fix wrong usage of n_montecarlo in inference tests

Author: Red-Portal

test/inference/repgradelbo_distributionsad.jl

@@ -15,9 +15,9 @@ end
                                                                      [Float64, Float32],
         (modelname, modelconstr) in Dict(:Normal => normal_meanfield),
         (objname, objective) in Dict(
-            :RepGradELBOClosedFormEntropy => RepGradELBO(n_montecarlo),
+            :RepGradELBOClosedFormEntropy => RepGradELBO(10),
             :RepGradELBOStickingTheLanding =>
-                RepGradELBO(n_montecarlo; entropy=StickingTheLandingEntropy()),
+                RepGradELBO(10; entropy=StickingTheLandingEntropy()),
         ),
         (adbackname, adtype) in AD_repgradelbo_distributionsad
 
@@ -30,7 +30,6 @@ end
         T = 1000
         η = 1e-3
         opt = Optimisers.Descent(realtype(η))
-        n_montecarlo = 10
 
         # For small enough η, the error of SGD, Δλ, is bounded as
         #     Δλ ≤ ρ^T Δλ0 + O(η),

test/inference/repgradelbo_locationscale.jl

@@ -16,9 +16,9 @@ end
         (modelname, modelconstr) in
         Dict(:Normal => normal_meanfield, :Normal => normal_fullrank),
         (objname, objective) in Dict(
-            :RepGradELBOClosedFormEntropy => RepGradELBO(n_montecarlo),
+            :RepGradELBOClosedFormEntropy => RepGradELBO(10),
             :RepGradELBOStickingTheLanding =>
-                RepGradELBO(n_montecarlo; entropy=StickingTheLandingEntropy()),
+                RepGradELBO(10; entropy=StickingTheLandingEntropy()),
         ),
         (adbackname, adtype) in AD_repgradelbo_locationscale
 
@@ -31,7 +31,6 @@ end
         T = 1000
         η = 1e-3
         opt = Optimisers.Descent(realtype(η))
-        n_montecarlo = 10
 
         # For small enough η, the error of SGD, Δλ, is bounded as
         #     Δλ ≤ ρ^T Δλ0 + O(η),

test/inference/repgradelbo_locationscale_bijectors.jl

@@ -16,9 +16,9 @@ end
         (modelname, modelconstr) in
         Dict(:NormalLogNormalMeanField => normallognormal_meanfield),
         (objname, objective) in Dict(
-            :RepGradELBOClosedFormEntropy => RepGradELBO(n_montecarlo),
+            :RepGradELBOClosedFormEntropy => RepGradELBO(10),
             :RepGradELBOStickingTheLanding =>
-                RepGradELBO(n_montecarlo; entropy=StickingTheLandingEntropy()),
+                RepGradELBO(10; entropy=StickingTheLandingEntropy()),
         ),
         (adbackname, adtype) in AD_repgradelbo_locationscale_bijectors
 
@@ -31,7 +31,6 @@ end
         T = 1000
         η = 1e-3
         opt = Optimisers.Descent(realtype(η))
-        n_montecarlo = 10
 
         b = Bijectors.bijector(model)
         b⁻¹ = inverse(b)

test/inference/scoregradelbo_distributionsad.jl

@@ -14,7 +14,7 @@ end
     @testset "$(modelname) $(objname) $(realtype) $(adbackname)" for realtype in
                                                                      [Float64, Float32],
         (modelname, modelconstr) in Dict(:Normal => normal_meanfield),
-        (objname, objective) in Dict(:ScoreGradELBO => ScoreGradELBO(n_montecarlo)),
+        (objname, objective) in Dict(:ScoreGradELBO => ScoreGradELBO(10)),
         (adbackname, adtype) in AD_scoregradelbo_distributionsad
 
         seed = (0x38bef07cf9cc549d)
@@ -26,7 +26,6 @@ end
         T = 1000
         η = 1e-4
         opt = Optimisers.Descent(realtype(η))
-        n_montecarlo = 10
 
         # For small enough η, the error of SGD, Δλ, is bounded as
         #     Δλ ≤ ρ^T Δλ0 + O(η),

test/inference/scoregradelbo_locationscale.jl

@@ -15,7 +15,7 @@ end
                                                                      [Float64, Float32],
         (modelname, modelconstr) in
         Dict(:Normal => normal_meanfield, :Normal => normal_fullrank),
-        (objname, objective) in Dict(:ScoreGradELBO => ScoreGradELBO(n_montecarlo)),
+        (objname, objective) in Dict(:ScoreGradELBO => ScoreGradELBO(10)),
         (adbackname, adtype) in AD_scoregradelbo_locationscale
 
         seed = (0x38bef07cf9cc549d)
@@ -27,7 +27,6 @@ end
         T = 1000
         η = 1e-4
         opt = Optimisers.Descent(realtype(η))
-        n_montecarlo = 10
 
         # For small enough η, the error of SGD, Δλ, is bounded as
         #     Δλ ≤ ρ^T Δλ0 + O(η),

test/inference/scoregradelbo_locationscale_bijectors.jl

@@ -15,7 +15,7 @@ end
                                                                      [Float64, Float32],
         (modelname, modelconstr) in
         Dict(:NormalLogNormalMeanField => normallognormal_meanfield),
-        (objname, objective) in Dict(:ScoreGradELBO => ScoreGradELBO(n_montecarlo)),
+        (objname, objective) in Dict(:ScoreGradELBO => ScoreGradELBO(10)),
         (adbackname, adtype) in AD_scoregradelbo_locationscale_bijectors
 
         seed = (0x38bef07cf9cc549d)
@@ -27,7 +27,6 @@ end
         T = 1000
         η = 1e-4
         opt = Optimisers.Descent(realtype(η))
-        n_montecarlo = 10
 
         b = Bijectors.bijector(model)
         b⁻¹ = inverse(b)