From c45dbae94a3a0bb03ab47fd78292b1c4547de458 Mon Sep 17 00:00:00 2001 From: Jan Boelts Date: Wed, 4 Sep 2024 11:53:13 +0200 Subject: [PATCH] fix #1260: include points in plotting limits --- sbi/analysis/plot.py | 58 ++++++++++++++++++------------ tutorials/00_getting_started.ipynb | 23 ++++++------ 2 files changed, 48 insertions(+), 33 deletions(-) diff --git a/sbi/analysis/plot.py b/sbi/analysis/plot.py index 45b072cea..e03fb2434 100644 --- a/sbi/analysis/plot.py +++ b/sbi/analysis/plot.py @@ -554,42 +554,56 @@ def handle_nan_infs(samples: List[np.ndarray]) -> List[np.ndarray]: return samples +def convert_to_list_of_numpy( + arr: Union[List[np.ndarray], List[torch.Tensor], np.ndarray, torch.Tensor], +) -> List[np.ndarray]: + """Converts a list of torch.Tensor to a list of np.ndarray.""" + if not isinstance(arr, list): + arr = ensure_numpy(arr) + return [arr] + return [ensure_numpy(a) for a in arr] + + +def infer_limits( + samples: List[np.ndarray], dim: int, points: Optional[List[np.ndarray]] = None +) -> List: + """Infer limits for the plot.""" + limits = [] + for d in range(dim): + min_val = min(np.min(sample[:, d]) for sample in samples) + max_val = max(np.max(sample[:, d]) for sample in samples) + if points is not None: + min_val = min(min_val, min(np.min(point[:, d]) for point in points)) + max_val = max(max_val, max(np.max(point[:, d]) for point in points)) + limits.append([min_val, max_val]) + return limits + + def prepare_for_plot( samples: Union[List[np.ndarray], List[torch.Tensor], np.ndarray, torch.Tensor], - limits: Optional[Union[List, torch.Tensor, np.ndarray]], + limits: Optional[Union[List, torch.Tensor, np.ndarray]] = None, + points: Optional[ + Union[List[np.ndarray], List[torch.Tensor], np.ndarray, torch.Tensor] + ] = None, ) -> Tuple[List[np.ndarray], int, torch.Tensor]: """ Ensures correct formatting for samples and limits, and returns dimension of the samples. """ - # Prepare samples - if not isinstance(samples, list): - samples = ensure_numpy(samples) - samples = [samples] - else: - samples = [ensure_numpy(sample) for sample in samples] + samples = convert_to_list_of_numpy(samples) + if points is not None: + points = convert_to_list_of_numpy(points) - # check if nans and infs samples = handle_nan_infs(samples) - # Dimensionality of the problem. dim = samples[0].shape[1] - # Prepare limits. Infer them from samples if they had not been passed. - if limits == [] or limits is None: - limits = [] - for d in range(dim): - min = +np.inf - max = -np.inf - for sample in samples: - min_ = np.min(sample[:, d]) - min = min_ if min_ < min else min - max_ = np.max(sample[:, d]) - max = max_ if max_ > max else max - limits.append([min, max]) + if limits is None or limits == []: + limits = infer_limits(samples, dim, points) else: limits = [limits[0] for _ in range(dim)] if len(limits) == 1 else limits + limits = torch.as_tensor(limits) return samples, dim, limits @@ -737,7 +751,7 @@ def pairplot( ) return fig, axes - samples, dim, limits = prepare_for_plot(samples, limits) + samples, dim, limits = prepare_for_plot(samples, limits, points) # prepate figure kwargs fig_kwargs_filled = _get_default_fig_kwargs() diff --git a/tutorials/00_getting_started.ipynb b/tutorials/00_getting_started.ipynb index 9c44f2636..a02e4123e 100644 --- a/tutorials/00_getting_started.ipynb +++ b/tutorials/00_getting_started.ipynb @@ -166,7 +166,7 @@ { "data": { "application/vnd.jupyter.widget-view+json": { - "model_id": "0b66b2e7830b43b7860e65e25c89b358", + "model_id": "deadabf3beb647c8b3a72ede380a2de8", "version_major": 2, "version_minor": 0 }, @@ -224,7 +224,7 @@ "name": "stdout", "output_type": "stream", "text": [ - " Neural network successfully converged after 158 epochs." + " Neural network successfully converged after 144 epochs." ] } ], @@ -252,7 +252,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "Posterior conditional density p(θ|x) of type DirectPosterior. It samples the posterior network and rejects samples that\n", + "Posterior p(θ|x) of type DirectPosterior. It samples the posterior network and rejects samples that\n", " lie outside of the prior bounds.\n" ] } @@ -305,7 +305,7 @@ { "data": { "application/vnd.jupyter.widget-view+json": { - "model_id": "1abb306e881d4c3090c1452c37da452b", + "model_id": "723f64befa4d491580b74ed759cb1a1f", "version_major": 2, "version_minor": 0 }, @@ -318,7 +318,7 @@ }, { "data": { - "image/png": "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", + "image/png": "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", "text/plain": [ "
" ] @@ -329,7 +329,8 @@ ], "source": [ "samples = posterior.sample((10000,), x=x_obs)\n", - "_ = pairplot(samples, limits=[[-2, 2], [-2, 2], [-2, 2]], figsize=(6, 6),labels=[r\"$\\theta_1$\", r\"$\\theta_2$\", r\"$\\theta_3$\"])" + "_ = pairplot(samples, limits=[[-2, 2], [-2, 2], [-2, 2]], figsize=(6, 6),labels=[r\"$\\theta_1$\", r\"$\\theta_2$\", r\"$\\theta_3$\"], \n", + " diag_kwargs=dict(mpl_kwargs=dict(bins=100)))" ] }, { @@ -354,7 +355,7 @@ { "data": { "application/vnd.jupyter.widget-view+json": { - "model_id": "14abd2077ac64e41a90fe42c63fbee57", + "model_id": "96bba6c5126e4df790a1f32a7e58c4e5", "version_major": 2, "version_minor": 0 }, @@ -367,7 +368,7 @@ }, { "data": { - "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfAAAAIgCAYAAAB6ejtUAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8hTgPZAAAACXBIWXMAAA9hAAAPYQGoP6dpAAAsW0lEQVR4nO3dfXBc9X3v8c/ZJ2m1Wj3YQjKWRXHiS1owtYOwXeNc6oADU8L1mEAg/GEeMiX1jM1AnSFm7gxjplMKM1CXEjN2b3oLjseASwfKYDJtweYaO2NjIgWC03uN04BD5foRa2WvVtqz5/zuHystFtigh12d/e2+XzPbSMfS9ovGq7d/Z8+DY4wxAgAAVgkFPQAAABg7Ag4AgIUIOAAAFiLgAABYiIADAGAhAg4AgIUIOAAAFiLgAABYKBL0AECxfSv03eI/qePICYclJ6RQfUJOfUKmJqpca4OyzTF5sZCyyZC8Gik8INX2egplfdWcHFD4yCmZwaxMf7/8zED++YwvcQ2lwLzuvzhp/79K8vexEjiOnEhUTjgkJ1EntUyRiceUnRJXf1tUftSRhl4iYdcoftxV5HRWoX5XoZO9MtmszMBg/jVl+etpvH8fWYEDE2Qc59x/cL7tAL6Y85n/PVtIUoh0SazAgdFxQlI4LCccllNbI9OQkF8bUaatVv2tIXkxR2695NVKkYyUqwsrPChJccXTg3KiESmXkzM4KOPbu1IAisYJyQmHpGhUTm2tvIa4comoMhdEdWZ6SH5M8iOSiRiFB0LyYjHFT4YUPRNR7MyAHM+TPE/KBP0fEhwCDoyCE3LkxGJywiGZ+jq5U+qUS0TUd1FY6Q5ffsyXGnKK1OSUTkc1eDKq8IAjPxxRtK9O4f6oQgOD0pm0nJAv4ym/Qrd4tx8wbo4jJ+Tk410Tk6mr1eDUWmWTIZ2ZEdKZWTk5dTnVxF0l44M63V+jvlhSXk1ENb0hRXrjCvl+/h/FISf/eqpCBBwYCyckhUIyEUd+xNGM+End6HVpRn+vjjoJvZb4mn5dP025gbBMKCSvxpEfCyvk+lI4HPT0QFlxHEdyQpre0KdrLvqVWhMpfRRv1vOxP9DR+no11WXUEk/reCShY3X1ytU6itQ4MpH861BOde9KJ+CYGGMktz//cbSust/39f38LjtjZBxHf9LyrtbEXpNJD71Vl5H+9Pg7+snlV2pd2x9rMB3TYG+Nso0RmZCj0KmYnGhExvPkSDKexyoc1Wco2pKkkKNv/cFv9ed/vFe+8q8j40h/+tHPtWnOXH08PakZsU/0m4E2PXe0SYP9tXK8kEw0XPj+akbAMTFuv/RX0/Mf/8/DUiwR7DwlYnwjxxgZY+RIaq/9RGs6XlNYZ8XX5A+a/cH7v9AvFs7QL2pnKHUiqmwy/8uqpiaWfx996PnkmPzRs0A1ckKa3pTWn1+9RyFn5BHVRtJd7/1Sv70krrb6Gv3fmv/U9tZLdLQvqtBgWH40LFPl8ZY4Ch0YPd/PP4zRn7T9SudaNw+vIK4/fPBzi4PzHq0OVKnrL/1Q5hyHmg+/jpoO5BQeet0Ykz+tzDGSjJHDwaCswIExM0YX1vSe8wyXYVP7+5XNheXkHIVco5Br5Pi+5Dj5kBuf1TeqlhNypJCjtoa0dM5/CucN9kV1PJfVR+6FSqXjipwJKZKWwoOe5Obyb2k5oaGq230u+HgQcGCUjMnvRnd8oyOZxi/4tSP11DTKN44cXwp5Ro5n8r9cQo509hGzVfYLBzj7wLOjfQkNrbc/92VG0om6uM540vFcUtnBiGoyjsKDRo7rDR2TUt3/CGYXOjAWQwey/duhS/O/dj7ze8cM/Z9najt15lSdoqcdhQeMwllfynmSb4g2qtvwniff6F9//XtyZM75OnKM9OK0OXrjzGV6u+8rMr0xRc9I0X4j5YYPKCXgAEbDzx/EZvoHdORwjf7m3SXy5ShnHHlylJMjX45+VL9U//kfHao7GFN9j1Eslb8EpDOQzR957vv5g9gIOaqRMfnXgfF1+FS9/ubNBTJylPMdeSb/evLl6IGG/6Gf/scf6dn3F+r//Or3lfxNWPX/6anuWE6hdEYml5Px/MJzVePriV3oKJrDqYymX1CZR6Hn/6U/dOqK78nJunrjN5fondr/put+71e6MNqrj6NN+sfY19VzeqpqUlIoaxRN+woN5uRkc4UD4EwV/qIBPssM/YP49QNf0a9TF+lbcw+prf60fhedoucartCHmSmKeJIfiyiacRRLGUXP+Iqkc/m9Wd7QQaVVjICjaG58arde/eH1am+KBz1KaRhf8kP5XXduTo7j6MSxKfp7fVN+1FE26circRQfNIqdzh+4VnMqp3A6K2Vdmaw7tALn9DFUOWPyb317npR1dfh4rTb9coH82ogGWut0ZnpEiZiGLoTkKDwgxU/lFOtz86+nXO7TVXeVrr4lAo4J+vXhPl029HHG9XQqna3MgBuTPxc85Mtk3fwtx3I5RY+cVjJdKz8allcXkRcLKZz1Fc7kFHJ9hdKDclJn8rv7BgYkz2P3OSBJxs/fUczNyfE8hUIhhWNRxQc9Rc/E5UdD8mpC8mP5Mzlqjw8o1JeRk3Xz3ze8+7yKEXBMyL8fThUCXg2MP3Q6mG8kL78rPZQJ54+KlRTKheS4vsL9ruT7+fe9c7n8iqEQb1bfwNn/KJbnybhu/iyPaEThTEQhN6SQG5YfCymU8+UMunLc3Kenj1X57nOJgAOjNxRe4/nS4GDhsrFO1pUTCslJR/PXO/f9/C8aY6TBrDQ4OPJgGwB5xpfxQ/n3szMD+dePMQr7RiYcUjgSlgmH5XienDMZmWx+97kZOPs1Vb17swg4MFpDvyhMzi3sunNyOTn9mfzdlaKRoYAbmbNj77r5o9eHow4gz+QvwGJy+Rg7IUdOLidl3fw1ExyncMGkwjEknic/60p+de8+lwg4MHZDv3Tk5A9oM8rfVclIhcs7Fo4097z8x1z2EfhSZuitKeXyB4mefT3i4Xgbw9tQwwg4MF7Du//83NAV1nwZuef4Gn7hAOdV2Cs1tNfKzeX3Ykkj7zY29DriINBPEXBgPD77S8eTRl4j9VxfC+C8hl8nxiu8BYUvxpXYgIn4sjgTb2DszBessnlNFbACByaKXyhAafDa+kKswAEAsBABx4Q0J2JBjwAAVYmAY0IubKzAy6YCgAUIOAAAFiLgAABYiIADAGAhAg4AgIUIOCbkk/7siM9PprPn+UoAQDERcIxbT29G9z3/y8Ln8WhYKzZ3qac3E+BUAFAdCDjG7VQ6q4z76fW///b2ryvjejrFKhwASo6AY9w+u9KeUsdFXQBgshBwjEtPb0Z/trkr6DEAoGoRcIwLu8kBIFgEHAAACxFwAAAsRMABALAQAQcAwEIEHAAACxFwTMg/rbgq6BEAoCoRcBQd10MHgNIj4Cia5kSU66EDwCQh4Cia6Y1xbVzeyfXQAWASEHAU1dQE10MHgMlAwAEAsBABBwDAQgQcAAALEXAAACxEwAEAsBABBwDAQgQcAAALEXAAACxEwAEAsBABBwDAQgQcAAALEXCUBLcUBYDSIuAoquZEjFuKAsAkIOAoqvYmbikKAJOBgKPouKUoAJQeAQcAwEIEHOPCQWoAECwCjjHr6c1oxeYuxaNhNSeiQY8DAFUpEvQAsM+pdFYZ19Om78/X9MZ40OMAQFViBY5x42A1AAgOAQcAwEIEHAAACxFwAAAsRMABALAQAQcAwEIEHAAACxFwAAAsRMABALAQAUfJcL10ACgdAo6ia07EFI+GtWJzl3p6M0GPAwAViYCj6Nqb4tq4vFMZ19MpVuEAUBIEHGM2ml3jXCcdAEqLgGNMRt5KlEgDQFAIOMZk+FaiG5d3qr2JW4kCQFAIOMaFXeQAECwCDgCAhQg4AAAWIuAAAFiIgAMAYCECDgCAhQg4AAAWIuAAAFiIgAMAYCECDgCAhQg4Sop7ggNAaRBwlAT3BAeA0iLgKAnuCQ4ApUXAUTLc8AQASoeAAwBgIQKOMeGgNAAoDwQco9bTm9GKzV2KR8NqZvc4AAQqEvQAsMepdFYZ19Om789Xe1M86HEAoKqxAseYcXAaAASPgAMAYCECDgCAhQg4AAAWIuAAAFiIgGPUxnsOOOeOA0DxEXCMynjOAT/7hiYAgOIi4BiV4XPANy7vHPU54Gff0AQAUFwEHGMy1nPAOWccAEqDgAMAYCECDgCAhQg4AAAWIuAAAFiIgAMAYCECDgCAhQg4AAAWIuAYFS6HCgDlhYDjS43nMqoAgNKKBD0Ayt/wZVQ3fX/+qC+jCgAoLVbgGDUuiwoA5YOAAwBgIQIOAICFCDgAABYi4PhSnEIGAOWHgOMLcQoZAJQnTiPDF+IUMgAoT6zAMSqcQgYA5YWAAwBgIQKOkmpOxBSPhoMeAwAqDgFHSbU3xbVxeWfQYwBAxSHgKDnePweA4iPg+EKcAw4A5YmA47w4BxwAyhfngeO8fnPsDOeAA0CZYgWOczp79T2rtT7ocQAAn8EKHOfEFdgAoLyxAsc5DR+8xhHkAFCeWIGjoKc3o10fHNeDL70vSRy8BgBlrCgBP9Y3oGOnB4vxVAjIyXRWd/7DvhHbXlyxkN3nAFCmHGOMCXoIAAAwNrwHDgCAhQg4AAAWIuAAAFiIgAMAYCECDgCAhSZ8GpkxRqdPny7GLKhwyWRSjuMEPQYAVIQJB/zEiRNqbW0txiyocMeOHdMFF1wQ9BgAUBEmHPBYLH+lro8//lgNDQ0THsh2fX196ujo4OdxluGfyfDfFQDAxE044MO7RBsaGgjWWfh5fB67zwGgeDiIDQAACxFwAAAsNOGA19TUaO3ataqpqSnGPNbj5/F5/EwAoPi4mQkAABZiFzoAABYi4AAAWIiAAwBgIQIOAICFJhRw13W1Zs0aXX755UokEpo+fbruuOMOHT58uFjzWefpp5/WxRdfrNraWi1YsED79u0LeqRAPProo5o3b56SyaRaW1u1bNkyHThwIOixAKBiTCjg/f396u7u1kMPPaTu7m699NJLOnDggJYuXVqs+ayydetWrV69WmvXrlV3d7fmzJmj66+/XseOHQt6tEm3c+dOrVy5Unv37tXrr78u13V13XXXKZ1OBz0aAFSEop9G9s4772j+/Pk6dOiQLrroomI+ddlbsGCB5s2bp/Xr10uSfN9XR0eH7r33Xj344IMBTxes48ePq7W1VTt37tTVV18d9DgAYL2ivweeSqXkOI6ampqK/dRlLZvNqqurS0uWLClsC4VCWrJkifbs2RPgZOUhlUpJkqZMmRLwJABQGYoa8IGBAa1Zs0a333571d3I48SJE/I8T21tbSO2t7W16ciRIwFNVR5839f999+vRYsWafbs2UGPAwAVYUwB37Jli+rr6wuPXbt2Ff7MdV3deuutMsZow4YNRR8U9lq5cqX279+vF154IehRAKBijOl2okuXLtWCBQsKn7e3t0v6NN6HDh3Sjh07qm71LUktLS0Kh8M6evToiO1Hjx7VtGnTApoqeKtWrdK2bdv01ltvacaMGUGPAwAVY0wr8GQyqVmzZhUe8Xi8EO+DBw/qjTfe0NSpU0s1a1mLxWLq7OzU9u3bC9t839f27du1cOHCACcLhjFGq1at0ssvv6wdO3Zo5syZQY8EABVlTCvwz3JdV7fccou6u7u1bds2eZ5XeL93ypQpisViRRnSFqtXr9add96pK6+8UvPnz9eTTz6pdDqtu+++O+jRJt3KlSv13HPP6ZVXXlEymSz8vWhsbFQ8Hg94OgCw34ROI/voo4/Ou7J68803tXjx4vE+tbXWr1+vxx9/XEeOHNHcuXP11FNPjXjboVo4jnPO7c8884zuuuuuyR0GACoQtxMFAMBCXAsdAAALTeg9cADAF/tW6LtBj4Ay97r/4ri+jxU4AAAWIuAAAFiIgAMAYCECDgCAhTiIDZgIYyS3P/9xtE46z/nvAFBsrMCBiXD7pb+ann8MhxwAJgEBBwDAQgR8gowxWrdunWbOnKm6ujotW7ZMqVQq6LEAABWOgE/QAw88oA0bNmjTpk3atWuXurq69PDDDwc9FgCgwhHwCXj77be1bt06bd26VVdffbU6Ozt1zz336Gc/+1nQowEAKhwBn4AnnnhC1157ra644orCtra2Np04cSLAqQAA1YCAj9Pg4KBee+013XTTTSO2DwwMqLGxMaCpEKTDqYx6ejNBjwGgShDwceru7lYmk9EPf/hD1dfXFx4/+tGPdMkll0iSbrrpJjU3N+uWW24JeFpMhhuf2q1Fj+3QxQ++pv09HMgIoLQI+Dh98MEHSiQSev/99/Xuu+8WHhdffLEWLVokSbrvvvv005/+NOBJMVkyrlf4mIADKDUCPk59fX1qaWnRrFmzCo9oNKqDBw/q5ptvliQtXrxYyWQy4EkRhOZELOgRAFQ4Aj5OLS0tSqVSMsYUtj3yyCO64YYbdOmllwY4GcpBe1M86BEAVDiuhT5O11xzjQYGBvTYY4/pe9/7nrZs2aJXX31V+/btC3o0AEAVYAU+Tm1tbXr22We1YcMGXXbZZdq7d692796tjo6OoEcDAFQBAj4Bt912m373u9+pv79f27Zt01e/+tWgR0LAHvvO5UGPAKBKsAu9hJYsWaL33ntP6XRaM2bM0IsvvqiFCxcGPRYAoAIQ8BJ64403gh4BAFCh2IUOAICFCDgAABYi4AAAWIiAAwBgIQIOAICFCDhQAjf+eDc3NAFQUgQcKJJ4NDziJiYEHEApcR44UCR/e/vX1dzITUwATA5W4ECRTKnjFqIAJg8BB4ro7F3o3BMcQCkRcKCI2pvi+rvlnYWPAaBUCDhQZIQbwGQg4AAAWIiAAwBgIQIOAICFCDgAABYi4AAAWIiAAwBgIQIOAICFCDhQIifT2aBHAFDBCDhQZM2JmOLRsFZs7lJPbybocQBUKAIOFFl7U1wbl3cq43o6xSocQIkQcKAEpnIjEwAlRsABALAQAQcAwEIEHAAACxFwAAAsRMABALAQAQcAwEIEHAAACxFwAAAsRMCBCTic+vRSqc2JaICTAKg2BByYgFNpt/Dx9MZ4gJMAqDYEHAAACxFwAAAsRMABALAQAQdK6CS3EwVQIgQcKIHmREzxaFgrNneppzfz5d8AAGNEwIESaG+Ka+PyTmVcT6dYhQMoAQIOlMjURCzoEQBUMAIOAICFCDgAABYi4AAAWIiAAwBgIQIOAICFCDgAABYi4AAAWIiAAwBgIQIOAICFCDgAABYi4AAAWIiAAwBgIQIOAICFCDgAABYi4AAAWIiAAyV2Mp0NegQAFYiAAyXSnIgpHg1rxeYu9fRmgh4HQIUh4MAEfNJ//tV1e1NcG5d3KuN6OsUqHECREXBgnHp6M7rv+V9+4ddMTcQmaRoA1YaAA+N0Kp1VxvWCHgNAlSLgAABYiIADAGAhAg4AgIUIOAAAFiLgAABYiIADAGAhAg4AgIUIOAAAFiLgAABYiIADAGAhAg4AgIUiQQ8AAMB5Oc7Iz40JZo4yxAocABA8x/n0cfa2c30dJLECBwCUC2doTekMf+p8us34Mr6RjB/MbGWIgAMAguXkQ10IdihfcCcSkUJDAXddyfMl48h4HrvSRcABAEEa2iV+drydcDi/LRyWwuH86tsL5aPthyR5AQ5cPgg4ACAYwyvvcFhObY2cSERObY1MIp5feUfCMtGw5BmF0hlpYFDGdWXOpGVyufwudb96Y07AAQCTyym8yS0n5MiJRhSqT0ixqPxkQm5LnfyIIz8Wkh9zFHKNYqdqFE5n5fQPyvF8aXBQcnMyxq/a3ekchQ5Mght/vFv7e1JBjwGUlcJu83BYikVlYlH5dVG5yYjchoiyDWENJsPKJsPK1UXkx6MyNTEpEsn/IyBU3UekswIHJsn+npRmtzcGPQZQHpyQnFgsv9s8WS/vgkZ58ajOdNQo9ZWQvFojPyb5MaPQgKP639UofjKqmt6Y6tKZ/K7zAcm4OUnVuQon4MA4nUxnv/RrmhOxc34MVDsn5OSPMq+pkamrVbapRm4yojPtIQ1+LaPauqzqalwlYlmd6o/rjJrkR0PyIxHV9sQVGhiUPE/KBP1fEhx2oQPj0NOb0YrNXYpHw1/4de1Ncf3d8s7Cx0DVO/tCLKGh3eChkPyakLwaR16tVFuXVVMio7b605pR36vW5Bn5wyvyqCOFnfxu9+FTz6oUK3BgHE6ls8q4nv7XHV+X/vGLv5ZwA5/hhArvfTuhkKZNPaPFX3lPrXUpfVjXoN80NKr2gqwuqf0vfTV2TO8PdOivj06V2xdXJOPIq4vJScfy54ZXMQIOTMCUOnaLA+MSyl829Vu//1vd/42fy1f+AmwmI4V2Gb39R9PUcOkZXRKtVcL5reoTAzpTE5cXk0zk0/jnr9RWnaeSsQsdADD5fKPpDX26f9FuhRyjiGMUdowiMnIkLdh7RP/v+DTtHZT2DczU6TP51Xd4QHJyvuQPHbhWxZdWZQUOjMNoDmADcA7GKH/UuKPrv/YfQ7keeQS5I8mXI/OrGv39ZX+s/0w3SUdqFDsl1fT5CmdcOVlXJpcL4D+gfBBwYIzOPoCtORENehzAWm0N/fpsvM+WPDOo351p1okzCYUzjsKDRuGs8qtv31TlqWNnI+DAGA0fwLbp+/M1vZED1IDxML7Rkd5a6RwrcA1t+W1fmz761XRFMo6SH0qJo55ip12FTmdkBoauxOZXb8QJODBOUzmvGxifod3o//LLdt0676CMGXl2mTH5rO/497ma5knhrKf44YzCvf35Xeep0zLZbP488CrGQWwAgEAc/qROf/2zOTJG8nyn8PDl6PHf3KDjJxsV6/MU7fMU6nflZAal4Qu4GCPDLnQAACaZMTKep397d7p+3TNVf9L5X2pr7NfR/gb9y+9m67/6m1UbSuWX5jlPzul0fre558kfzP+v8YeOQq/SkBNwYJJw5DrwGcZIxlPP8Rr979e/+um9wMNZOc6x/NeEHMk38nM5Gc/Lh3toBV7t2IUOlFhzIqZ4NKwVm7vU01vFF24Gzsf4+YdvRkTaeJ6Mmyt8rCo+YO1cCDhQYu1NcW1c3qmM6+kUq3Dg84Z2p5ucmw92NiszODjy4ebyf87qu4Bd6MAk4Ih14EsMR9l4MsYZuQ3nRMABAOWFcI8Ku9ABALAQAQcAwEIEHAAACxFwAAAsRMABALAQAQcAwEIEHAAACxFwYIy4pjmAckDAgTHo6c1oxeYuxaNhNY/j6mrEH0CxEHBgDE6ls8q4njYu71R7U3zU38cNTQAUGwEHxmGs1zbnhiYAio2AA2MwkV3g3NAEQDERcGCUJvr+99nPAwATRcCBURrv+9/DhsP9Z7wPDqAICDgwRuPdFX529HkfHMBEEXBgksxub9S2e78R9BgAKgQBB0apmOdwcz44gIki4MAo7O9JFeUAtrPPB9/fkyrihACqTSToAYBiO9Y3oGOnB4v2fCfTWa3Y3CVJenHFwnEdwDasvSmuF1cs1Hc37tF3N+7RxuWdnF4WgNntjUGPAEyYY4wxQQ8BAADGhl3oAABYiIADAGAhAg4AgIUIOAAAFiLgAABYiNPIUFGMMTp9+nTQY8ACyWRSjuMEPQYwbgQcFeXEiRNqbW0NegxY4NixY7rggguCHgMYNwKOihKL5S+K8vHHH6uhoSHgaYLX19enjo4Ofh5nGf6ZDP9dAWxFwFFRhneJNjQ0EKyz8PP4PHafw3YcxAYAgIUIOAAAFiLgqCg1NTVau3atampqgh6lLPDz+Dx+JqgU3MwEAAALsQIHAMBCBBwAAAsRcAAALETAAQCwEAEHAMBCBBwVw3VdrVmzRpdffrkSiYSmT5+uO+64Q4cPHw56tMA8/fTTuvjii1VbW6sFCxZo3759QY8UiEcffVTz5s1TMplUa2urli1bpgMHDgQ9FjAhBBwVo7+/X93d3XrooYfU3d2tl156SQcOHNDSpUuDHi0QW7du1erVq7V27Vp1d3drzpw5uv7663Xs2LGgR5t0O3fu1MqVK7V37169/vrrcl1X1113ndLpdNCjAePGeeCoaO+8847mz5+vQ4cO6aKLLgp6nEm1YMECzZs3T+vXr5ck+b6vjo4O3XvvvXrwwQcDni5Yx48fV2trq3bu3Kmrr7466HGAcWEFjoqWSqXkOI6ampqCHmVSZbNZdXV1acmSJYVtoVBIS5Ys0Z49ewKcrDykUilJ0pQpUwKeBBg/Ao6KNTAwoDVr1uj222+vujtxnThxQp7nqa2tbcT2trY2HTlyJKCpyoPv+7r//vu1aNEizZ49O+hxgHEj4LDWli1bVF9fX3js2rWr8Geu6+rWW2+VMUYbNmwIcEqUm5UrV2r//v164YUXgh4FmBDuBw5rLV26VAsWLCh83t7eLunTeB86dEg7duyoutW3JLW0tCgcDuvo0aMjth89elTTpk0LaKrgrVq1Stu2bdNbb72lGTNmBD0OMCGswGGtZDKpWbNmFR7xeLwQ74MHD+qNN97Q1KlTgx4zELFYTJ2dndq+fXthm+/72r59uxYuXBjgZMEwxmjVqlV6+eWXtWPHDs2cOTPokYAJYwWOiuG6rm655RZ1d3dr27Zt8jyv8H7vlClTFIvFAp5wcq1evVp33nmnrrzySs2fP19PPvmk0um07r777qBHm3QrV67Uc889p1deeUXJZLLw96KxsVHxeDzg6YDx4TQyVIyPPvrovCurN998U4sXL57cgcrA+vXr9fjjj+vIkSOaO3eunnrqqRFvO1QLx3HOuf2ZZ57RXXfdNbnDAEVCwAEAsBDvgQMAYCECDgCAhQg4AAAWIuAAAFiIgAMAYCHOAwcABMsYye3Pfxytk85z2h9GYgUOAAiW2y/91fT8Yzjk+FIEHAAACxFwAAAsRMCBCmKM0bp16zRz5kzV1dVp2bJlSqVSQY8FoAQIOFBBHnjgAW3YsEGbNm3Srl271NXVpYcffjjosQCUAAEHKsTbb7+tdevWaevWrbr66qvV2dmpe+65Rz/72c+CHg1ACRBwoEI88cQTuvbaa3XFFVcUtrW1tenEiRMBTgWgVAg4UAEGBwf12muv6aabbhqxfWBgQI2NjQFNBUzM/p6ULn7wNe3v4TiOcyHgQAXo7u5WJpPRD3/4Q9XX1xceP/rRj3TJJZfo448/1uLFi3XppZfqD//wD/Xiiy8GPTLwpYbDTcDPjSuxARXggw8+UCKR0Lvvvjti+7e//W0tWrRIkUhETz75pObOnasjR46os7NTN9xwgxKJRDADA5gwAg5UgL6+PrW0tGjWrFmFbYcOHdLBgwd1880368ILL9SFF14oSZo2bZpaWlr0ySefEHBY4WQ6G/QIZYld6EAFaGlpUSqVkjGmsO2RRx7RDTfcoEsvvXTE13Z1dcnzPHV0dEz2mMCYNCdikqTH//WAenozAU9TfliBAxXgmmuu0cDAgB577DF973vf05YtW/Tqq69q3759I77uk08+0R133KGf/OQnAU0KjF57U7zw8al0dsTnYAUOVIS2tjY9++yz2rBhgy677DLt3btXu3fvHrHKHhwc1LJly/Tggw/qqquuCnBaYHTYdf7FWIEDFeK2227Tbbfdds4/M8borrvu0jXXXKPly5dP8mTA2PX0ZrRic1fQY5Q1VuBAFfj5z3+urVu36p//+Z81d+5czZ07V++//37QYwHndSqdVcb19MD1Xwt6lLLFChyoAt/4xjfk+37QYwBjNnXoQDZ8HitwAAAsRMABALAQAQcAlJX9PSnd+OPdQY9R9gg4AKCscO3z0SHgAICy0nzWgWuHuQLbeRFwAEBZOfuKa9O5+tp5cRoZAKAsbbv3G0GPUNZYgQMAYCECDgCAhQg4AAAWIuAAAFiIgAMAYCECDgAoezf+eDcXePkMAg4AKFtnX9Slh4u6jEDAAQBlq70prr9b3ln4GJ8i4ACAska4z42AAwBgIQIOAICFCDgAABYi4AAAWIiAAwBgIQIOAICFCDgAABYi4AAAWIiAAwBgIQIOAICFCDgAABYi4AAAWIiAAwBgIQIOACgrJ9PZoEewAgEHAJSNw6mMVmzuUjwaVnMiFvQ4ZS0S9AAAAAw7lXaVcT1t+v587gP+JViBAwDKzlRW31+KgAMAYCECDgCAhQg4AAAWIuAAACtwetlIBBwAUNaaEzHFo2Gt2Nylnt5M0OOUDQIOAChr7U1xbVzeqYzr6RSr8AICDgAoe5xW9nkEHAAACxFwAAAsRMABALAQAQcAwEIEHAAACxFwAAAsRMABALAQAQcAwEIEHAAACxFwAAAsRMABALAQAQcAwEIEHABQNj7p525jo0XAAQBl477nf6l4NKxm7j72pQg4AKBsZFxPG5d3qr0pHvQoZY+AAwDKCvf+Hh0CDgCAhQg4AAAWIuAAAFiIgAMAYCECDgCAhQg4AAAWIuAAAFiIgAMAYCECDgCAhQg4AKBscB300SPgAICy8be3f/0Lr4N+Ms3dyoYRcABA2ZhSd+7Vd3Mipng0rBWbu9TTm5nkqcoTAQcAlL32prg2Lu9UxvV0ilW4JAIOALAEdykbiYADAGAhAg4AgIUIOAAAFiLgAABYiIADAGAhAg4AsAoXc8kj4AAAK3Axl5EIOADAClzMZSQCDgCwBhdz+RQBBwDAQgQcAAALEXAAACxEwAEAsBABBwDAQgQcAAALEXAAACxEwAEAsBABBwAE6nDq08uiNieiAU5iFwIOAAjUqbRb+Hh6YzzASexCwAEAsBABBwDAQgQcAAALEXAAACxEwAEAsBABBwBY52Q6G/QIgSPgAABrNCdiikfDWrG5Sz29mS//hgpGwAEA1mhvimvj8k5lXE+nqnwVTsABAFaZmogFPUJZIOAAAFiIgAMAAvVJf3XvCh8vAg4ACExPb0b3Pf/LoMewEgEHAATmVDqrjOsFPYaVCDgAABYi4AAAWIiAAwBgIQIOAAhMtV9NbSIIOAAgED29Gf3Z5q5xf3+1Xw+dgAMAAjHeS6FyPfQ8Ag4ACNQ/rbhqTF/P9dDzCDgAwDpcD52AAwBgJQIOAJh0Oz84rht/vHvCz1PNB7IRcADApOrpzejOf9hX+Lw5ER3zc5x9INv+nlQxx7NGJOgBAKCSHesb0LHTg0GPUVZ+c+yMJOmx71yu/37JBZpe54/5Odqb4npxxUJ9d+MefXfjHm1c3mnt++Kz2xvH9X2OMcYUeRYAAFBi7EIHAMBCBBwAAAsRcAAALETAAQCwEAEHAMBCnEYGACVijNHp06eDHgMWSCaTchxnTN9DwAGgRE6cOKHW1tagx4AFjh07pgsuuGBM30PAAaBEYrH8hUU+/vhjNTQ0BDxN8Pr6+tTR0cHP4yzDP5PhvytjQcABoESGd4k2NDQQrLPw8/i8se4+lziIDQAAKxFwAAAsRMABoERqamq0du1a1dTUBD1KWeDn8XkT+ZlwMxMAACzEChwAAAsRcAAALETAAQCwEAEHAMBCBBwAAAsRcAAoAdd1tWbNGl1++eVKJBKaPn267rjjDh0+fDjo0QLz9NNP6+KLL1Ztba0WLFigffv2BT1SIB599FHNmzdPyWRSra2tWrZsmQ4cODDm5yHgAFAC/f396u7u1kMPPaTu7m699NJLOnDggJYuXRr0aIHYunWrVq9erbVr16q7u1tz5szR9ddfr2PHjgU92qTbuXOnVq5cqb179+r111+X67q67rrrlE6nx/Q8nAcOAJPknXfe0fz583Xo0CFddNFFQY8zqRYsWKB58+Zp/fr1kiTf99XR0aF7771XDz74YMDTBev48eNqbW3Vzp07dfXVV4/6+1iBA8AkSaVSchxHTU1NQY8yqbLZrLq6urRkyZLCtlAopCVLlmjPnj0BTlYeUqmUJGnKlClj+j4CDgCTYGBgQGvWrNHtt99edXfiOnHihDzPU1tb24jtbW1tOnLkSEBTlQff93X//fdr0aJFmj179pi+l4ADQBFs2bJF9fX1hceuXbsKf+a6rm699VYZY7Rhw4YAp0S5Wblypfbv368XXnhhzN/L/cABoAiWLl2qBQsWFD5vb2+X9Gm8Dx06pB07dlTd6luSWlpaFA6HdfTo0RHbjx49qmnTpgU0VfBWrVqlbdu26a233tKMGTPG/P2swAGgCJLJpGbNmlV4xOPxQrwPHjyoN954Q1OnTg16zEDEYjF1dnZq+/bthW2+72v79u1auHBhgJMFwxijVatW6eWXX9aOHTs0c+bMcT0PK3AAKAHXdXXLLbeou7tb27Ztk+d5hfd7p0yZolgsFvCEk2v16tW68847deWVV2r+/Pl68sknlU6ndffddwc92qRbuXKlnnvuOb3yyitKJpOFvxeNjY2Kx+Ojfh5OIwOAEvjoo4/Ou7J68803tXjx4skdqAysX79ejz/+uI4cOaK5c+fqqaeeGvG2Q7VwHOec25955hndddddo38eAg4AgH14DxwAAAsRcAAALETAAQCwEAEHAMBCBBwAAAsRcAAALETAAQCwEAEHAMBCBBwAAAsRcABASRhjtG7dOs2cOVN1dXVatmyZUqlU0GNVDAIOACiJBx54QBs2bNCmTZu0a9cudXV16eGHHw56rIrBtdABAEX39ttva+HChfrFL36hK664QpL0F3/xF9qyZYsOHDgQ8HSVgRU4AKDonnjiCV177bWFeEtSW1ubTpw4EeBUlYWAAwCKanBwUK+99ppuuummEdsHBgbU2NgY0FSVh13oAICi2rNnj6666irV1tYqHA4Xtruuq29+85t64YUXtGTJEuVyOeVyOd1333265557ApzYTpGgBwAAVJYPPvhAiURC77777ojt3/72t7Vo0SIlk0m99dZbqqurUzqd1uzZs/Wd73xHU6dODWZgSxFwAEBR9fX1qaWlRbNmzSpsO3TokA4ePKibb75Z4XBYdXV1kvK7240xYmfw2PEeOACgqFpaWpRKpUZE+ZFHHtENN9ygSy+9VJLU29urOXPmaMaMGXrggQfU0tIS1LjWIuAAgKK65pprNDAwoMcee0wffvih/vIv/1KvvvqqNmzYUPiapqYmvffee/rwww/13HPP6ejRowFObCcCDgAoqra2Nj377LPasGGDLrvsMu3du1e7d+9WR0fHOb92zpw52rVrVwCT2o2j0AEAk+ro0aOqq6tTMplUKpXSokWL9Pzzz+vyyy8PejSrcBAbAGBSHTp0SD/4wQ8KB6/de++9xHscWIEDAGAh3gMHAMBCBBwAAAsRcAAALETAAQCwEAEHAMBCBBwAAAsRcAAALETAAQCwEAEHAMBCBBwAAAsRcAAALPT/AY7M9G9OaXYKAAAAAElFTkSuQmCC", + "image/png": "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", "text/plain": [ "
" ] @@ -409,9 +410,9 @@ "name": "stdout", "output_type": "stream", "text": [ - "high for true theta : tensor([2.7324])\n", - "low for different theta : tensor([-201.8379])\n", - "range of posterior samples: min: tensor(-6.6965) max : tensor(4.2212)\n" + "high for true theta : tensor([2.5505])\n", + "low for different theta : tensor([-251.4675])\n", + "range of posterior samples: min: tensor(-9.2097) max : tensor(4.0836)\n" ] } ],