docs(neuralnetworks): convert hotfix rationale to code comments

CodeRabbit nitpicks (PRRT_kwDOKSXUF858Ek3i UNet3D, PRRT_kwDOKSXUF858Ek3m
VGG, same pattern across the other 9): the XML <summary> blocks I added
next to each Predict override duplicated the existing public API doc and
leaked compiled-replay / tracer internals into IntelliSense / published
docs. Converted all 11 to plain `// Hotfix:` comments so only one
user-facing XML summary remains per method, and pointed the
revert-me note at the upstream root-cause issue
(ooples/AiDotNet.Tensors#228).

Files: ResNetNetwork, VGGNetwork, MobileNetV2Network, EfficientNetNetwork,
ConvolutionalNeuralNetwork, FastText, GloVe, Word2Vec,
SiameseNeuralNetwork, UNet3D, VoxelCNN.

Build: 0 errors.

Author: ooples

src/NeuralNetworks/ConvolutionalNeuralNetwork.cs

@@ -234,13 +234,11 @@ public override void UpdateParameters(Vector<T> parameters)
     /// about what's in the image.
     /// </para>
     /// </remarks>
-    /// <summary>
-    /// Runs <see cref="Forward"/> directly rather than routing through the
-    /// compiled-replay path. <see cref="Forward"/> contains shape-conditional
-    /// control flow that the tracer does not capture faithfully — routing
-    /// through <see cref="NeuralNetworkBase{T}.PredictCompiled"/> was observed
-    /// to return an intermediate feature-map shape instead of the final logits.
-    /// </summary>
+    // Hotfix: bypass the compiled-replay path in NeuralNetworkBase.Predict.
+    // Forward has shape-conditional control flow the tracer does not capture
+    // correctly — routing through PredictCompiled returned an intermediate
+    // feature-map shape instead of the final logits. Tracked upstream at
+    // ooples/AiDotNet.Tensors#228; remove this override once that lands.
     public override Tensor<T> Predict(Tensor<T> input) => Forward(input);
 
     /// <summary>

src/NeuralNetworks/EfficientNetNetwork.cs

@@ -293,13 +293,11 @@ public Tensor<T> Forward(Tensor<T> input)
     }
 
     /// <inheritdoc />
-    /// <summary>
-    /// Runs <see cref="Forward"/> directly rather than routing through the
-    /// compiled-replay path. <see cref="Forward"/> contains shape-conditional
-    /// control flow that the tracer does not capture faithfully — routing
-    /// through <see cref="NeuralNetworkBase{T}.PredictCompiled"/> was observed
-    /// to return an intermediate feature-map shape instead of the final logits.
-    /// </summary>
+    // Hotfix: bypass the compiled-replay path in NeuralNetworkBase.Predict.
+    // Forward has shape-conditional control flow the tracer does not capture
+    // correctly — routing through PredictCompiled returned an intermediate
+    // feature-map shape instead of the final logits. Tracked upstream at
+    // ooples/AiDotNet.Tensors#228; remove this override once that lands.
     public override Tensor<T> Predict(Tensor<T> input) => Forward(input);
 
     /// <inheritdoc />

src/NeuralNetworks/FastText.cs

@@ -244,13 +244,12 @@ public override void UpdateParameters(Vector<T> parameters)
             }
         }
 
-        /// <summary>
-        /// Runs <see cref="Forward"/> directly rather than routing through the
-        /// compiled-replay path. <see cref="Forward"/> contains shape-conditional
-        /// control flow that the tracer does not capture faithfully — routing
-        /// through <see cref="NeuralNetworkBase{T}.PredictCompiled"/> was observed
-        /// to return an intermediate feature-map shape instead of the final output.
-        /// </summary>
+        // Hotfix: bypass the compiled-replay path in NeuralNetworkBase.Predict.
+        // Forward has shape-conditional control flow the tracer does not
+        // capture correctly — routing through PredictCompiled returned an
+        // intermediate shape instead of the final output. Tracked upstream
+        // at ooples/AiDotNet.Tensors#228; remove this override once that
+        // lands.
         public override Tensor<T> Predict(Tensor<T> input) => Forward(input);
 
         /// <summary>

src/NeuralNetworks/GloVe.cs

@@ -293,13 +293,12 @@ public override void UpdateParameters(Vector<T> parameters)
         /// <b>For Beginners:</b> This is identical to the Forward pass—it takes word IDs and 
         /// returns their addresses (embeddings).
         /// </remarks>
-        /// <summary>
-        /// Runs <see cref="Forward"/> directly rather than routing through the
-        /// compiled-replay path. <see cref="Forward"/> contains shape-conditional
-        /// control flow that the tracer does not capture faithfully — routing
-        /// through <see cref="NeuralNetworkBase{T}.PredictCompiled"/> was observed
-        /// to return an intermediate shape instead of the final output.
-        /// </summary>
+        // Hotfix: bypass the compiled-replay path in NeuralNetworkBase.Predict.
+        // Forward has shape-conditional control flow the tracer does not
+        // capture correctly — routing through PredictCompiled returned an
+        // intermediate shape instead of the final embedding output. Tracked
+        // upstream at ooples/AiDotNet.Tensors#228; remove this override once
+        // that lands.
         public override Tensor<T> Predict(Tensor<T> input) => Forward(input);
 
         /// <summary>

src/NeuralNetworks/MobileNetV2Network.cs

@@ -288,13 +288,12 @@ public override Dictionary<string, Tensor<T>> GetNamedLayerActivations(Tensor<T>
     }
 
     /// <inheritdoc />
-    /// <summary>
-    /// Runs <see cref="Forward"/> directly rather than routing through the
-    /// compiled-replay path. <see cref="Forward"/> contains shape-conditional
-    /// control flow that the tracer does not capture faithfully — routing
-    /// through <see cref="NeuralNetworkBase{T}.PredictCompiled"/> was observed
-    /// to return an intermediate feature-map shape instead of the final logits.
-    /// </summary>
+    // Hotfix: bypass the compiled-replay path in NeuralNetworkBase.Predict.
+    // Forward has shape-conditional control flow (3D → 4D promotion +
+    // trailing squeeze) that the tracer does not capture correctly — routing
+    // through PredictCompiled returned an intermediate feature-map shape
+    // instead of the final logits. Tracked upstream at
+    // ooples/AiDotNet.Tensors#228; remove this override once that lands.
     public override Tensor<T> Predict(Tensor<T> input) => Forward(input);
 
     /// <inheritdoc />

src/NeuralNetworks/ResNetNetwork.cs

@@ -493,17 +493,17 @@ public override void UpdateParameters(Vector<T> parameters)
     }
 
     /// <summary>
-    /// Makes a prediction using the ResNet network for the given input. Runs
-    /// <see cref="Forward"/> directly rather than routing through the
-    /// compiled-replay path, because <see cref="Forward"/> contains
-    /// shape-conditional control flow (rank-3 → rank-4 batch promotion and the
-    /// final <c>Reshape</c> that strips the synthetic batch dim) that the
-    /// tracer does not capture faithfully — routing through
-    /// <see cref="NeuralNetworkBase{T}.PredictCompiled"/> was observed to
-    /// return an intermediate feature-map shape instead of the final logits.
+    /// Makes a prediction using the ResNet network for the given input.
     /// </summary>
     /// <param name="input">The input tensor to make a prediction for.</param>
     /// <returns>The predicted output tensor containing class probabilities.</returns>
+    // Hotfix: bypass the compiled-replay path in NeuralNetworkBase.Predict.
+    // Forward has shape-conditional control flow (rank-3 → rank-4 batch
+    // promotion and the trailing Reshape that strips the synthetic batch
+    // dim) that the tracer does not capture correctly — routing through
+    // PredictCompiled returned an intermediate feature-map shape instead of
+    // the final logits. Tracked upstream at ooples/AiDotNet.Tensors#228;
+    // remove this override once that lands.
     public override Tensor<T> Predict(Tensor<T> input) => Forward(input);
 
     /// <summary>

src/NeuralNetworks/SiameseNeuralNetwork.cs

@@ -301,13 +301,12 @@ public override void UpdateParameters(Vector<T> parameters)
             }
         }
 
-        /// <summary>
-        /// Runs <see cref="Forward"/> directly rather than routing through the
-        /// compiled-replay path. <see cref="Forward"/> contains shape-conditional
-        /// control flow that the tracer does not capture faithfully — routing
-        /// through <see cref="NeuralNetworkBase{T}.PredictCompiled"/> was observed
-        /// to return an intermediate shape instead of the final output.
-        /// </summary>
+        // Hotfix: bypass the compiled-replay path in NeuralNetworkBase.Predict.
+        // Forward has shape-conditional control flow the tracer does not
+        // capture correctly — routing through PredictCompiled returned an
+        // intermediate shape instead of the final output. Tracked upstream
+        // at ooples/AiDotNet.Tensors#228; remove this override once that
+        // lands.
         public override Tensor<T> Predict(Tensor<T> input) => Forward(input);
 
         /// <summary>

src/NeuralNetworks/UNet3D.cs

@@ -235,14 +235,12 @@ public Tensor<T> Forward(Tensor<T> input)
     /// </summary>
     /// <param name="input">The input voxel grid tensor.</param>
     /// <returns>The predicted segmentation map.</returns>
-    /// <inheritdoc />
-    /// <summary>
-    /// Runs <see cref="Forward"/> directly rather than routing through the
-    /// compiled-replay path. <see cref="Forward"/> contains shape-conditional
-    /// control flow that the tracer does not capture faithfully — routing
-    /// through <see cref="NeuralNetworkBase{T}.PredictCompiled"/> was observed
-    /// to return an intermediate feature-map shape instead of the final output.
-    /// </summary>
+    // Hotfix: bypass the compiled-replay path in NeuralNetworkBase.Predict.
+    // Forward has shape-conditional control flow the tracer does not capture
+    // correctly — routing through PredictCompiled returned an intermediate
+    // feature-map shape instead of the final segmentation map. Tracked
+    // upstream at ooples/AiDotNet.Tensors#228; remove this override once
+    // that lands.
     public override Tensor<T> Predict(Tensor<T> input) => Forward(input);
 
     /// <summary>

src/NeuralNetworks/VGGNetwork.cs

@@ -339,14 +339,12 @@ public override void UpdateParameters(Vector<T> parameters)
     /// probability is the network's prediction.
     /// </para>
     /// </remarks>
-    /// <summary>
-    /// Runs <see cref="Forward"/> directly rather than routing through the
-    /// compiled-replay path. <see cref="Forward"/> contains shape-conditional
-    /// control flow (batch-dim promotion, post-pipeline reshape) that the
-    /// tracer does not capture faithfully — routing through
-    /// <see cref="NeuralNetworkBase{T}.PredictCompiled"/> was observed to
-    /// return an intermediate feature-map shape instead of the final logits.
-    /// </summary>
+    // Hotfix: bypass the compiled-replay path in NeuralNetworkBase.Predict
+    // because Forward has shape-conditional control flow (batch-dim
+    // promotion, post-pipeline reshape) that the tracer does not capture
+    // correctly — routing through PredictCompiled returned an intermediate
+    // feature-map shape instead of the final logits. Tracked upstream at
+    // ooples/AiDotNet.Tensors#228; remove this override once that lands.
     public override Tensor<T> Predict(Tensor<T> input) => Forward(input);
 
     /// <summary>

src/NeuralNetworks/VoxelCNN.cs

@@ -228,14 +228,11 @@ public Tensor<T> Forward(Tensor<T> input)
     /// </summary>
     /// <param name="input">The input voxel grid tensor.</param>
     /// <returns>The predicted class probabilities or scores.</returns>
-    /// <inheritdoc />
-    /// <summary>
-    /// Runs <see cref="Forward"/> directly rather than routing through the
-    /// compiled-replay path. <see cref="Forward"/> contains shape-conditional
-    /// control flow that the tracer does not capture faithfully — routing
-    /// through <see cref="NeuralNetworkBase{T}.PredictCompiled"/> was observed
-    /// to return an intermediate feature-map shape instead of the final output.
-    /// </summary>
+    // Hotfix: bypass the compiled-replay path in NeuralNetworkBase.Predict.
+    // Forward has shape-conditional control flow the tracer does not capture
+    // correctly — routing through PredictCompiled returned an intermediate
+    // feature-map shape instead of the final logits. Tracked upstream at
+    // ooples/AiDotNet.Tensors#228; remove this override once that lands.
     public override Tensor<T> Predict(Tensor<T> input) => Forward(input);
 
     /// <summary>