Methods, devices, and computer program products for processing target data
A target data processing method comprises: in response to receiving, from a client, a request for processing target data, obtaining, at a server and from a storage device, the target data, the target data being predetermined to be obtained by the client from the storage device and a first set of operations being predetermined to be performed on the target data; and processing, at the server, the target data by performing the first set of operations on the target data such that at least one dedicated processing resource of the server performs, on the processed target data, a second set of operations associated with the first set of operations. As such, the data moving path may be shortened significantly, the network traffic pressure may be reduced significantly, the efficiency of the target data processing system may be improved, and moreover, the number of dedicated processing resources in the server may be easily scaled.
1. A method of processing target data by a server comprising a first processing device and a plurality of dedicated processing devices that are different than the first processing device, the method comprising:
obtaining, from a client device, a request for processing target data for use in a machine learning operation, the request comprising an indication of at least one pre-processing process to be performed on the target data;
based at least in part on the obtained request for processing the target data, obtaining the target data from a storage device, the storage device comprising a repository of data for use in machine learning operations;
performing the at least one pre-processing process on the target data using the first processing device of the server to generate pre-processed data;
storing the pre-processed target data into a first dedicated processing device of the plurality of dedicated processing devices;
determining that a given dedicated processing device of the plurality of dedicated processing devices is specified to perform the at least one machine learning operation on the pre-processed data;
moving the pre-processed target data from the first dedicated processing device to the given dedicated processing device based at least in part on the determination that the given dedicated processing device of the plurality of dedicated processing devices is specified to perform the at least one machine learning operation on the pre-processed data;
performing at least one machine learning operation on the pre-processed data using the given dedicated processing device of the server to generate machine learning output data; and
providing the generated machine learning output data to the client device.
2. The method according to claim 1 , wherein:
the indication of the at least one pre-processing process to be performed on the target data comprises a coordination message, the coordination message indicating that a batch-processing scheme be performed on the target data; and
performing the at least one pre-processing process on the target data using the first processing device of the server to generate pre-processed data comprises processing, based at least in part on the batch-processing scheme, the target data in batches and synthesizing, based on the batch-processing scheme, the batch-processed target data.
3. The method according to claim 1 , wherein the method further comprises:
receiving, from the client device, a message that specifies the given dedicated processing device of the plurality of dedicated processing devices for performing the at least one machine learning operation on the pre-processed data, wherein determining that the given dedicated processing device of the plurality of dedicated processing devices is specified to perform the at least one machine learning operation on the pre-processed data is based at least in part on the received message; and
determining that the first dedicated processing device is different than the given dedicated processing device, wherein the moving of the pre-processed data from the first dedicated processing device to the given dedicated processing device is based at least in part on the determination that the first dedicated processing device is different than the given dedicated processing device.
4. A computer program product being tangibly stored on a non-transient computer readable medium and comprising machine executable instructions which, when being executed, causing a machine to perform steps of the method according to claim 1 .
5. A device for processing target data, comprising:
a first processing device;
a plurality of dedicated processing devices that are different than the first processing device;
at least one memory coupled to the first processing device and the dedicated processing devices and storing instructions to be executed by at least one of the first processing device and the dedicated processing devices, the instructions, when being executed by the at least one of the first processing device and the dedicated processing devices, causing the device to perform acts comprising:
obtaining, from a client device, a request for processing target data for use in a machine learning operation, the request comprising an indication of at least one pre-processing process to be performed on the target data;
based at least in part on the obtained request for processing the target data, obtaining the target data from a storage device, storage device comprising a repository of data for use in machine learning operations;
performing the at least one pre-processing process on the target data using the first processing device of the server to generate pre-processed data;
storing the pre-processed target data into a first dedicated processing device of the plurality of dedicated processing devices;
determining that a given dedicated processing device of the plurality of dedicated processing devices is specified to perform the at least one machine learning operation on the pre-processed data;
moving the pre-processed target data from the first dedicated processing device to the given dedicated processing device based at least in part on the determination that the given dedicated processing device of the plurality of dedicated processing devices is specified to perform the at least one machine learning operation on the pre-processed data; and
performing at least one machine learning operation on the pre-processed data using the given dedicated processing device of the server to generate machine learning output data.
6. The device according to claim 5 , wherein:
the indication of the at least one pre-processing process to be performed on the target data comprises a coordination message, the coordination message indicating that a batch-processing scheme be performed on the target data; and
performing the at least one pre-processing process on the target data using the first processing device of the server to generate pre-processed data comprises processing, based at least in part on the batch-processing scheme, the target data in batches and synthesizing, based on the batch-processing scheme, the batch-processed target data.
7. The device according to claim 5 , wherein the acts further comprise:
receiving, from the client device, a message that specifies the given dedicated processing device of the plurality of dedicated processing devices for performing the at least one machine learning operation on the pre-processed data, wherein determining that the given dedicated processing device of the plurality of dedicated processing devices is specified to perform the at least one machine learning operation on the pre-processed data is based at least in part on the received message; and
determining that the first dedicated processing device is different than the given dedicated processing device, wherein the moving of the pre-processed data from the first dedicated processing device to the given dedicated processing device is based at least in part on the determination that the first dedicated processing device is different than the given dedicated processing device.
8. A method of processing target data, comprising:
determining, at a client device, whether target data is to be processed by the client device, the target data being predetermined to be obtained by the client device from a storage device comprising a repository of data for use in machine learning operations; and
in response to determining that the target data is not to be processed by the client device, sending, from the client device to a server, a request for processing the target data, the request comprising an indication of at least one pre-processing process to be performed on the target data by the server, wherein the server is configurable, based at least in part on the request:
to obtain the target data from the storage device;
to perform the at least one pre-processing process on the target data using a first processing device of the server to generate pre-processed data;
to store the pre-processed target data into a first dedicated processing device of a plurality of dedicated processing devices of the server that are different than the first processing device;
to determine that a given dedicated processing device of the plurality of dedicated processing devices is specified to perform the at least one machine learning operation on the pre-processed data;
to move the pre-processed target data from the first dedicated processing device to the given dedicated processing device based at least in part on the determination that the given dedicated processing device of the plurality of dedicated processing devices is specified to perform the at least one machine learning operation on the pre-processed data; and
to perform at least one machine learning operation on the pre-processed data using the given dedicated processing device of the server to generate machine learning output data; and
obtaining, from the server, the generated machine learning output data;
wherein the method is performed by at least one processing device of the client device comprising at least one processor coupled to memory.
9. The method according to claim 8 wherein:
the indication of the at least one pre-processing process to be performed on the target data comprises a coordination message, the coordination message indicating that a batch-processing scheme be performed on the target data; and
performing the at least one pre-processing process on the target data using the first processing device of the server to generate pre-processed data comprises processing, based at least in part on the batch-processing scheme, the target data in batches and synthesizing, based on the batch-processing scheme, the batch-processed target data.
10. The method according to claim 8 , wherein the method further comprises:
sending, to the server, a message that specifies the given dedicated processing device of the plurality of dedicated processing devices for performing the at least one machine learning operation on the pre-processed data, wherein determining that the given dedicated processing device of the plurality of dedicated processing devices is specified to perform the at least one machine learning operation on the pre-processed data is based at least in part on the received message, the server being further configurable to;
determine that the first dedicated processing device is different than the given dedicated processing device, wherein the moving of the pre-processed data from the first dedicated processing device to the given dedicated processing device is based at least in part on the determination that the first dedicated processing device is different than the given dedicated processing device.
11. A computer program product being tangibly stored on a non-transient computer readable medium and comprising machine executable instructions which, when being executed, causing a machine to perform steps of the method according to claim 8 .
12. A device for processing target data, comprising at least one processing unit, and at least one memory coupled to the at least one processing unit and storing instructions to be executed by the at least one processing unit, the instructions, when being executed by the at least one processing unit, causing the device to perform steps of claim 8 .
13. The method according to claim 1 , wherein the first processing device comprises a central processing unit of the server and the plurality of dedicated processing devices comprise at least one graphics processing unit of the server.
14. The device according to claim 5 , wherein the first processing device comprises a central processing unit of the server and the plurality of dedicated processing devices comprise at least one graphics processing unit of the server.
15. The method according to claim 8 , wherein the first processing device comprises a central processing unit of the server and the plurality of dedicated processing devices comprise at least one graphics processing unit of the server.
16. The method according to claim 1 , wherein:
the repository of data stored in the storage device comprises a repository of training data; and
the at least one machine learning operation comprises a training operation.
17. The device according to claim 5 , wherein:
the repository of data stored in the storage device comprises a repository of training data; and
the at least one machine learning operation comprises a training operation.
18. The method according to claim 8 , wherein:
the repository of data stored in the storage device comprises a repository of training data; and
the at least one machine learning operation comprises a training operation.