Method and apparatus for processing data sequence
The present application discloses a method and apparatus for processing a data sequence. A specific implementation of the method includes: receiving an inputted to-be-processed data sequence; copying a weight matrix in a recurrent neural network model to an embedded block random access memory (RAM) of a field-programmable gate array (FPGA); processing sequentially each piece of to-be-processed data in the to-be-processed data sequence by using an activation function in the recurrent neural network model and the weight matrix stored in the embedded block RAM; and outputting a processed data sequence corresponding to the to-be-processed data sequence. This implementation improves the data sequence processing efficiency of the recurrent neural network model.
1. A method for processing a data sequence, comprising:
receiving an inputted to-be-processed data sequence;
copying a weight matrix in a recurrent neural network model to an embedded block random access memory (RAM) of a field-programmable gate array (FPGA);
processing sequentially each piece of to-be-processed data in the to-be-processed data sequence by using an activation function in the recurrent neural network model and the weight matrix stored in the embedded block RAM; and
outputting a processed data sequence corresponding to the to-be-processed data sequence,
wherein before the copying, the method further comprises:
calling an address assignment interface to assign a storage address in the embedded block RAM to the weight matrix, and
wherein the copying comprises:
calling a copying interface to copy the weight matrix stored in a double data rate synchronous dynamic random access memory to the storage address in the embedded block RAM that is assigned to the weight matrix,
wherein the copying of the weight matrix is performed only once during the process of processing the to-be-processed data sequence.
2. The method according to claim 1 , further comprising:
deleting the weight matrix stored in the embedded block RAM after the processed data sequence is output.
3. The method according to claim 2 , wherein the deleting the weight matrix stored in the embedded block RAM comprises:
calling a deletion interface to delete the weight matrix stored in the embedded block RAM.
4. The method according to claim 1 , wherein the embedded block RAM is a static random access memory.
5. An apparatus for processing a data sequence, comprising:
at least one processor; and
a memory storing instructions, which when executed by the at least one processor, cause the at least one processor to perform operations, the operations comprising:
receiving an inputted to-be-processed data sequence;
copying a weight matrix in a recurrent neural network model to an embedded block random access memory (RAM) of a field-programmable gate array (FPGA);
processing sequentially each piece of to-be-processed data in the to-be-processed data sequence by using an activation function in the recurrent neural network model and the weight matrix stored in the embedded block RAM; and
outputting a processed data sequence corresponding to the to-be-processed data sequence,
wherein, the operations further comprises, before the copying:
calling an address assignment interface to assign a storage address in the embedded block RAM to the weight matrix, and
wherein the copying comprises;
calling a copying interface to copy the weight matrix stored in a double data rate synchronous dynamic random access memory to the storage address in the embedded block RAM that is assigned to the weight matrix,
wherein the copying of the weight matrix is performed only once during the process of processing the to-be-processed data sequence.
6. The apparatus according to claim 5 , wherein the operations further comprises:
deleting the weight matrix stored in the embedded block RAM after the processed data sequence is output.
7. The apparatus according to claim 6 , wherein the deleting the weight matrix stored in the embedded block RAM comprises:
calling a deletion interface to delete the weight matrix stored in the embedded block RAM.
8. The apparatus according to claim 5 , wherein the embedded block RAM is a static random access memory.
9. A non-transitory storage medium storing one or more programs, the one or more programs when executed by an apparatus, causing the apparatus to perform operations, the operations comprising:
receiving an inputted to-be-processed data sequence;
copying a weight matrix in a recurrent neural network model to an embedded block random access memory (RAM) of a field-programmable gate array (FPGA);
processing sequentially each piece of to-be-processed data in the to-be-processed data sequence by using an activation function in the recurrent neural network model and the weight matrix stored in the embedded block RAM; and
outputting a processed data sequence corresponding to the to-be-processed data sequence,
wherein the operations further comprises, before the copying:
calling an address assignment interface to assign a storage address in the embedded block RAM to the weight matrix, and
wherein the copying comprises:
calling a copying interface to copy the weight matrix stored in a double data rate synchronous dynamic random access memory to the storage address in the embedded block RAM that is assigned to the weight matrix,
wherein the copying of the weight matrix is performed only once during the process of processing the to-be-processed data sequence.
10. The non-transitory storage medium according to claim 9 , wherein the operations further comprises:
deleting the weight matrix stored in the embedded block RAM after the processed data sequence is output.
11. The non-transitory storage medium according to claim 10 , wherein the deleting the weight matrix stored in the embedded block RAM comprises:
calling a deletion interface to delete the weight matrix stored in the embedded block RAM.
12. The non-transitory storage medium according to claim 9 , wherein the embedded block RAM is a static random access memory.