diff --git a/Sources/DeepLearning/Operators.swift b/Sources/DeepLearning/Operators.swift
index 51b2406a8..827890b3d 100644
--- a/Sources/DeepLearning/Operators.swift
+++ b/Sources/DeepLearning/Operators.swift
@@ -98,12 +98,15 @@ public extension Tensor where Scalar: BinaryFloatingPoint {
 }
 
 //===------------------------------------------------------------------------------------------===//
-// Convolution and pooling
+// Convolution and Pooling
 //===------------------------------------------------------------------------------------------===//
 
 /// A padding scheme. Used by padding, convolution, and pooling ops.
 // @_frozen // SR-9739
 public enum Padding {
+    /// The "explicit" padding scheme, which is defined by an array indicating the explicit padding 
+    /// sizes at the start and end of each dimension.
+    case explicit([Int32])
     /// The "valid" padding scheme.
     case valid
     /// The "same" padding scheme.
@@ -112,12 +115,22 @@ public enum Padding {
 
 public extension Padding {
     @inlinable
-    var raw: Raw.Padding {
+    var raw: Raw.Padding2 {
         switch self {
+        case .explicit: return .explicit
         case .same: return .same
         case .valid: return .valid
         }
     }
+
+    @inlinable
+    internal var explicitPaddings: [Int32] {
+        switch self {
+        case .explicit(let paddings): return paddings
+        case .same: return []
+        case .valid: return []
+        }
+    }
 }
 
 public extension Tensor where Scalar: TensorFlowFloatingPoint {
@@ -135,7 +148,8 @@ public extension Tensor where Scalar: TensorFlowFloatingPoint {
             filter: filter,
             outBackprop: self,
             strides: [strides.0, strides.1, strides.2, strides.3],
-            padding: padding.raw)
+            padding: padding.raw,
+            explicitPaddings: padding.explicitPaddings)
     }
 
     /// TensorFlow builtin conv2d gradient helper for the filter.
@@ -152,7 +166,8 @@ public extension Tensor where Scalar: TensorFlowFloatingPoint {
             filterSizes: filterSizes,
             outBackprop: self,
             strides: [strides.0, strides.1, strides.2, strides.3],
-            padding: padding.raw)
+            padding: padding.raw,
+            explicitPaddings: padding.explicitPaddings)
     }
 
     @inlinable
@@ -282,7 +297,8 @@ public extension Tensor where Scalar: FloatingPoint {
             self,
             filter: filter,
             strides: [strides.0, strides.1, strides.2, strides.3],
-            padding: padding.raw)
+            padding: padding.raw,
+            explicitPaddings: padding.explicitPaddings)
     }
 
     /// Computes a 2-D max pooling, with the specified kernel sizes, strides, and