CUQIpy 1.2.0.post0.dev8__tar.gz → 1.2.0.post0.dev15__tar.gz

{cuqipy-1.2.0.post0.dev8 → cuqipy-1.2.0.post0.dev15}/CUQIpy.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: CUQIpy
-Version: 1.2.0.post0.dev8
+Version: 1.2.0.post0.dev15
 Summary: Computational Uncertainty Quantification for Inverse problems in Python
 Maintainer-email: "Nicolai A. B. Riis" <nabr@dtu.dk>, "Jakob S. Jørgensen" <jakj@dtu.dk>, "Amal M. Alghamdi" <amaal@dtu.dk>, Chao Zhang <chaz@dtu.dk>
 License:                                  Apache License

{cuqipy-1.2.0.post0.dev8 → cuqipy-1.2.0.post0.dev15}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: CUQIpy
-Version: 1.2.0.post0.dev8
+Version: 1.2.0.post0.dev15
 Summary: Computational Uncertainty Quantification for Inverse problems in Python
 Maintainer-email: "Nicolai A. B. Riis" <nabr@dtu.dk>, "Jakob S. Jørgensen" <jakj@dtu.dk>, "Amal M. Alghamdi" <amaal@dtu.dk>, Chao Zhang <chaz@dtu.dk>
 License:                                  Apache License

{cuqipy-1.2.0.post0.dev8 → cuqipy-1.2.0.post0.dev15}/cuqi/_version.py

@@ -8,11 +8,11 @@ import json
 
 version_json = '''
 {
- "date": "2024-09-30T09:16:05+0200",
+ "date": "2024-09-30T18:06:47+0200",
  "dirty": false,
  "error": null,
- "full-revisionid": "915f2f14ca7a4f588b561d25a1c4a47b8e295835",
- "version": "1.2.0.post0.dev8"
+ "full-revisionid": "59a6730dfe1227a35a2def467daf63265bdba87f",
+ "version": "1.2.0.post0.dev15"
 }
 '''  # END VERSION_JSON
 

{cuqipy-1.2.0.post0.dev8 → cuqipy-1.2.0.post0.dev15}/cuqi/experimental/mcmc/_hmc.py

@@ -88,14 +88,16 @@ class NUTS(Sampler):
     """
 
     _STATE_KEYS = Sampler._STATE_KEYS.union({'_epsilon', '_epsilon_bar',
-                                                '_H_bar', '_mu',
-                                                '_alpha', '_n_alpha'})
+                                                '_H_bar',
+                                                'current_target_logd',
+                                                'current_target_grad',
+                                                'max_depth'})
 
     _HISTORY_KEYS = Sampler._HISTORY_KEYS.union({'num_tree_node_list',
                                                     'epsilon_list',
                                                     'epsilon_bar_list'})
 
-    def __init__(self, target=None, initial_point=None, max_depth=15,
+    def __init__(self, target=None, initial_point=None, max_depth=None,
                  step_size=None, opt_acc_rate=0.6, **kwargs):
         super().__init__(target, initial_point=initial_point, **kwargs)
 
@@ -104,30 +106,23 @@ class NUTS(Sampler):
         self.step_size = step_size
         self.opt_acc_rate = opt_acc_rate
 
-        # Initialize epsilon and epsilon_bar
-        # epsilon is the step size used in the current iteration
-        # after warm up and one sampling step, epsilon is updated
-        # to epsilon_bar for the remaining sampling steps.
-        self._epsilon = None
-        self._epsilon_bar = None
-        self._H_bar = None
-
-        # Extra parameters for tuning
-        self._n_alpha = None
-        self._alpha = None
-
 
     def _initialize(self):
 
         # Arrays to store acceptance rate
         self._acc = [None] # Overwrites acc from Sampler. TODO. Check if this is necessary
 
-        self._alpha = 0 # check if meaningful value
-        self._n_alpha = 0 # check if meaningful value
+        self._current_alpha_ratio = np.nan # Current alpha ratio is set to some
+                                           # value (other than np.nan) before 
+                                           # being used
 
         self.current_target_logd, self.current_target_grad = self._nuts_target(self.current_point)
 
-        # parameters dual averaging
+        # Parameters dual averaging
+        # Initialize epsilon and epsilon_bar
+        # epsilon is the step size used in the current iteration
+        # after warm up and one sampling step, epsilon is updated
+        # to epsilon_bar for the remaining sampling steps.
         if self.step_size is None:
             self._epsilon = self._FindGoodEpsilon()
         else:
@@ -155,6 +150,8 @@ class NUTS(Sampler):
 
     @max_depth.setter
     def max_depth(self, value):
+        if value is None:
+            value = 15 # default value
         if not isinstance(value, int):
             raise TypeError('max_depth must be an integer.')
         if value < 0:
@@ -273,8 +270,7 @@ class NUTS(Sampler):
             s = s_prime *\
                 int((dpoints @ r_minus.T) >= 0) * int((dpoints @ r_plus.T) >= 0)
             j += 1
-            self._alpha = alpha
-            self._n_alpha = n_alpha
+            self._current_alpha_ratio = alpha/n_alpha
 
         # update run diagnostic attributes
         self._update_run_diagnostic_attributes(
@@ -293,7 +289,7 @@ class NUTS(Sampler):
 
         eta1 = 1/(k + t_0)
         self._H_bar = (1-eta1)*self._H_bar +\
-            eta1*(self.opt_acc_rate - (self._alpha/self._n_alpha))
+            eta1*(self.opt_acc_rate - (self._current_alpha_ratio))
         self._epsilon = np.exp(self._mu - (np.sqrt(k)/gamma)*self._H_bar)
         eta = k**(-kappa)
         self._epsilon_bar =\
@@ -309,12 +305,12 @@ class NUTS(Sampler):
         #    self._epsilon
         #  Parameters that does not change during the run
         #    self._mu
-
+        self._ensure_initialized()
         if self._epsilon_bar == "unset": # Initial value of epsilon_bar for tuning
             self._epsilon_bar = 1
 
     def _pre_sample(self):
-
+        self._ensure_initialized()
         if self._epsilon_bar == "unset": # Initial value of epsilon_bar for sampling
             self._epsilon_bar = self._epsilon