Method and system with deep learning model generation
Provided is a method and system with deep learning model generation. The method includes identifying a plurality of connections in a neural network that is pre-associated with a deep learning model, generating a plurality of pruned neural networks by pruning different sets of one or more of the plurality of connections to respectively generate each of the plurality of pruned neural networks, generating a plurality of intermediate deep learning models by generating a respective intermediate deep learning model corresponding to each of the plurality of pruned neural networks, and selecting one of the plurality of intermediate deep learning models, having a determined greatest accuracy among the plurality of intermediate deep learning models, to be an optimized deep learning model.
1. A processor implemented method, the method comprising:
identifying a plurality of connections in a neural network that is pre-associated with a deep learning model;
determining different sets, each set comprising a different combination of one or more of the identified plurality of connections;
generating a plurality of pruned neural networks by iteratively pruning the determined different sets of one or more of the plurality of connections to respectively generate each of the plurality of pruned neural networks;
generating a plurality of intermediate deep learning models by generating a respective intermediate deep learning model corresponding to each of the plurality of pruned neural networks;
determining an accuracy level of each of the plurality of intermediate deep learning models to determine a greatest accuracy among the plurality of intermediate deep learning models; and
selecting, based on the determined accuracy level of each of the plurality of the intermediate deep learning models, one of the plurality of intermediate deep learning models, having the determined greatest accuracy from among the plurality of intermediate deep learning models, as an optimized deep learning model,
wherein a first intermediate deep learning model is generated corresponding to a first pruned neural network and a second intermediate deep learning model is generated corresponding to a second pruned neural network, and
wherein the first intermediate deep learning model has a greater accuracy level than the second intermediate deep learning model and is selected as the optimized deep learning model.
2. The method of claim 1 , wherein the pruning of the different sets of the one or more of the plurality of connections is performed based on predetermined pruning policies.
3. The method of claim 2 , further comprising updating the predetermined pruning policies based on a determined accuracy level of the optimized deep learning model.
4. The method of claim 2 , wherein the predetermined pruning policies comprise at least one of a policy of pruning one or more connections for a predetermined time period or a policy of pruning connections until a threshold number of connections are pruned.
5. The method of claim 2 , wherein the pruning of the different sets of the one or more of the plurality of connections comprises:
selecting, at random, respective combinations of two or more connections for pruning; and
pruning each of the respective combinations based on the predetermined pruning policies.
6. The method of claim 1 , wherein the determining of the accuracy levels includes using a predetermined validation technique to determine the accuracy levels.
7. The method of claim 6 , wherein the predetermined validation technique comprises determining an error level corresponding to each of the plurality of intermediate deep learning models.
8. The method of claim 1 , wherein the pruning includes assigning a zero value to each weight corresponding to each pruned connection.
9. The method of claim 1 , wherein each of the plurality of intermediate deep learning models is a subset of the deep learning model.
10. The method of claim 9 , wherein a total number of connections in an intermediate deep learning model, of the plurality of intermediate deep learning models, is less than or equal to a total number of connections in the deep learning model.
11. The method of claim 1 , further comprising implementing the optimized deep learning model.
12. The method of claim 11 , further comprising determining the greatest accuracy based on an implementing of the one of the plurality of intermediate deep learning models.
13. A computing system, the system comprising:
one or more processors; and
a memory storing instructions, which when executed by the one or more processors, configure the one or more processors to:
identify a plurality of connections in a neural network that is pre-associated with a deep learning model;
determine different sets, each set comprising a different combination of one or more of the identified plurality of connections;
generate a plurality of pruned neural networks by implementing an iterative pruning of the determined different sets of one or more of the plurality of connections to respectively generate each of the plurality of pruned neural networks;
generate a plurality of intermediate deep learning models by implementing a generation of a respective intermediate deep learning model corresponding to each of the plurality of pruned neural networks;
determine accuracy levels of each of the plurality of intermediate deep learning models to determine a greatest accuracy among the plurality of intermediate deep learning models; and
select one of the plurality of intermediate deep learning models, having the determined greatest accuracy among the plurality of intermediate deep learning models, as an optimized deep learning model,
wherein a first intermediate deep learning model is generated corresponding to a first pruned neural network and a second intermediate deep learning model is generated corresponding to a second pruned neural network, and
wherein the first intermediate deep learning model has the greater accuracy than the second intermediate deep learning model and is selected as the optimized deep learning model.
14. The system of claim 13 , wherein the one or more processors are configured to perform the pruning of the different sets of the one or of the plurality of connections based on predetermined pruning policies.
15. The system of claim 14 , wherein the one or more processors are configured to update the predetermined pruning policies based on a determined accuracy level of the optimized deep learning model.
16. The system of claim 14 , wherein the predetermined pruning policies comprise at least one of a policy of pruning one or more connections for a predetermined time period or a policy of pruning connections until a threshold number of connections are pruned.
17. The system of claim 14 , wherein, to perform the pruning of the different sets of the one or more of the plurality of connections, the one or more processors are configured to:
select, at random, respective combinations of two or more connections for pruning; and
prune each of the respective combinations based on the predetermined pruning policies.
18. The system of claim 13 , wherein, for the determining of the accuracy levels, the one or more processors are configured to determine the accuracy levels using a predetermined validation technique.
19. The system of claim 18 , wherein the predetermined validation technique comprises a determination of an error level corresponding to each of the plurality of intermediate deep learning models.
20. The system of claim 13 , wherein, for the pruning, the one or more processors are configured to assign a zero value to each weight corresponding to each pruned connection.
21. The system of claim 13 , wherein each of the plurality of intermediate deep learning models is a subset of the deep learning model.
22. The system of claim 21 , wherein a total number of connections in an intermediate deep learning model, of the plurality of intermediate deep learning models, is less than or equal to a total number of connections in the deep learning model.