diff --git a/ovf/core/GVFManager.py b/ovf/core/GVFManager.py
new file mode 100644
index 0000000..addcb41
--- /dev/null
+++ b/ovf/core/GVFManager.py
@@ -0,0 +1,123 @@
+from ovf.core.VFManager import VFManager
+from dataclasses import dataclass
+from typing import List,Dict,Literal
+import numpy as np
+import json 
+import skrf as rf
+import os
+from ovf.core.sample import auto_select_multple_ports
+from ovf.core.basis.MultiPortOrthonormalBasis import MultiPortOrthonormalBasis
+from ovf.core.geometry.plot_poles import plot_poles_in_3d
+from ovf.schemas.geometry.poles import PolesPlot3dUnit,PolesPlot3dDataUnit
+import pickle
+
+@dataclass
+class GVFDataUnit:
+    geometries: Dict[str,float]
+    vf_manager:VFManager
+    enabled:bool=True
+
+class GVFManager:
+    def __init__(self):
+        self.parameter_type: Literal["s","y","z"] = "s"
+        self.datasets: List[GVFDataUnit] = []
+        
+    def save(self,filepath:str):
+        os.makedirs(os.path.dirname(filepath),exist_ok=True)
+        with open(filepath,"wb") as f:
+            pickle.dump(self,f)
+            
+    @classmethod
+    def load(cls,filepath:str):
+        with open(filepath,"rb") as f:
+            obj = pickle.load(f)
+        assert isinstance(obj,GVFManager),"The loaded object is not a GVFManager instance."
+        return obj
+    
+    def load_from_datasets(self,jsonfile:str,npoles_cplx:int=2,max_iterations:int=5,max_points:int|None=20,basis:type=MultiPortOrthonormalBasis,parameter_type:Literal["s","y","z"]="s"):
+        self.parameter_type = parameter_type
+        with open(jsonfile,"r") as f:
+            datas = json.load(f)
+
+        for i in range(len(datas)):
+            print(f"Loading dataset {i+1}/{len(datas)+1}")
+            network = rf.Network(f"{os.path.dirname(jsonfile)}/{datas[i]['result_dir']}")
+            ports = network.nports
+            K = max_iterations
+
+            full_freqences = network.f
+            
+            if parameter_type == "s":
+                sampled_points = network.s.reshape(-1,ports,ports)
+            elif parameter_type == "y":
+                sampled_points = network.y.reshape(-1,ports,ports)
+            elif parameter_type == "z":
+                sampled_points = network.z.reshape(-1,ports,ports)
+
+            if max_points:
+                H,freqs = auto_select_multple_ports(sampled_points,full_freqences,max_points=max_points)
+            else:
+                H,freqs = sampled_points,full_freqences
+            vf = VFManager(npoles_cplx=npoles_cplx,freqs=freqs,H=H,model=basis,iterations=K,verbose=False)
+            vf.fit()
+            
+            geometries = datas[i]["parameters"]
+            self.datasets.append(GVFDataUnit(geometries=geometries, vf_manager=vf))
+            
+    def plot_poles(self,save_path,degree:int,geometry_1:str,geometry_2:str):
+        unit = PolesPlot3dUnit(
+            title="Poles Distribution",
+            datas=[
+                PolesPlot3dDataUnit(
+                    poles=ds.vf_manager.poles,
+                    geometry_1=ds.geometries[geometry_1],
+                    geometry_2=ds.geometries[geometry_2]
+                ) for ds in self.datasets
+            ],
+            x_scale="linear",
+            y_scale="linear",
+            z_scale="linear",
+            ridge_degree=degree,
+            ridge_alpha=1.0,
+            real_part=True,
+            imag_part=True,
+            magnitude=True,
+            phase=False,
+            pole_label="pole",
+            geometry_1_label=geometry_1,
+            geometry_2_label=geometry_2,
+            pole_scale="linear",
+            geometry_1_scale="linear",
+            geometry_2_scale="linear"
+        )
+
+        plot_poles_in_3d(unit, save_path)
+        
+    def plot_vf_responses_with_index(self,save_dir:str,index:int):
+        ds = self.datasets[index]
+        id = f"{index}"
+        os.makedirs(save_dir,exist_ok=True)
+        vf = ds.vf_manager
+        vf.plot_metrics(show=False,save_path=f"{save_dir}/{id}")
+        full_freqences = vf.freqs
+        model_responses = vf.get_model_responses(full_freqences)
+        vf.plot_model_responses(show=False,save_path=f"{save_dir}/{id}")
+        
+    def plot_all_vf_responses(self,save_dir:str):
+        for index,ds in enumerate(self.datasets):
+            id = f"{index}"
+            os.makedirs(save_dir,exist_ok=True)
+            vf = ds.vf_manager
+            vf.plot_metrics(show=False,save_path=f"{save_dir}/{id}")
+            full_freqences = vf.freqs
+            model_responses = vf.get_model_responses(full_freqences)
+            vf.plot_model_responses(show=False,save_path=f"{save_dir}/{id}")
+        
+    def evaluate_dataset(self):
+        for ds in self.datasets:
+            rmse = ds.vf_manager.rms_error
+            condition = ds.vf_manager.condition
+            row_condition = ds.vf_manager.row_condition
+            col_condition = ds.vf_manager.col_condition
+            
+            
diff --git a/ovf/core/VFManager.py b/ovf/core/VFManager.py
index 692cf11..d9b3840 100644
--- a/ovf/core/VFManager.py
+++ b/ovf/core/VFManager.py
@@ -48,6 +48,47 @@ class VFManager():
         self.model_responses_H = None
 
         self.model=model
+        
+    @property
+    def residuals(self):
+        assert self.model_instance is not None ,"Please run levi() and sk_iteration() first."
+        return self.model_instance.residuals
+    
+    @property
+    def constant(self):
+        assert self.model_instance is not None ,"Please run levi() and sk_iteration() first."
+        return self.model_instance.D
+    
+    @property
+    def proportional(self):
+        assert self.model_instance is not None ,"Please run levi() and sk_iteration() first."
+        return self.model_instance.proportional
+    
+    @property
+    def poles(self):
+        assert self.model_instance is not None ,"Please run levi() and sk_iteration() first."
+        return list(self.model_instance.poles)
+    
+    @property
+    def rms_error(self):
+        assert self.model_instance is not None ,"Please run levi() and sk_iteration() first."
+        return self.eigenval_rms_error[-1]
+    
+    @property
+    def condition(self):
+        assert self.model_instance is not None ,"Please run levi() and sk_iteration() first."
+        return self.eigenval_condition[-1]
+    
+    @property
+    def row_condition(self):
+        assert self.model_instance is not None ,"Please run levi() and sk_iteration() first."
+        return self.eigenval_row_condition[-1]
+    
+    @property
+    def col_condition(self):
+        assert self.model_instance is not None ,"Please run levi() and sk_iteration() first."
+        return self.eigenval_col_condition[-1]
+    
 
     @classmethod
     def load(cls,dirname):
@@ -146,11 +187,11 @@ class VFManager():
         return self.model
 
     def levi(self):
-        self.poles = generate_starting_poles(self.npoles_cplx,beta_min=max(self.freqs[0],1e4),beta_max=self.freqs[-1]*1.1,verbose=self.verbose)
+        self._poles = generate_starting_poles(self.npoles_cplx,beta_min=max(self.freqs[0],1e4),beta_max=self.freqs[-1]*1.1,verbose=self.verbose)
         self.model_instance=self.model(
             H=self.H,
             freqs=self.freqs,
-            pre_poles=self.poles,
+            pre_poles=self._poles,
             fit_constant=self.fit_constant,
             fit_proportional=self.fit_proportional,
             dc_enforce=self.dc_enforce,
@@ -161,11 +202,11 @@ class VFManager():
     def sk_iteration(self):
         for i in range(self.iterations):
             assert self.model_instance is not None ,"Please run levi() first."
-            self.poles = self.model_instance.poles
+            self._poles = self.model_instance.poles
             self.model_instance = self.model(
                 H=self.H,
                 freqs=self.freqs,
-                pre_poles=self.poles,
+                pre_poles=self._poles,
                 fit_constant=self.fit_constant,
                 fit_proportional=self.fit_proportional,
                 dc_enforce=self.dc_enforce,
diff --git a/ovf/core/basis/MultiPortOrthonormalBasis.py b/ovf/core/basis/MultiPortOrthonormalBasis.py
index 3dc8b58..6275083 100644
--- a/ovf/core/basis/MultiPortOrthonormalBasis.py
+++ b/ovf/core/basis/MultiPortOrthonormalBasis.py
@@ -96,7 +96,12 @@ class MultiPortOrthonormalBasis:
         """Return condition number and RMS error of eigenvalues of A-BC."""
         z = self.poles
         cond = np.linalg.cond(self.A - self.B @ self.C)
-        rms = np.sqrt(np.mean(np.abs(np.real(z) - np.real(self.pre_poles))**2 + np.abs(np.imag(z) - np.imag(self.pre_poles))**2))
+        # rms = np.sqrt(np.mean(np.abs(np.real(z) - np.real(self.pre_poles))**2 + np.abs(np.imag(z) - np.imag(self.pre_poles))**2))
+        
+        Hk = self.H
+        Hk_fitted = self.get_model_responses(self.freqs)
+        
+        rms = np.sqrt(np.mean(np.abs(Hk - Hk_fitted)**2))
 
         row_cond = cond_row_inf(self.A - self.B @ self.C)
         col_cond = cond_col_one(self.A - self.B @ self.C)
diff --git a/ovf/core/geometry/plot2d.py b/ovf/core/geometry/plot2d.py
new file mode 100644
index 0000000..4d825f5
--- /dev/null
+++ b/ovf/core/geometry/plot2d.py
@@ -0,0 +1,32 @@
+from plotly import graph_objects as go
+from plotly.subplots import make_subplots
+import numpy as np
+from sklearn.linear_model import Ridge
+import os
+from ovf.schemas.geometry.basic import GeometryPlot2dUnit, GeometryPlot2dComplexDataUnit
+
+def plot_2d(unit: GeometryPlot2dUnit, dirname: str):
+    os.makedirs(dirname, exist_ok=True)
+
+    fig = make_subplots(rows=1, cols=1)
+
+    for index in range(len(unit.datas[0].x)):
+        x = [unit.datas[i].x[index] for i in range(len(unit.datas))]
+        y = [unit.datas[i].y[index] for i in range(len(unit.datas))]
+        fig.add_trace(
+            go.Scatter(
+                x=x,
+                y=y,
+                mode='markers',
+                name=data.label
+            ),
+            row=1, col=1
+        )
+    
+    fig.update_layout(
+        title=unit.title,
+        xaxis_title=unit.x_label,
+        yaxis_title=unit.y_label,
+    )
+    
+    fig.write_html(f"{dirname}/plot_2d.html")
\ No newline at end of file
diff --git a/ovf/core/geometry/plot3d.py b/ovf/core/geometry/plot3d.py
new file mode 100644
index 0000000..0dc1aae
--- /dev/null
+++ b/ovf/core/geometry/plot3d.py
@@ -0,0 +1,72 @@
+from ovf.schemas.geometry.basic import GeometryPlot3dDataUnit
+from ovf.schemas.geometry.basic import GeometryPlot3dUnit
+import numpy as np
+import plotly.graph_objs as go
+from sklearn.linear_model import Ridge
+from sklearn.preprocessing import PolynomialFeatures
+from ovf.core.utils import _generate_ridge_plane
+
+
+def get_plot_instance_in_3d(datas:GeometryPlot3dUnit):
+    # 显示所有pole所在的点和拟合得到的平面
+    def predict(ridge:Ridge, x, y, poly):
+        features = poly.transform(np.array([[x, y]]))
+        result = float(ridge.predict(features)[0])
+        return result
+        
+    poly = PolynomialFeatures(degree=datas.ridge_degree, include_bias=False)
+        
+    if datas.x_scale == "log":
+        datas.x_label = "log10(" + datas.x_label + ")"
+    if datas.y_scale == "log":
+        datas.y_label = "log10(" + datas.y_label + ")"
+    if datas.z_scale == "log":
+        datas.z_label = "log10(" + datas.z_label + ")"
+        
+    pt = [(
+        np.float64((np.log10(d.x) if d.x > 0 else 0) if datas.x_scale == "log" else np.float64(d.x)), 
+        np.float64((np.log10(d.y) if d.y > 0 else 0) if datas.y_scale == "log" else np.float64(d.y)), 
+        np.float64((np.log10(d.z) if d.z > 0 else 0) if datas.z_scale == "log" else np.float64(d.z))
+        ) for d in datas.datas]
+    pt = np.array(pt)
+
+    model,poly = _generate_ridge_plane([tuple(row) for row in pt], alpha=datas.ridge_alpha, poly=poly)
+
+    # z轴取log10
+    trace = go.Scatter3d(
+        x=pt[:,0], 
+        y=pt[:,1], 
+        z=pt[:,2],
+        mode='markers',
+        marker=dict(size=6, colorscale='Viridis', opacity=0.8),
+        text=[f'{datas.x_label}:{pt[p, 0]}<br>{datas.y_label}:{pt[p, 1]}<br>{datas.z_label}:{pt[p, 2]}' for p in range(len(pt))],
+        hoverinfo='text'
+    )
+    # --- 拟合平面 ---
+
+    X_range = np.linspace(min(pt[:,0]), max(pt[:,0]), 30)
+    Y_range = np.linspace(min(pt[:,1]), max(pt[:,1]), 30)
+    Z_range = np.linspace(min(pt[:,2]), max(pt[:,2]), 30)
+    X_grid, Y_grid = np.meshgrid(X_range, Y_range)
+    X_flat = X_grid.ravel()
+    Y_flat = Y_grid.ravel()
+    Z_flat = np.array([predict(model,x,y,poly) for x, y in zip(X_flat, Y_flat)])
+    Z_grid = Z_flat.reshape(X_grid.shape)
+    surface = go.Surface(
+        x=X_range, y=Y_range, z=Z_grid,
+        opacity=0.5, colorscale='Reds', showscale=False,
+        name='Fitted Plane'
+    )
+    
+    layout = go.Layout(
+        title=datas.title,
+        scene=dict(
+            xaxis=dict(title=datas.x_label),
+            yaxis=dict(title=datas.y_label),
+            zaxis=dict(title=datas.z_label),
+        ),
+        showlegend=False
+    )
+    return trace, surface, layout
+    # fig = go.Figure(data=[trace, surface], layout=layout)
+    # plot(fig, filename=filename)
\ No newline at end of file
diff --git a/ovf/core/geometry/plot_poles.py b/ovf/core/geometry/plot_poles.py
new file mode 100644
index 0000000..0d6c57c
--- /dev/null
+++ b/ovf/core/geometry/plot_poles.py
@@ -0,0 +1,88 @@
+import plotly as px
+import plotly.graph_objs as go
+from plotly.offline import plot
+import numpy as np
+from sklearn.linear_model import Ridge
+from ovf.schemas.geometry.poles import PolesPlot3dUnit, PolesPlot3dDataUnit
+from ovf.schemas.geometry.basic import GeometryPlot3dDataUnit, GeometryPlot3dUnit
+from ovf.core.geometry.plot3d import get_plot_instance_in_3d
+from plotly.subplots import make_subplots
+import os
+
+def plot_poles_in_3d(poles:PolesPlot3dUnit,dirname:str):
+    os.makedirs(dirname,exist_ok=True)
+    
+    index = 0
+    
+    while index < len(poles.datas[0].poles):
+        if poles.real_part:
+            plotunit = GeometryPlot3dUnit(
+                title="Real Part",
+                datas=[GeometryPlot3dDataUnit(x=p.geometry_1, y=p.geometry_2, z=np.real(p.poles[index])) for p in poles.datas],
+                x_label=poles.geometry_1_label,
+                y_label=poles.geometry_2_label,
+                z_label=poles.pole_label,
+                x_scale=poles.geometry_1_scale,
+                y_scale=poles.geometry_2_scale,
+                z_scale=poles.pole_scale,
+                ridge_alpha=poles.ridge_alpha,
+                ridge_degree=poles.ridge_degree
+            )
+            trace, surface, layout = get_plot_instance_in_3d(plotunit)
+            fig = go.Figure(data=[trace, surface], layout=layout)
+            fig.write_html(f"{dirname}/pole_{index}_real.html")
+            
+        if poles.imag_part:
+            plotunit = GeometryPlot3dUnit(
+                title="Imaginary Part",
+                datas=[GeometryPlot3dDataUnit(x=p.geometry_1, y=p.geometry_2, z=np.imag(p.poles[index])) for p in poles.datas],
+                x_label=poles.geometry_1_label,
+                y_label=poles.geometry_2_label,
+                z_label=poles.pole_label,
+                x_scale=poles.geometry_1_scale,
+                y_scale=poles.geometry_2_scale,
+                z_scale=poles.pole_scale,
+                ridge_alpha=poles.ridge_alpha
+            )
+            trace, surface, layout = get_plot_instance_in_3d(plotunit)
+            fig = go.Figure(data=[trace, surface], layout=layout)
+            fig.write_html(f"{dirname}/pole_{index}_imag.html")
+        if poles.magnitude:
+            plotunit = GeometryPlot3dUnit(
+                title="Magnitude",
+                datas=[GeometryPlot3dDataUnit(x=p.geometry_1, y=p.geometry_2, z=np.abs(p.poles[index])) for p in poles.datas],
+                x_label=poles.geometry_1_label,
+                y_label=poles.geometry_2_label,
+                z_label=poles.pole_label,
+                x_scale=poles.geometry_1_scale,
+                y_scale=poles.geometry_2_scale,
+                z_scale=poles.pole_scale,
+                ridge_alpha=poles.ridge_alpha
+            )
+            trace, surface, layout = get_plot_instance_in_3d(plotunit)
+            fig = go.Figure(data=[trace, surface], layout=layout)
+            fig.write_html(f"{dirname}/pole_{index}_magnitude.html")
+        if poles.phase:
+            plotunit = GeometryPlot3dUnit(
+                title="Phase",
+                datas=[GeometryPlot3dDataUnit(x=p.geometry_1, y=p.geometry_2, z=np.float64(np.angle(p.poles[index]))) for p in poles.datas],
+                x_label=poles.geometry_1_label,
+                y_label=poles.geometry_2_label,
+                z_label=poles.pole_label,
+                x_scale=poles.geometry_1_scale,
+                y_scale=poles.geometry_2_scale,
+                z_scale=poles.pole_scale,
+                ridge_alpha=poles.ridge_alpha
+            )
+            trace, surface, layout = get_plot_instance_in_3d(plotunit)
+            fig = go.Figure(data=[trace, surface], layout=layout)
+            fig.write_html(f"{dirname}/pole_{index}_phase.html")
+
+        
+        if index + 1 < len(poles.datas[0].poles):
+            if np.isclose(poles.datas[0].poles[index] , np.conj(poles.datas[0].poles[index + 1])):
+                index += 1  # 跳过共轭点
+                
+        index += 1
+        
+        
\ No newline at end of file
diff --git a/ovf/core/utils.py b/ovf/core/utils.py
index 4859462..defc2d5 100644
--- a/ovf/core/utils.py
+++ b/ovf/core/utils.py
@@ -1,6 +1,26 @@
 import numpy as np
 from dataclasses import dataclass
 from typing import List, Dict, Tuple
+from sklearn.linear_model import Ridge
+from sklearn.preprocessing import PolynomialFeatures
+
+def _generate_ridge_plane(points:List[tuple[np.float64,np.float64,np.float64]],poly:PolynomialFeatures, alpha:float=1.0) -> tuple[Ridge,PolynomialFeatures]:
+    # 使用Ridge回归拟合平面
+    X_poly = poly.fit_transform(np.array([[d[0], d[1]] for d in points]))
+    model = Ridge(alpha=alpha)
+    model.fit(X_poly, np.array([d[2] for d in points]))
+    return model,poly
+
+def grey_to_rgb(value):
+    """Convert a grey value (0-255) to an RGB tuple."""
+    if not (0 <= value <= 255):
+        raise ValueError("Value must be between 0 and 255")
+    # color refers to rainbow scale
+    ratio = value / 255
+    b = int(max(0, 255 * (1 - 2 * ratio)))
+    r = int(max(0, 255 * (2 * ratio - 1)))
+    g = 255 - b - r
+    return (r, g, b)
 
 def cond_row_inf(A, use_pinv=True):
     """行条件数 κ∞(A) = ||A||∞ * ||A^{-1}||∞；矩形阵用广义逆。"""
diff --git a/ovf/schemas/geometry/basic.py b/ovf/schemas/geometry/basic.py
new file mode 100644
index 0000000..d53a0a8
--- /dev/null
+++ b/ovf/schemas/geometry/basic.py
@@ -0,0 +1,36 @@
+from dataclasses import dataclass
+from typing import List, Literal
+from typing import Dict, Union
+
+@dataclass
+class GeometryPlot3dDataUnit:
+    x: float
+    y: float
+    z: float
+    
+@dataclass
+class GeometryPlot3dUnit:
+    title: str
+    datas: List[GeometryPlot3dDataUnit]
+    ridge_alpha: float = 1.0  # Ridge回归的正则化参数
+    ridge_degree: int = 2     # Ridge回归的多项式特征的度
+    x_label: str = "x"
+    y_label: str = "y"
+    z_label: str = "z"
+    x_scale: Literal["linear", "log"] = "linear" 
+    y_scale: Literal["linear", "log"] = "linear"
+    z_scale: Literal["linear", "log"] = "linear"
+    
+@dataclass
+class GeometryPlot2dComplexDataUnit:
+    x: List[float]
+    y: List[float]
+    geometries: Dict[str,Union[float,int,str]]
+
+
+@dataclass
+class GeometryPlot2dUnit:
+    title: str
+    datas: List[GeometryPlot2dComplexDataUnit]
+    x_label: str = "x"
+    y_label: str = "y"
diff --git a/ovf/schemas/geometry/poles.py b/ovf/schemas/geometry/poles.py
new file mode 100644
index 0000000..3301596
--- /dev/null
+++ b/ovf/schemas/geometry/poles.py
@@ -0,0 +1,30 @@
+from dataclasses import dataclass
+from typing import List, Literal, Union
+import numpy as np
+
+@dataclass
+class PolesPlot3dDataUnit:
+    poles: List[Union[complex,np.complex128]]
+    geometry_1: float
+    geometry_2: float
+
+    
+@dataclass
+class PolesPlot3dUnit:
+    title: str
+    datas: List[PolesPlot3dDataUnit]
+    x_scale: Literal["linear", "log"] = "linear" 
+    y_scale: Literal["linear", "log"] = "linear"
+    z_scale: Literal["linear", "log"] = "linear"
+    ridge_degree: int = 2     # Ridge回归的多项式特征的度
+    ridge_alpha: float = 1.0  # Ridge回归的正则化参数
+    real_part: bool = True  # 是否显示实部
+    imag_part: bool = True  # 是否显示虚部
+    magnitude: bool = True  # 是否显示幅值
+    phase: bool = False  # 是否显示相位
+    pole_label: str = "pole"
+    geometry_1_label: str = "geometry_1"
+    geometry_2_label: str = "geometry_2"
+    pole_scale: Literal["linear", "log"] = "linear"
+    geometry_1_scale: Literal["linear", "log"] = "linear"
+    geometry_2_scale: Literal["linear", "log"] = "linear"
\ No newline at end of file