Merge branch 'master' of github.com:0xdata/h2o

subramaniano · Aug 7, 2014 · adca60f · adca60f
2 parents 67e6ae6 + 194c951
commit adca60f
Show file tree

Hide file tree

Showing 11 changed files with 27 additions and 25 deletions.
diff --git a/py/testdir_single_jvm/test_GLM2_tweedie_rand2.py b/py/testdir_single_jvm/test_GLM2_tweedie_rand2.py
@@ -37,7 +37,7 @@ def setUpClass(cls):
     def tearDownClass(cls):
         h2o.tear_down_cloud()
 
-    def test_GLM2_tweedie_rand2(self):
+    def test_NOPASS_GLM2_tweedie_rand2(self):
         h2o.beta_features = True
         if 1==1:
             csvPathname = 'standard/covtype.data'

diff --git a/py/testdir_single_jvm/test_GLM2_weight_nan_fail.py b/py/testdir_single_jvm/test_GLM2_weight_nan_fail.py
@@ -20,7 +20,7 @@ def setUpClass(cls):
     def tearDownClass(cls):
         h2o.tear_down_cloud()
 
-    def test_GLM2_weight_nan_fail(self):
+    def test_NOPASS_GLM2_weight_nan_fail(self):
         h2o.beta_features = True
         csvPathname = 'covtype/covtype.20k.data'
         hex_key = 'covtype.20k.hex'

diff --git a/py/testdir_single_jvm/test_exec2_row_range.py b/py/testdir_single_jvm/test_exec2_row_range.py
@@ -84,7 +84,7 @@ def test_exec2_row_range(self):
 
                 execExpr = "%s=%s[1:%s,]" % (hex_key_i, hex_key, rowCount-10)
                 resultExec, result = h2e.exec_expr(execExpr=execExpr, timeoutSecs=30)
-                inspect = h2o_cmd.runInspect(None, hex_key_i, timeoutSecs=timeoutSecs2)
+                inspect = h2o_cmd.runInspect(None, hex_key_i, timeoutSecs=timeoutSecs)
                 h2o_cmd.infoFromInspect(inspect, hex_key_i)
                 print "\n" + hex_key_i, \
                     "    numRows:", "{:,}".format(inspect['numRows']), \

diff --git a/py/testdir_single_jvm/test_exec2_ternary_phrases.py b/py/testdir_single_jvm/test_exec2_ternary_phrases.py
@@ -139,20 +139,17 @@ def test_exec2_ternary_phrases(self):
             h2e.exec_expr(h2o.nodes[0], execExpr, resultKey=None, timeoutSecs=4)
 
         # now do some double concats of the expressions created
-        for j in range (50):
-            execExpr = "a=" + random.choice(bigExprList) + "; b= " + random.choice(bigExprList)
-            # don't allow double ternary
-            if re.search("\?.*\?", execExpr): 
-                break
-            h2e.exec_expr(h2o.nodes[0], execExpr, resultKey=None, timeoutSecs=4)
+        #for j in range (50):
+        #    execExpr = "a=" + random.choice(bigExprList) + "; b= " + random.choice(bigExprList)
+        #    # don't allow double ternary
+        #    if re.search("\?.*\?", execExpr): 
+        #        break
+        #    h2e.exec_expr(h2o.nodes[0], execExpr, resultKey=None, timeoutSecs=4)
 
         # now do some triple concats of the expressions created
-        for j in range (50):
-            execExpr = "a=" + random.choice(bigExprList) + "; b= " + random.choice(bigExprList) + "; c=" + random.choice(bigExprList)
-            h2e.exec_expr(h2o.nodes[0], execExpr, resultKey=None, timeoutSecs=4)
+        #for j in range (50):
+        #    execExpr = "a=" + random.choice(bigExprList) + "; b= " + random.choice(bigExprList) + "; c=" + random.choice(bigExprList)
+        #    h2e.exec_expr(h2o.nodes[0], execExpr, resultKey=None, timeoutSecs=4)
 
-
-
-
 if __name__ == '__main__':
     h2o.unit_main()
diff --git a/py/testdir_single_jvm/test_mixed_causes_NA.py b/py/testdir_single_jvm/test_mixed_causes_NA.py
@@ -19,7 +19,7 @@ def setUpClass(cls):
     def tearDownClass(cls):
         h2o.tear_down_cloud()
 
-    def test_mixed_causes_NA(self):
+    def test_NOPASS_mixed_causes_NA(self):
         csvFilename = 'mixed_causes_NA.csv'
         parseResult = h2i.import_parse(bucket='smalldata', path=csvFilename, timeoutSecs=15, schema='put')
         inspect = h2o_cmd.runInspect(None, parseResult['destination_key'])

diff --git a/py/testdir_single_jvm/test_mixed_int_enum.py b/py/testdir_single_jvm/test_mixed_int_enum.py
@@ -52,7 +52,7 @@ def setUpClass(cls):
     def tearDownClass(cls):
         h2o.tear_down_cloud()
 
-    def test_mixed_int_enum(self):
+    def test_NOPASS_mixed_int_enum(self):
         h2o.beta_features = True
         SYNDATASETS_DIR = h2o.make_syn_dir()
 

diff --git a/py/testdir_single_jvm/test_parse_500_cols_spill_fvec.py b/py/testdir_single_jvm/test_parse_500_cols_spill_fvec.py
@@ -43,7 +43,7 @@ def tearDownClass(cls):
         # h2o.sleep(3600)
         h2o.tear_down_cloud()
 
-    def test_parse_500_cols_fvec(self):
+    def test_NOPASS_parse_500_cols_fvec(self):
         h2o.beta_features = True
         SYNDATASETS_DIR = h2o.make_syn_dir()
         tryList = [

diff --git a/py/testdir_single_jvm_fvec/test_speedrf_float_bigexp.py b/py/testdir_single_jvm_fvec/test_speedrf_float_bigexp.py
@@ -89,12 +89,12 @@ def test_rf_float_bigexp_fvec(self):
             ntree = 2
             kwargs = {
                 'response': 'AGE',
-                'num_trees': ntree,
-                'mtry': None,
+                'ntrees': ntree,
+                'mtries': None,
                 'max_depth': 20,
-                'sample': 0.67,
+                'sample_rate': 0.67,
                 'destination_key': None,
-                'bin_limit': 1024,
+                'nbins': 1024,
                 'seed': 784834182943470027,
             }
             parseResult = h2i.import_parse(path=csvPathname, schema='put', hex_key=hex_key, doSummary=True)

diff --git a/src/main/java/hex/deeplearning/DeepLearningTask.java b/src/main/java/hex/deeplearning/DeepLearningTask.java
@@ -44,7 +44,12 @@ private DeepLearningTask(hex.deeplearning.DeepLearningModel.DeepLearningModelInf
 
   @Override public final void processRow(long seed, final double [] nums, final int numcats, final int [] cats, double [] responses){
     if(_output.get_params().self() != null && !Job.isRunning(_output.get_params().self())) throw new Job.JobCancelledException();
-    seed = new Random().nextLong(); //no point in being reproducible - better to be "good" at being random
+//    if (H2O.CLOUD.size()==1) {
+//      seed += model_info().get_processed_global(); //avoid periodicity
+//    } else {
+//      seed = new Random().nextLong(); //multi-node: no point in being reproducible - better to be "good" at being random
+//    }
+    seed = new Random().nextLong();
     ((Neurons.Input)_neurons[0]).setInput(seed, nums, numcats, cats);
     step(seed, _neurons, _output, _training, responses);
   }

diff --git a/src/main/java/hex/glm/GLMModelView.java b/src/main/java/hex/glm/GLMModelView.java
@@ -89,7 +89,7 @@ public static Response redirect(Request req, Key modelKey, Key job_key) {
           firstRow.append("\t\t<td><b>" + DFORMAT2.format(glm_model.submodels[i].lambda_value) + "</b></td>\n");
         else
           firstRow.append("\t\t<td>" + link(DFORMAT2.format(glm_model.submodels[i].lambda_value), glm_model._key, glm_model.submodels[i].lambda_value) + "</td>\n");
-        secondRow.append("\t\t<td>" + (sm.rank - 1) + "</td>\n");
+        secondRow.append("\t\t<td>" + Math.max(0,(sm.rank - 1)) + "</td>\n"); // rank counts intercept, that's why -1 is there, however, intercept can be 0 as well, so just prevent -1
         if(sm.xvalidation != null){
           thirdRow.append("\t\t<td>"  + DFORMAT.format(1 - sm.xvalidation.residual_deviance / sm.xvalidation.null_deviance) + "<sub>x</sub>(" + DFORMAT.format(1 - sm.validation.residual_deviance / sm.validation.null_deviance) + ")" + "</td>\n");
           fourthRow.append("\t\t<td>" + DFORMAT.format(glm_model.glm.family == Family.binomial ? sm.xvalidation.auc : sm.xvalidation.aic) + "<sub>x</sub>("+ DFORMAT.format(glm_model.glm.family == Family.binomial ? sm.validation.auc : sm.validation.aic) + ")</td>\n");

diff --git a/src/test/java/hex/DeepLearningSpiralsTest.java b/src/test/java/hex/DeepLearningSpiralsTest.java
@@ -47,7 +47,7 @@ public class DeepLearningSpiralsTest extends TestUtil {
           p.source = frame;
           p.response = frame.lastVec();
           p.validation = null;
-          p.score_interval = 10;
+          p.score_interval = 2;
           p.ignored_cols = null;
           p.train_samples_per_iteration = 0; //sync once per period
           p.quiet_mode = true;