OPTIMIZATION OF TUNING FOR MODELS USED FOR MULTIPLE PREDICTION GENERATION
The present application discloses a method, system, and computer system for tuning a set of models. The method includes determining a set of one or more models to optimize, determining a plurality of optimizer modules with which to optimize the set of one or more models, causing the plurality of optimizer modules to respectively perform a respective optimizing process with respect to at least one model of the set of one or more models, and deploying an optimized model obtained based at least in part on optimizing metrics of the set of the one or more models.
1 . A system, comprising:
one or more processors configured to:
determine a set of one or more models to optimize;
determine a plurality of optimizer modules with which to optimize the set of the one or more models;
cause the plurality of optimizer modules to respectively perform a respective optimizing process with respect to at least one model of the set of the one or more models; and
deploy an optimized model obtained based at least in part on optimizing metrics of the set of the one or more models; and
a memory coupled to the one or more processors and configured to provide the one or more processors with instructions.
2 . The system of claim 1 , wherein each of the plurality of optimizer modules is run on a different thread or compute node.
3 . The system of claim 1 , wherein the set of the one or more models are optimized according to a predetermined time interval.
4 . The system of claim 1 , wherein optimizing the set of the one or more models comprises:
obtaining a respective starting set of parameters with which to optimize the set of the one or more models; and
running, by each of the plurality of optimizer modules, the respective optimizing process in connection with optimizing, based at least in part on the respective starting set of parameters, a particular model of the set of one or more models to optimize.
5 . The system of claim 4 , wherein:
at least two of the plurality of optimizer modules optimize a same model selected from among the plurality of respective models to be optimized; and
the at least two of the plurality of optimizer modules optimize the same model using a different starting set of parameters.
6 . The system of claim 5 , wherein the at least two of the plurality of optimizer modules optimize the same model in parallel.
7 . The system of claim 6 , an optimized model is selected from among a set of optimized models that is optimized by the plurality of optimizer modules.
8 . The system of claim 4 , wherein each of the plurality of optimizer modules is run on a different thread or compute node.
9 . The system of claim 4 , wherein running the respective optimizing process includes performing at least a number of iterations of training the particular model of the plurality of respective models to be optimized.
10 . The system of claim 4 , wherein a training dataset used in connection with optimizing the particular model is stored in cache while a particular optimizer module runs the respective optimization process.
11 . The system of claim 1 , wherein determining the plurality of optimizer modules comprises:
determining a set of available threads; and
creating a pool of threads with which to implement the plurality of optimizer modules.
12 . The system of claim 12 , wherein:
the plurality of optimizer modules are respectively implemented on different threads from the pool of threads; and
each optimizer module iterates over a tuning process within an associated thread to tune a respective model of the set of one or more models, and an optimized model is selected from among a set of tuned models comprising models obtained at each iteration of the tuning process.
13 . The system of claim 12 , wherein:
a same model is tuned using different optimizer modules; and
at least two different optimizer modules that tune the same model use different optimizing processes.
14 . The system of claim 13 , wherein at least two different optimizer modules that tune the same model use different starting sets of parameters.
15 . The system of claim 1 , wherein the at least one of the one or more models is stored in cache during optimization with respect to at least one model.
16 . The system of claim 1 , wherein at least a subset of the set of the one or more models are obtained and stored in a cache in response to determining to optimize the subset of the set of the one or more models.
17 . The system of claim 1 , wherein:
at least two of the plurality of optimizer modules optimize different models associated with a same expected prediction; and
the optimized model deployed for the expected prediction is selected from among the different models associated with the same expected prediction.
18 . A method, comprising:
determining, by one or more processors, a set of one or more models to optimize;
determining a plurality of optimizer modules with which to optimize the set of the one or more models;
causing the plurality of optimizer modules to respectively perform a respective optimizing process with respect to at least one model of the set of the one or more models; and
deploying an optimized model obtained based at least in part on optimizing metrics of the set of the one or more models.
19 . A computer program product embodied in a non-transitory computer readable medium and comprising computer instructions for:
determining, by one or more processors, a set of one or more models to optimize;
determining a plurality of optimizer modules with which to optimize the set of the one or more models;
causing the plurality of optimizer modules to respectively perform a respective optimizing process with respect to at least one model of the set of the one or more models; and
deploying an optimized model obtained based at least in part on optimizing metrics of the set of the one or more models.