diff --git a/fitspy/spectra.py b/fitspy/spectra.py
index 7404806..0a8fae8 100644
--- a/fitspy/spectra.py
+++ b/fitspy/spectra.py
@@ -3,18 +3,16 @@
 "Spectra"
 """
 import os
-import csv
 from copy import deepcopy
 from concurrent.futures import ProcessPoolExecutor
 import matplotlib.pyplot as plt
-from lmfit import Model, fit_report, Parameters
+from lmfit import Model, Parameters
 from lmfit.model import ModelResult
 
-from fitspy.utils import fileparts, check_or_rename
-from fitspy.utils import save_to_json, load_from_json
+from fitspy.utils import fileparts, save_to_json, load_from_json
 from fitspy.models import gaussian
 from fitspy.spectrum import Spectrum
-from fitspy import MODELS, PARAMS
+from fitspy import MODELS
 
 
 def fit(params):
@@ -129,38 +127,11 @@ def save_results(self, dirname_res, fnames=None):
         if fnames is None:
             fnames = self.fnames
 
-        def write_params(fname_params, labels, models):
-            with open(fname_params, 'w', newline='') as fid:
-                writer = csv.writer(fid, delimiter=';')
-                writer.writerow(['label', 'model'] + PARAMS)
-                for label, model in zip(labels, models):
-                    vals = [label, Spectrum.get_model_name(model)]
-                    for key in PARAMS:
-                        params = model.param_hints
-                        if key in params.keys():
-                            vals.append(params[key]['value'])
-                        else:
-                            vals.append('')
-                    writer.writerow(vals)
-
         for fname in fnames:
             spectrum, _ = self.get_objects(fname)
             if spectrum.result_fit is not None:
-                # TODO : use Path, move write method in spectrum
-                _, name, _ = fileparts(fname)
-
-                # results saving
-                fname_params = os.path.join(dirname_res, name + '.csv')
-                fname_params = check_or_rename(fname_params)
-                labels, models = spectrum.models_labels, spectrum.models
-                if len(models) > 0:
-                    write_params(fname_params, labels, models)
-
-                # statistics saving
-                fname_stats = os.path.join(dirname_res, name + '_stats.txt')
-                fname_stats = check_or_rename(fname_stats)
-                with open(fname_stats, 'w') as fid:
-                    fid.write(fit_report(spectrum.result_fit))
+                spectrum.save_params(dirname_res)
+                spectrum.save_stats(dirname_res)
 
     def save_figures(self, dirname_fig, fnames=None, bounds=None):
         """
diff --git a/fitspy/spectrum.py b/fitspy/spectrum.py
index 326db15..3510b75 100644
--- a/fitspy/spectrum.py
+++ b/fitspy/spectrum.py
@@ -1,21 +1,23 @@
 """
 Class dedicated to spectrum processing
 """
+import os
+import csv
 import itertools
 from copy import deepcopy
 import numpy as np
 import pandas as pd
 from scipy.signal import find_peaks
-from lmfit import Model, report_fit
+from lmfit import Model, report_fit, fit_report
 from lmfit.models import ConstantModel, LinearModel, ParabolicModel, \
     ExponentialModel, ExpressionModel  # pylint:disable=unused-import
 
-from fitspy.utils import closest_index
+from fitspy.utils import closest_index, fileparts, check_or_rename
 from fitspy.utils import save_to_json, load_from_json
 from fitspy.app.utils import convert_dict_from_tk_variables
 from fitspy.app.utils import dict_has_tk_variable
 from fitspy.baseline import BaseLine
-from fitspy import MODELS, BKG_MODELS
+from fitspy import MODELS, PARAMS, BKG_MODELS
 
 
 class Spectrum:
@@ -121,7 +123,7 @@ def set_attributes(self, dict_attrs, **fit_kwargs):
                         model.prefix = pfx
                     else:
                         model = Model(model, independent_vars=['x'], prefix=pfx)
-                    model.param_hints = param_hints
+                    model.param_hints = deepcopy(param_hints)
                     self.models.append(model)
 
         if 'bkg_model' in keys and dict_attrs['bkg_model']:
@@ -135,11 +137,11 @@ def set_attributes(self, dict_attrs, **fit_kwargs):
                 model = Model(model, independent_vars=['x'])
             self.bkg_model = model
             self.bkg_model.name2 = model_name
-            self.bkg_model.param_hints = param_hints
+            self.bkg_model.param_hints = deepcopy(param_hints)
 
         if 'baseline' in keys:
             self.baseline = BaseLine()
-            for key, val in vars(self.baseline).items():
+            for key in vars(self.baseline).keys():
                 if key in dict_attrs['baseline'].keys():
                     setattr(self.baseline, key, dict_attrs['baseline'][key])
 
@@ -529,6 +531,33 @@ def plot_residual(self, ax, factor=1):
             ax.plot(x, factor * residual, 'r', label=f"residual (x{factor})")
             ax.legend()
 
+    def save_params(self, dirname_params):
+        """ Save fit parameters in a '.csv' file located in 'dirname_params' """
+        _, name, _ = fileparts(self.fname)
+        fname_params = os.path.join(dirname_params, name + '.csv')
+        fname_params = check_or_rename(fname_params)
+        if len(self.models) > 0:
+            with open(fname_params, 'w', newline='') as fid:
+                writer = csv.writer(fid, delimiter=';')
+                writer.writerow(['label', 'model'] + PARAMS)
+                for label, model in zip(self.models_labels, self.models):
+                    vals = [label, self.get_model_name(model)]
+                    for key in PARAMS:
+                        params = model.param_hints
+                        if key in params.keys():
+                            vals.append(params[key]['value'])
+                        else:
+                            vals.append('')
+                    writer.writerow(vals)
+
+    def save_stats(self, dirname_stats):
+        """ Save statistics in a '.txt' file located in 'dirname_stats' """
+        _, name, _ = fileparts(self.fname)
+        fname_stats = os.path.join(dirname_stats, name + '_stats.txt')
+        fname_stats = check_or_rename(fname_stats)
+        with open(fname_stats, 'w') as fid:
+            fid.write(fit_report(self.result_fit))
+
     def save(self, fname_json=None):
         """ Return a 'dict_attrs' dictionary from the spectrum attributes and
             Save it if a 'fname_json' is given """