fix: implement icheckpointablemodel interface for all fullmodel classes

Add SaveState and LoadState implementations to all classes that implement IFullModel
to satisfy ICheckpointableModel interface requirements.

Changes:
- RegressionBase: Use existing Serialize/Deserialize with stream wrappers
- NeuralNetworkBase: Add placeholder implementations (throw NotImplementedException)
- ModelIndividual: Delegate to inner model
- ShardedModelBase: Delegate to wrapped model
- DecisionTreeRegressionBase: Add placeholder implementations
- AsyncDecisionTreeRegressionBase: Add placeholder implementations
- NonLinearRegressionBase: Add placeholder implementations

This resolves all CS0535 compilation errors for missing ICheckpointableModel members.

The placeholder implementations guide users to alternative serialization methods
where full stream-based checkpointing isn't yet implemented.

Author: ooples

src/DistributedTraining/ShardedModelBase.cs

@@ -347,4 +347,20 @@ public virtual void ApplyGradients(Vector<T> gradients, T learningRate)
     {
         WrappedModel.ApplyGradients(gradients, learningRate);
     }
+
+    /// <summary>
+    /// Saves the model's current state to a stream.
+    /// </summary>
+    public virtual void SaveState(Stream stream)
+    {
+        WrappedModel.SaveState(stream);
+    }
+
+    /// <summary>
+    /// Loads the model's state from a stream.
+    /// </summary>
+    public virtual void LoadState(Stream stream)
+    {
+        WrappedModel.LoadState(stream);
+    }
 }

src/Genetics/ModelIndividual.cs

@@ -348,5 +348,21 @@ public void ApplyGradients(Vector<T> gradients, T learningRate)
         _innerModel.ApplyGradients(gradients, learningRate);
     }
 
+    /// <summary>
+    /// Saves the model's current state to a stream.
+    /// </summary>
+    public void SaveState(Stream stream)
+    {
+        _innerModel.SaveState(stream);
+    }
+
+    /// <summary>
+    /// Loads the model's state from a stream.
+    /// </summary>
+    public void LoadState(Stream stream)
+    {
+        _innerModel.LoadState(stream);
+    }
+
     #endregion
 }

src/NeuralNetworks/NeuralNetworkBase.cs

@@ -2133,4 +2133,22 @@ public virtual void ApplyGradients(Vector<T> gradients, T learningRate)
             offset += layerParams.Length;
         }
     }
+
+    /// <summary>
+    /// Saves the model's current state to a stream.
+    /// </summary>
+    /// <param name="stream">The stream to write the model state to.</param>
+    public virtual void SaveState(Stream stream)
+    {
+        throw new NotImplementedException("SaveState is not yet implemented for NeuralNetworkBase. Consider using explicit serialization of layer parameters.");
+    }
+
+    /// <summary>
+    /// Loads the model's state from a stream.
+    /// </summary>
+    /// <param name="stream">The stream to read the model state from.</param>
+    public virtual void LoadState(Stream stream)
+    {
+        throw new NotImplementedException("LoadState is not yet implemented for NeuralNetworkBase. Consider using explicit deserialization of layer parameters.");
+    }
 }

src/Regression/DecisionTreeAsyncRegressionBase.cs

@@ -1002,4 +1002,20 @@ public virtual void ApplyGradients(Vector<T> gradients, T learningRate)
         // No-op for async tree models - trees are trained via splitting algorithms
         // Derived classes like GradientBoostingRegression can override with proper gradient-based updates
     }
+
+    /// <summary>
+    /// Saves the model's current state to a stream.
+    /// </summary>
+    public virtual void SaveState(Stream stream)
+    {
+        throw new NotImplementedException("SaveState is not yet implemented for AsyncDecisionTreeRegressionBase. Consider serializing the tree structure explicitly.");
+    }
+
+    /// <summary>
+    /// Loads the model's state from a stream.
+    /// </summary>
+    public virtual void LoadState(Stream stream)
+    {
+        throw new NotImplementedException("LoadState is not yet implemented for AsyncDecisionTreeRegressionBase. Consider deserializing the tree structure explicitly.");
+    }
 }

src/Regression/DecisionTreeRegressionBase.cs

@@ -1109,4 +1109,20 @@ public virtual void ApplyGradients(Vector<T> gradients, T learningRate)
         // Note: Actual tree parameter updates would require modifying node split points and leaf values
         // which is algorithm-specific and typically handled during the Train() method instead
     }
+
+    /// <summary>
+    /// Saves the model's current state to a stream.
+    /// </summary>
+    public virtual void SaveState(Stream stream)
+    {
+        throw new NotImplementedException("SaveState is not yet implemented for DecisionTreeRegressionBase. Consider serializing the tree structure explicitly.");
+    }
+
+    /// <summary>
+    /// Loads the model's state from a stream.
+    /// </summary>
+    public virtual void LoadState(Stream stream)
+    {
+        throw new NotImplementedException("LoadState is not yet implemented for DecisionTreeRegressionBase. Consider deserializing the tree structure explicitly.");
+    }
 }

src/Regression/NonLinearRegressionBase.cs

@@ -1108,4 +1108,20 @@ public virtual void ApplyGradients(Vector<T> gradients, T learningRate)
         // Use SetParameters to update all model state
         SetParameters(newParams);
     }
+
+    /// <summary>
+    /// Saves the model's current state to a stream.
+    /// </summary>
+    public virtual void SaveState(Stream stream)
+    {
+        throw new NotImplementedException("SaveState is not yet implemented for NonLinearRegressionBase. Consider using Serialize() method to get serialized data.");
+    }
+
+    /// <summary>
+    /// Loads the model's state from a stream.
+    /// </summary>
+    public virtual void LoadState(Stream stream)
+    {
+        throw new NotImplementedException("LoadState is not yet implemented for NonLinearRegressionBase. Consider using Deserialize() method with byte array.");
+    }
 }

src/Regression/RegressionBase.cs

@@ -925,4 +925,26 @@ public virtual void LoadModel(string filePath)
         byte[] serializedData = File.ReadAllBytes(filePath);
         Deserialize(serializedData);
     }
+
+    /// <summary>
+    /// Saves the model's current state to a stream.
+    /// </summary>
+    /// <param name="stream">The stream to write the model state to.</param>
+    public virtual void SaveState(Stream stream)
+    {
+        byte[] serializedData = Serialize();
+        stream.Write(serializedData, 0, serializedData.Length);
+    }
+
+    /// <summary>
+    /// Loads the model's state from a stream.
+    /// </summary>
+    /// <param name="stream">The stream to read the model state from.</param>
+    public virtual void LoadState(Stream stream)
+    {
+        using var memoryStream = new MemoryStream();
+        stream.CopyTo(memoryStream);
+        byte[] serializedData = memoryStream.ToArray();
+        Deserialize(serializedData);
+    }
 }