Reproduce with N to 10000 and minutes to 10

Author: Jamesrsum

dvc.lock

@@ -271,11 +271,11 @@ stages:
       nfiles: 0
   cgpm-infer-hyperparameters:
     cmd: "find data/cgpm/hydrated -type f | parallel jsonschema --instance {} schemas/cgpm.json
-      && mkdir -p data/cgpm/complete && echo 1 >> data/cgpm/inf.log && find data/cgpm/hydrated
+      && mkdir -p data/cgpm/complete && echo 10 >> data/cgpm/inf.log && find data/cgpm/hydrated
       -type f | sort | parallel --verbose 'python scripts/cgpm_infer.py {} --kernel
       alpha --kernel view_alphas --kernel column_hypers --kernel rows --kernel columns
       --output data/cgpm/complete/{/} --data data/numericalized.csv --params params.yaml
-      --seed $((0 + {#} - 1)) --minutes 1' #--iterations 1\n"
+      --seed $((0 + {#} - 1)) --minutes 10' #--iterations 1\n"
     deps:
     - path: data/cgpm/hydrated
       hash: md5
@@ -298,7 +298,7 @@ stages:
       params.yaml:
         cgpm:
           iterations: 1
-          minutes: 1
+          minutes: 10
         parallel.flags: --verbose
         seed: 0
     outs:

params.yaml

@@ -47,7 +47,7 @@ loom:
   extra_passes: 100
 cgpm:
   iterations: 1
-  minutes: 1
+  minutes: 10
   #dependence:
   #  # While the CrossCat implementation in CGPM takes dependence constraints, doing so
   #  # throws a not-implemented-error. Hence, we apply a workaround: