Skip to content

Commit da19e71

Browse files
arnocandelarnocandel
committed
Modify learning rate from 0.02 to 0.01.
1 parent 7677958 commit da19e71

6 files changed

Lines changed: 15 additions & 15 deletions

File tree

‎tutorials/deeplearning/.ipynb_checkpoints/deeplearning-checkpoint.ipynb‎

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -5307,7 +5307,7 @@
53075307
" score_validation_samples=10000, # downsample validation set for faster scoring\n",
53085308
" score_duty_cycle=0.025, # don't score more than 2.5% of the wall time\n",
53095309
" adaptive_rate=False, # manually tuned learning rate\n",
5310-
" rate=0.02, \n",
5310+
" rate=0.01, \n",
53115311
" rate_annealing=0.000002, \n",
53125312
" momentum_start=0.2, # manually tuned momentum\n",
53135313
" momentum_stable=0.4, \n",
@@ -5776,7 +5776,7 @@
57765776
"name": "python",
57775777
"nbconvert_exporter": "python",
57785778
"pygments_lexer": "ipython2",
5779-
"version": "2.7.10"
5779+
"version": "2.7.8"
57805780
}
57815781
},
57825782
"nbformat": 4,

‎tutorials/deeplearning/README.md‎

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -51,7 +51,7 @@ We start with a small dataset representing red and black dots on a plane, arrang
5151
We visualize the nature of H2O Deep Learning (DL), H2O's tree methods (GBM/DRF) and H2O's generalized linear modeling (GLM) by plotting the decision boundary between the red and black spirals:
5252

5353
```r
54-
#setwd("path/to/tutorials/") ##For RStudio
54+
#setwd("~/h2o-world-2015-training/tutorials") ##For RStudio
5555
spiral <- h2o.importFile(path = normalizePath("../data/spiral.csv"))
5656
grid <- h2o.importFile(path = normalizePath("../data/grid.csv"))
5757
# Define helper to plot contours
@@ -227,7 +227,7 @@ m3 <- h2o.deeplearning(
227227
score_validation_samples=10000, ## downsample validation set for faster scoring
228228
score_duty_cycle=0.025, ## don't score more than 2.5% of the wall time
229229
adaptive_rate=F, ## manually tuned learning rate
230-
rate=0.02,
230+
rate=0.01,
231231
rate_annealing=2e-6,
232232
momentum_start=0.2, ## manually tuned momentum
233233
momentum_stable=0.4,
@@ -397,7 +397,7 @@ m_cont <- h2o.deeplearning(
397397
score_validation_samples=10000, ## downsample validation set for faster scoring
398398
score_duty_cycle=0.025, ## don't score more than 2.5% of the wall time
399399
adaptive_rate=F, ## manually tuned learning rate
400-
rate=0.02,
400+
rate=0.01,
401401
rate_annealing=2e-6,
402402
momentum_start=0.2, ## manually tuned momentum
403403
momentum_stable=0.4,

‎tutorials/deeplearning/deeplearning.R‎

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -44,7 +44,7 @@ example(h2o.deeplearning)
4444
#
4545
#We visualize the nature of H2O Deep Learning (DL), H2O's tree methods (GBM/DRF) and H2O's generalized linear modeling (GLM) by plotting the decision boundary between the red and black spirals:
4646
#
47-
#setwd("path/to/tutorials/") ##For RStudio
47+
#setwd("~/h2o-world-2015-training/tutorials") ##For RStudio
4848
spiral <- h2o.importFile(path = normalizePath("../data/spiral.csv"))
4949
grid <- h2o.importFile(path = normalizePath("../data/grid.csv"))
5050
# Define helper to plot contours
@@ -198,7 +198,7 @@ m3 <- h2o.deeplearning(
198198
score_validation_samples=10000, ## downsample validation set for faster scoring
199199
score_duty_cycle=0.025, ## don't score more than 2.5% of the wall time
200200
adaptive_rate=F, ## manually tuned learning rate
201-
rate=0.02,
201+
rate=0.01,
202202
rate_annealing=2e-6,
203203
momentum_start=0.2, ## manually tuned momentum
204204
momentum_stable=0.4,
@@ -352,7 +352,7 @@ m_cont <- h2o.deeplearning(
352352
score_validation_samples=10000, ## downsample validation set for faster scoring
353353
score_duty_cycle=0.025, ## don't score more than 2.5% of the wall time
354354
adaptive_rate=F, ## manually tuned learning rate
355-
rate=0.02,
355+
rate=0.01,
356356
rate_annealing=2e-6,
357357
momentum_start=0.2, ## manually tuned momentum
358358
momentum_stable=0.4,

‎tutorials/deeplearning/deeplearning.Rmd‎

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -51,7 +51,7 @@ We start with a small dataset representing red and black dots on a plane, arrang
5151
We visualize the nature of H2O Deep Learning (DL), H2O's tree methods (GBM/DRF) and H2O's generalized linear modeling (GLM) by plotting the decision boundary between the red and black spirals:
5252

5353
```{r setup_spiral}
54-
#setwd("path/to/tutorials/") ##For RStudio
54+
#setwd("~/h2o-world-2015-training/tutorials") ##For RStudio
5555
spiral <- h2o.importFile(path = normalizePath("../data/spiral.csv"))
5656
grid <- h2o.importFile(path = normalizePath("../data/grid.csv"))
5757
# Define helper to plot contours
@@ -227,7 +227,7 @@ m3 <- h2o.deeplearning(
227227
score_validation_samples=10000, ## downsample validation set for faster scoring
228228
score_duty_cycle=0.025, ## don't score more than 2.5% of the wall time
229229
adaptive_rate=F, ## manually tuned learning rate
230-
rate=0.02,
230+
rate=0.01,
231231
rate_annealing=2e-6,
232232
momentum_start=0.2, ## manually tuned momentum
233233
momentum_stable=0.4,
@@ -397,7 +397,7 @@ m_cont <- h2o.deeplearning(
397397
score_validation_samples=10000, ## downsample validation set for faster scoring
398398
score_duty_cycle=0.025, ## don't score more than 2.5% of the wall time
399399
adaptive_rate=F, ## manually tuned learning rate
400-
rate=0.02,
400+
rate=0.01,
401401
rate_annealing=2e-6,
402402
momentum_start=0.2, ## manually tuned momentum
403403
momentum_stable=0.4,

‎tutorials/deeplearning/deeplearning.ipynb‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -5307,7 +5307,7 @@
53075307
" score_validation_samples=10000, # downsample validation set for faster scoring\n",
53085308
" score_duty_cycle=0.025, # don't score more than 2.5% of the wall time\n",
53095309
" adaptive_rate=False, # manually tuned learning rate\n",
5310-
" rate=0.02, \n",
5310+
" rate=0.01, \n",
53115311
" rate_annealing=0.000002, \n",
53125312
" momentum_start=0.2, # manually tuned momentum\n",
53135313
" momentum_stable=0.4, \n",

‎tutorials/deeplearning/deeplearning.md‎

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -51,7 +51,7 @@ We start with a small dataset representing red and black dots on a plane, arrang
5151
We visualize the nature of H2O Deep Learning (DL), H2O's tree methods (GBM/DRF) and H2O's generalized linear modeling (GLM) by plotting the decision boundary between the red and black spirals:
5252

5353
```r
54-
#setwd("path/to/tutorials/") ##For RStudio
54+
#setwd("~/h2o-world-2015-training/tutorials") ##For RStudio
5555
spiral <- h2o.importFile(path = normalizePath("../data/spiral.csv"))
5656
grid <- h2o.importFile(path = normalizePath("../data/grid.csv"))
5757
# Define helper to plot contours
@@ -227,7 +227,7 @@ m3 <- h2o.deeplearning(
227227
score_validation_samples=10000, ## downsample validation set for faster scoring
228228
score_duty_cycle=0.025, ## don't score more than 2.5% of the wall time
229229
adaptive_rate=F, ## manually tuned learning rate
230-
rate=0.02,
230+
rate=0.01,
231231
rate_annealing=2e-6,
232232
momentum_start=0.2, ## manually tuned momentum
233233
momentum_stable=0.4,
@@ -397,7 +397,7 @@ m_cont <- h2o.deeplearning(
397397
score_validation_samples=10000, ## downsample validation set for faster scoring
398398
score_duty_cycle=0.025, ## don't score more than 2.5% of the wall time
399399
adaptive_rate=F, ## manually tuned learning rate
400-
rate=0.02,
400+
rate=0.01,
401401
rate_annealing=2e-6,
402402
momentum_start=0.2, ## manually tuned momentum
403403
momentum_stable=0.4,

0 commit comments

Comments
 (0)