.

Systems and techniques are provided for monitoring and managing the performance of services accessed by sites on a computing platform. When a performance issue is identified, a service is monitored to determine if calls to the service exceed a threshold completion time. If so, a resource available to call the service is adaptively throttled by the platform.

Software performance can be automatically managed in a distributed computing environment. In one example, a system that can receive metrics information describing resource usage by a first instance of a service in a distributed computing environment. The system can also determine a quality-of-service (QoS) constraint for the service. The system can then modify a definition file based on the metrics information and the QoS constraint, the definition file being configured for deploying instances of the service in the distributed computing environment. The system can deploy a second instance of the service in the distributed computing environment using the modified definition file. As a result, the second instance can more closely satisfy the QoS constraint than the first instance.

In one embodiment, an apparatus comprises a communication interface and a processor. The communication interface is to communicate with a plurality of devices. The processor is to: receive compressed data from a first device, wherein the compressed data is associated with visual data captured by sensor(s); perform a current stage of processing on the compressed data using a current CNN, wherein the current stage of processing corresponds to one of a plurality of processing stages associated with the visual data, and wherein the current CNN corresponds to one of a plurality of CNNs associated with the plurality of processing stages; obtain an output associated with the current stage of processing; determine, based on the output, whether processing associated with the visual data is complete; if the processing is complete, output a result associated with the visual data; if the processing is incomplete, transmit the compressed data to a second device.