Hotfix misleading xratio proportion nomenclature (it is train instead of test)

cappelletto · cappelletto · commit 308472f18096 · 2023-07-21T11:51:25.000+01:00
diff --git a/src/bnn_inference/train.py b/src/bnn_inference/train.py
@@ -227,7 +227,7 @@ def train_impl(
     )
 
     X_train, X_valid, y_train, y_valid = train_test_split(
-        X_norm, y_norm, test_size=xratio, shuffle=True  # 8:2 ratio
+        X_norm, y_norm, train_size=xratio, shuffle=True  # 8:2 ratio
     )
     # Convert train and test vectors to tensors
     X_train, y_train = torch.Tensor(X_train).float(), torch.Tensor(y_train).float()

Original file line number	Diff line number	Diff line change
`@@ -227,7 +227,7 @@ def train_impl(`
`227`	`227`	`)`
`228`	`228`
`229`	`229`	`X_train, X_valid, y_train, y_valid = train_test_split(`
`230`		`- X_norm, y_norm, test_size=xratio, shuffle=True # 8:2 ratio`
	`230`	`+ X_norm, y_norm, train_size=xratio, shuffle=True # 8:2 ratio`
`231`	`231`	`)`
`232`	`232`	`# Convert train and test vectors to tensors`
`233`	`233`	`X_train, y_train = torch.Tensor(X_train).float(), torch.Tensor(y_train).float()`