CUQIpy 1.2.0.post0.dev8__py3-none-any.whl → 1.2.0.post0.dev27__py3-none-any.whl

{CUQIpy-1.2.0.post0.dev8.dist-info → CUQIpy-1.2.0.post0.dev27.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: CUQIpy
-Version: 1.2.0.post0.dev8
+Version: 1.2.0.post0.dev27
 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.dist-info → CUQIpy-1.2.0.post0.dev27.dist-info}/RECORD

@@ -1,6 +1,6 @@
 cuqi/__init__.py,sha256=LsGilhl-hBLEn6Glt8S_l0OJzAA1sKit_rui8h-D-p0,488
 cuqi/_messages.py,sha256=fzEBrZT2kbmfecBBPm7spVu7yHdxGARQB4QzXhJbCJ0,415
-cuqi/_version.py,sha256=4J105k1-kR-3pwUoZ1l_3exw9H13lnsyupul8rBPwlQ,508
+cuqi/_version.py,sha256=XUNd3Y4LTHLhvH0Xz5AA7zswxVQgfbbyvLs1ONnxum0,509
 cuqi/config.py,sha256=wcYvz19wkeKW2EKCGIKJiTpWt5kdaxyt4imyRkvtTRA,526
 cuqi/diagnostics.py,sha256=5OrbJeqpynqRXOe5MtOKKhe7EAVdOEpHIqHnlMW9G_c,3029
 cuqi/array/__init__.py,sha256=-EeiaiWGNsE3twRS4dD814BIlfxEsNkTCZUc5gjOXb0,30
@@ -40,13 +40,13 @@ cuqi/experimental/mcmc/_conjugate_approx.py,sha256=uEnY2ea9su5ivcNagyRAwpQP2gBY9
 cuqi/experimental/mcmc/_cwmh.py,sha256=50v3uZaWhlVnfrEB5-lB_7pn8QoUVBe-xWxKGKbmNHg,7234
 cuqi/experimental/mcmc/_direct.py,sha256=9pQS_2Qk2-ybt6m8WTfPoKetcxQ00WaTRN85-Z0FrBY,777
 cuqi/experimental/mcmc/_gibbs.py,sha256=a_Zr007F233R6A3CJ_0zlQppsQrVuX8-oWCbNJWJxmA,12745
-cuqi/experimental/mcmc/_hmc.py,sha256=UAPM9tGs8tneUcDiywXKEBNsKazQMLSjWsTemJ7QwnQ,19377
+cuqi/experimental/mcmc/_hmc.py,sha256=AR8ucnjPe9Q78em3IMoihLoifdBxu9oAjrQ51AYPHH4,19390
 cuqi/experimental/mcmc/_langevin_algorithm.py,sha256=yNO7ABxmkixzcLG-lv57GOTyeTr7HwFs2DrrhuZW9OI,8398
 cuqi/experimental/mcmc/_laplace_approximation.py,sha256=rdiE3cMQFq6FLQcOQwPpuGIxrTAp3aoGPxMDSdeopV0,5688
 cuqi/experimental/mcmc/_mh.py,sha256=MXo0ahXP4KGFkaY4HtvcBE-TMQzsMlTmLKzSvpz7drU,2941
 cuqi/experimental/mcmc/_pcn.py,sha256=wqJBZLuRFSwxihaI53tumAg6AWVuceLMOmXssTetd1A,3374
 cuqi/experimental/mcmc/_rto.py,sha256=OtzgiYCxDoTdXp7y4mkLa2upj74qadesoqHYpr11ZCg,10061
-cuqi/experimental/mcmc/_sampler.py,sha256=S7sY7ORAPz2d7lvUhCRSg1PCbC745YXn4KVIqRYAZtw,20040
+cuqi/experimental/mcmc/_sampler.py,sha256=lWb0ORdkYcVkZ6G0ypHfmrWXZPfaXnsUj1wAncKxa7g,20339
 cuqi/experimental/mcmc/_utilities.py,sha256=kUzHbhIS3HYZRbneNBK41IogUYX5dS_bJxqEGm7TQBI,525
 cuqi/geometry/__init__.py,sha256=Tz1WGzZBY-QGH3c0GiyKm9XHN8MGGcnU6TUHLZkzB3o,842
 cuqi/geometry/_geometry.py,sha256=SDRZdiN2CIuS591lXxqgFoPWPIpwY-MHk75116QvdYY,46901
@@ -85,8 +85,8 @@ cuqi/testproblem/_testproblem.py,sha256=x769LwwRdJdzIiZkcQUGb_5-vynNTNALXWKato7s
 cuqi/utilities/__init__.py,sha256=H7xpJe2UinjZftKvE2JuXtTi4DqtkR6uIezStAXwfGg,428
 cuqi/utilities/_get_python_variable_name.py,sha256=QwlBVj2koJRA8s8pWd554p7-ElcI7HUwY32HknaR92E,1827
 cuqi/utilities/_utilities.py,sha256=Jc4knn80vLoA7kgw9FzXwKVFGaNBOXiA9kgvltZU3Ao,11777
-CUQIpy-1.2.0.post0.dev8.dist-info/LICENSE,sha256=kJWRPrtRoQoZGXyyvu50Uc91X6_0XRaVfT0YZssicys,10799
-CUQIpy-1.2.0.post0.dev8.dist-info/METADATA,sha256=NMx8mHLa7g4rr3PxGSYekxoBUWhE2m9MMyiyjMTXgYs,18409
-CUQIpy-1.2.0.post0.dev8.dist-info/WHEEL,sha256=GV9aMThwP_4oNCtvEC2ec3qUYutgWeAzklro_0m4WJQ,91
-CUQIpy-1.2.0.post0.dev8.dist-info/top_level.txt,sha256=AgmgMc6TKfPPqbjV0kvAoCBN334i_Lwwojc7HE3ZwD0,5
-CUQIpy-1.2.0.post0.dev8.dist-info/RECORD,,
+CUQIpy-1.2.0.post0.dev27.dist-info/LICENSE,sha256=kJWRPrtRoQoZGXyyvu50Uc91X6_0XRaVfT0YZssicys,10799
+CUQIpy-1.2.0.post0.dev27.dist-info/METADATA,sha256=e4sjOzPLNqifeVV_6sUvLx64pIOfhsZLw42i2Xi58UU,18410
+CUQIpy-1.2.0.post0.dev27.dist-info/WHEEL,sha256=GV9aMThwP_4oNCtvEC2ec3qUYutgWeAzklro_0m4WJQ,91
+CUQIpy-1.2.0.post0.dev27.dist-info/top_level.txt,sha256=AgmgMc6TKfPPqbjV0kvAoCBN334i_Lwwojc7HE3ZwD0,5
+CUQIpy-1.2.0.post0.dev27.dist-info/RECORD,,

cuqi/_version.py

@@ -8,11 +8,11 @@ import json
 
 version_json = '''
 {
- "date": "2024-09-30T09:16:05+0200",
+ "date": "2024-10-04T09:06:08+0200",
  "dirty": false,
  "error": null,
- "full-revisionid": "915f2f14ca7a4f588b561d25a1c4a47b8e295835",
- "version": "1.2.0.post0.dev8"
+ "full-revisionid": "f5a74472dfb8e7d2e07e829b2a03b811d554dc67",
+ "version": "1.2.0.post0.dev27"
 }
 '''  # END VERSION_JSON
 

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 +106,20 @@ 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 will be 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 +147,8 @@ 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:
@@ -236,6 +230,7 @@ class NUTS(Sampler):
         r_minus, r_plus = np.copy(r_k), np.copy(r_k)
 
         # run NUTS
+        acc = 0
         while (s == 1) and (j <= self.max_depth):
             # sample a direction
             v = int(2*(np.random.rand() < 0.5)-1)
@@ -263,9 +258,8 @@ class NUTS(Sampler):
                 self.current_point = point_prime
                 self.current_target_logd = logd_prime
                 self.current_target_grad = np.copy(grad_prime)
-                self._acc.append(1)
-            else:
-                self._acc.append(0)
+                acc = 1
+
 
             # update number of particles, tree level, and stopping criterion
             n += n_prime
@@ -273,8 +267,7 @@ 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(
@@ -284,6 +277,8 @@ class NUTS(Sampler):
         if np.isnan(self.current_target_logd):
             raise NameError('NaN potential func')
 
+        return acc
+
     def tune(self, skip_len, update_count):
         """ adapt epsilon during burn-in using dual averaging"""
         k = update_count+1
@@ -293,7 +288,7 @@ 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 +304,12 @@ 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
             

cuqi/experimental/mcmc/_sampler.py

@@ -220,7 +220,8 @@ class Sampler(ABC):
         if hasattr(self, "_pre_sample"): self._pre_sample()
 
         # Draw samples
-        for _ in tqdm( range(Ns), "Sample: "):
+        pbar = tqdm(range(Ns), "Sample: ")
+        for idx in pbar:
             
             # Perform one step of the sampler
             acc = self.step()
@@ -229,6 +230,9 @@ class Sampler(ABC):
             self._acc.append(acc)
             self._samples.append(self.current_point)
 
+            # display acc rate at progress bar
+            pbar.set_postfix_str(f"acc rate: {np.mean(self._acc[-1-idx:]):.2%}")
+
             # Add sample to batch
             if batch_size > 0:
                 batch_handler.add_sample(self.current_point)
@@ -260,7 +264,8 @@ class Sampler(ABC):
         if hasattr(self, "_pre_warmup"): self._pre_warmup()
 
         # Draw warmup samples with tuning
-        for idx in tqdm(range(Nb), "Warmup: "):
+        pbar = tqdm(range(Nb), "Warmup: ")
+        for idx in pbar:
 
             # Perform one step of the sampler
             acc = self.step()
@@ -273,6 +278,9 @@ class Sampler(ABC):
             self._acc.append(acc)
             self._samples.append(self.current_point)
 
+            # display acc rate at progress bar
+            pbar.set_postfix_str(f"acc rate: {np.mean(self._acc[-1-idx:]):.2%}")
+
             # Call callback function if specified
             self._call_callback(self.current_point, len(self._samples)-1)