fixup

72627e42 · Philipp Frank · 873e5a29 · 72627e42
Commit 72627e42 authored 2 years ago by Philipp Frank
--- a/demo_poisson.ipynb
+++ b/demo_poisson.ipynb
@@ -98,7 +98,7 @@
   "outputs": [],
   "source": [
    "# Compare evidence\n",
-    "print(evidences)"
+    "print(evidences[:2])"
   ]
  },
  {

 %% Cell type:markdown id: tags:

 Nifty tutorial for Poisson count data
 =====================================

 %% Cell type:markdown id: tags:

 Setup
 -----

 %% Cell type:code id: tags:

 ``` python
 import numpy as np
 import matplotlib.pyplot as plt
 import nifty8 as ift
 from utils import plot_2D, load_psf, geovi_sampling, plot_posterior

 ift.random.push_sseq_from_seed(42)
 evidences = np.zeros(3)
 ```

 %% Cell type:code id: tags:

 ``` python
 # Load data and visualize
 data = np.load('data/poisson.npz')
 ```

 %% Cell type:code id: tags:

 ``` python
 position_space = ift.RGSpace([128, 128])

 # Homogeneous poisson process


 print(model1)
 ```

 %% Cell type:code id: tags:

 ``` python
 # Set up likelihood & PSF
 ```

 %% Cell type:code id: tags:

 ``` python
 # Inference model 1


 print(evidence)
 evidences[0] = evidence
 ```

 %% Cell type:code id: tags:

 ``` python
 # Independent poisson process

 # Inference model 2


 evidences[1] = evidence
 ```

 %% Cell type:code id: tags:

 ``` python
 # Compare evidence
-print(evidences)
+print(evidences[:2])
 ```

 %% Cell type:code id: tags:

 ``` python
 # Diffuse poisson process
 args = {
    'offset_mean': .5,
    'offset_std': (1., 1E-5),

    # Amplitude of field fluctuations
    'fluctuations': (1.5, 0.5),  # 1.0, 1e-2

    # Exponent of power law power spectrum component
    'loglogavgslope': (-4., 1),  # -6.0, 1

    # Amplitude of integrated Wiener process power spectrum component
    'flexibility': (1., 0.2),  # 2.0, 1.0

    # How ragged the integrated Wiener process component is
    'asperity': (0.1, 0.01),  # 0.1, 0.5

    # Name of the input keys
    'prefix' : 'diffuse'
 }
 correlated_field = ift.SimpleCorrelatedField(position_space, **args)
 pspec = correlated_field.power_spectrum

 ```

 %% Cell type:code id: tags:

 ``` python
 # Prior samples
 ```

 %% Cell type:code id: tags:

 ``` python
 # Inference model 3

 evidences[2] = evidence
 ```

 %% Cell type:code id: tags:

 ``` python
 # Compare evidence
 print(evidences)
 print(evidences - evidences[-1])
 ```

 %% Cell type:markdown id: tags:

 Posterior visualization
 -----------------------

 %% Cell type:code id: tags:

 ``` python
 plot_posterior(samples, data, model3, diffuse, model2, pspec)
 ```