You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@madlib.apache.org by GitBox <gi...@apache.org> on 2021/01/22 00:32:28 UTC
[GitHub] [madlib] fmcquillan99 commented on pull request #531: DL: print timing for evaluate operation in fit multiple
fmcquillan99 commented on pull request #531:
URL: https://github.com/apache/madlib/pull/531#issuecomment-765030529
```
madlib=# SELECT madlib.madlib_keras_fit_multiple_model('iris_train_packed', -- source_table
madlib(# 'iris_multi_model', -- model_output_table
madlib(# 'mst_table', -- model_selection_table
madlib(# 10, -- num_iterations
madlib(# FALSE, -- use gpus
madlib(# 'iris_test_packed', -- validation dataset
madlib(# 1, -- metrics compute frequency
madlib(# FALSE, -- warm start
madlib(# 'Sophie L.', -- name
madlib(# 'Model selection for iris dataset' -- description
madlib(# );
INFO:
Time for training in iteration 1: 5.08531904221 sec
DETAIL:
Training set after iteration 1:
mst_key=1: metric=0.658333361149, loss=1.1474006176
mst_key=3: metric=0.10000000149, loss=1.38732385635
mst_key=2: metric=0.333333343267, loss=1.0933380127
mst_key=4: metric=0.533333361149, loss=1.7163516283
Time for evaluating training dataset in iteration 1: 1.03906393051
Validation set after iteration 1:
mst_key=1: metric=0.699999988079, loss=1.10724759102
mst_key=3: metric=0.166666671634, loss=1.39472186565
mst_key=2: metric=0.333333343267, loss=1.08243787289
mst_key=4: metric=0.533333361149, loss=1.8540776968
Time for evaluating validation dataset in iteration 1: 1.13996315002
CONTEXT: PL/Python function "madlib_keras_fit_multiple_model"
INFO:
Time for training in iteration 2: 3.52112603188 sec
DETAIL:
Training set after iteration 2:
mst_key=1: metric=0.658333361149, loss=1.14587247372
mst_key=3: metric=0.108333334327, loss=1.32700014114
mst_key=2: metric=0.333333343267, loss=1.08806085587
mst_key=4: metric=0.625, loss=1.65616095066
Time for evaluating training dataset in iteration 2: 1.04296994209
Validation set after iteration 2:
mst_key=1: metric=0.699999988079, loss=1.10603606701
mst_key=3: metric=0.166666671634, loss=1.3331912756
mst_key=2: metric=0.333333343267, loss=1.07862055302
mst_key=4: metric=0.566666662693, loss=1.79157197475
Time for evaluating validation dataset in iteration 2: 1.17975687981
etc.
```
```
madlib=# SELECT madlib.madlib_keras_fit_multiple_model('iris_train_packed', -- source_table
madlib(# 'iris_multi_model', -- model_output_table
madlib(# 'mst_table', -- model_selection_table
madlib(# 10, -- num_iterations
madlib(# FALSE, -- use gpus
madlib(# 'iris_test_packed', -- validation dataset
madlib(# 3, -- metrics compute frequency
madlib(# FALSE, -- warm start
madlib(# 'Sophie L.', -- name
madlib(# 'Model selection for iris dataset' -- description
madlib(# );
INFO:
Time for training in iteration 1: 5.23196315765 sec
CONTEXT: PL/Python function "madlib_keras_fit_multiple_model"
INFO:
Time for training in iteration 2: 3.06432795525 sec
CONTEXT: PL/Python function "madlib_keras_fit_multiple_model"
INFO:
Time for training in iteration 3: 3.11480093002 sec
DETAIL:
Training set after iteration 3:
mst_key=1: metric=0.341666668653, loss=1.00513792038
mst_key=3: metric=0.341666668653, loss=1.14255702496
mst_key=2: metric=0.333333343267, loss=1.30739760399
mst_key=4: metric=0.341666668653, loss=1.01052212715
Time for evaluating training dataset in iteration 3: 1.07409715652
Validation set after iteration 3:
mst_key=1: metric=0.300000011921, loss=1.03619933128
mst_key=3: metric=0.300000011921, loss=1.15146791935
mst_key=2: metric=0.333333343267, loss=1.27005898952
mst_key=4: metric=0.300000011921, loss=1.08076310158
Time for evaluating validation dataset in iteration 3: 1.44217610359
etc
```
```
madlib=# SELECT madlib.madlib_keras_fit_multiple_model('iris_train_packed', -- source_table
madlib(# 'iris_multi_model', -- model_output_table
madlib(# 'mst_table', -- model_selection_table
madlib(# 5, -- num_iterations
madlib(# FALSE -- use gpus
madlib(# );
INFO:
Time for training in iteration 1: 3.21357607841 sec
CONTEXT: PL/Python function "madlib_keras_fit_multiple_model"
INFO:
Time for training in iteration 2: 3.24296498299 sec
CONTEXT: PL/Python function "madlib_keras_fit_multiple_model"
INFO:
Time for training in iteration 3: 3.2322010994 sec
CONTEXT: PL/Python function "madlib_keras_fit_multiple_model"
INFO:
Time for training in iteration 4: 3.33785796165 sec
CONTEXT: PL/Python function "madlib_keras_fit_multiple_model"
INFO:
Time for training in iteration 5: 3.1472928524 sec
DETAIL:
Training set after iteration 5:
mst_key=1: metric=0.675000011921, loss=1.01427876949
mst_key=3: metric=0.341666668653, loss=0.958754897118
mst_key=2: metric=0.324999988079, loss=1.13000130653
mst_key=4: metric=0.341666668653, loss=1.0127518177
Time for evaluating training dataset in iteration 5: 1.03426170349
CONTEXT: PL/Python function "madlib_keras_fit_multiple_model"
madlib_keras_fit_multiple_model
---------------------------------
(1 row)
```
3.214 | training in iteration 1
3.243 | training in iteration 2
3.232 | training in iteration 3
3.338 | training in iteration 4
3.147 | training in iteration 5
1.034 | evaluating training dataset in iteration 5
17.21 | total
18.89 | E2E run time
1.681 | other overhead time
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org