update week 46

Author: mhjensen

doc/pub/week46/html/._week46-bs001.html

@@ -323,7 +323,10 @@ <h2 id="plan-for-week-46" class="anchor">Plan for week 46 </h2>
 <div class="panel-body">
 <!-- subsequent paragraphs come in larger fonts, so start with a paragraph -->
 <ol>
- <li> Work on and discussions of project 3</li>
+<li> Intro to and mathematics of  Recurrent Neural Networks (RNNs)</li>
+<li> Lecture notes at <a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/pub/week46/ipynb/week46.ipynb" target="_self"><tt>https://github.com/CompPhysics/MachineLearning/blob/master/doc/pub/week46/ipynb/week46.ipynb</tt></a>
+<!-- o Video of lecture at <a href="https://youtu.be/TfRKUfdJwB4" target="_self"><tt>https://youtu.be/TfRKUfdJwB4</tt></a> -->
+<!-- o Whiteboard notes at <a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2024/NotesNovember11.pdf" target="_self"><tt>https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2024/NotesNovember11.pdf</tt></a> --></li>
 </ol>
 </div>
 </div>
@@ -333,10 +336,7 @@ <h2 id="plan-for-week-46" class="anchor">Plan for week 46 </h2>
 <div class="panel-body">
 <!-- subsequent paragraphs come in larger fonts, so start with a paragraph -->
 <ol>
-<li> Intro to and mathematics of  Recurrent Neural Networks (RNNs)</li>
-<li> Lecture notes at <a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/pub/week46/ipynb/week46.ipynb" target="_self"><tt>https://github.com/CompPhysics/MachineLearning/blob/master/doc/pub/week46/ipynb/week46.ipynb</tt></a>
-<!-- o Video of lecture at <a href="https://youtu.be/TfRKUfdJwB4" target="_self"><tt>https://youtu.be/TfRKUfdJwB4</tt></a> -->
-<!-- o Whiteboard notes at <a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2024/NotesNovember11.pdf" target="_self"><tt>https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2024/NotesNovember11.pdf</tt></a> --></li>
+<li> Work on and discussions of project 3, project available from November 10,(evening), deadline December 15 at midnight.</li>
 </ol>
 </div>
 </div>

doc/pub/week46/html/._week46-bs013.html

@@ -447,15 +447,6 @@ <h2 id="corresponding-example-using-pytorch" class="anchor">Corresponding exampl
        <span style="color: #408080; font-style: italic"># Update sequence by removing first element and adding new prediction</span>
        current_sequence <span style="color: #666666">=</span> torch<span style="color: #666666">.</span>cat([current_sequence[:, <span style="color: #666666">1</span>:, :], pred<span style="color: #666666">.</span>unsqueeze(<span style="color: #666666">1</span>)], dim<span style="color: #666666">=1</span>)
 
-<span style="color: #408080; font-style: italic"># Plot results</span>
-plt<span style="color: #666666">.</span>figure(figsize<span style="color: #666666">=</span>(<span style="color: #666666">12</span>, <span style="color: #666666">6</span>))
-plt<span style="color: #666666">.</span>plot(t[<span style="color: #666666">800</span>:]<span style="color: #666666">.</span>numpy(), val_data<span style="color: #666666">.</span>numpy(), label<span style="color: #666666">=</span><span style="color: #BA2121">&#39;True values&#39;</span>)
-plt<span style="color: #666666">.</span>plot(t[<span style="color: #666666">800</span>:]<span style="color: #666666">.</span>numpy(), predictions, label<span style="color: #666666">=</span><span style="color: #BA2121">&#39;Predictions&#39;</span>)
-plt<span style="color: #666666">.</span>title(<span style="color: #BA2121">&#39;Sine Wave Prediction&#39;</span>)
-plt<span style="color: #666666">.</span>xlabel(<span style="color: #BA2121">&#39;Time&#39;</span>)
-plt<span style="color: #666666">.</span>ylabel(<span style="color: #BA2121">&#39;Value&#39;</span>)
-plt<span style="color: #666666">.</span>legend()
-plt<span style="color: #666666">.</span>show()
 
 <span style="color: #408080; font-style: italic"># Plot training and validation loss</span>
 plt<span style="color: #666666">.</span>figure(figsize<span style="color: #666666">=</span>(<span style="color: #666666">10</span>, <span style="color: #666666">5</span>))

doc/pub/week46/html/week46-reveal.html

@@ -198,22 +198,22 @@ <h4>Week 46, November 10-14</h4>
 <h2 id="plan-for-week-46">Plan for week 46 </h2>
 
 <div class="alert alert-block alert-block alert-text-normal">
-<b>Lab sessions on Tuesday and Wednesday</b>
+<b>Material for the lecture on Monday November 10, 2025</b>
 <p>
 <ol>
- <p><li> Work on and discussions of project 3</li>
+<p><li> Intro to and mathematics of  Recurrent Neural Networks (RNNs)</li>
+<p><li> Lecture notes at <a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/pub/week46/ipynb/week46.ipynb" target="_blank"><tt>https://github.com/CompPhysics/MachineLearning/blob/master/doc/pub/week46/ipynb/week46.ipynb</tt></a>
+<!-- o Video of lecture at <a href="https://youtu.be/TfRKUfdJwB4" target="_blank"><tt>https://youtu.be/TfRKUfdJwB4</tt></a> -->
+<!-- o Whiteboard notes at <a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2024/NotesNovember11.pdf" target="_blank"><tt>https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2024/NotesNovember11.pdf</tt></a> --></li>
 </ol>
 </div>
 
 
 <div class="alert alert-block alert-block alert-text-normal">
-<b>Material for the lecture on Monday November 10, 2025</b>
+<b>Lab sessions on Tuesday and Wednesday</b>
 <p>
 <ol>
-<p><li> Intro to and mathematics of  Recurrent Neural Networks (RNNs)</li>
-<p><li> Lecture notes at <a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/pub/week46/ipynb/week46.ipynb" target="_blank"><tt>https://github.com/CompPhysics/MachineLearning/blob/master/doc/pub/week46/ipynb/week46.ipynb</tt></a>
-<!-- o Video of lecture at <a href="https://youtu.be/TfRKUfdJwB4" target="_blank"><tt>https://youtu.be/TfRKUfdJwB4</tt></a> -->
-<!-- o Whiteboard notes at <a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2024/NotesNovember11.pdf" target="_blank"><tt>https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2024/NotesNovember11.pdf</tt></a> --></li>
+<p><li> Work on and discussions of project 3, project available from November 10,(evening), deadline December 15 at midnight.</li>
 </ol>
 </div>
 </section>
@@ -592,15 +592,6 @@ <h2 id="corresponding-example-using-pytorch">Corresponding example using PyTorch
        <span style="color: #228B22"># Update sequence by removing first element and adding new prediction</span>
        current_sequence = torch.cat([current_sequence[:, <span style="color: #B452CD">1</span>:, :], pred.unsqueeze(<span style="color: #B452CD">1</span>)], dim=<span style="color: #B452CD">1</span>)
 
-<span style="color: #228B22"># Plot results</span>
-plt.figure(figsize=(<span style="color: #B452CD">12</span>, <span style="color: #B452CD">6</span>))
-plt.plot(t[<span style="color: #B452CD">800</span>:].numpy(), val_data.numpy(), label=<span style="color: #CD5555">&#39;True values&#39;</span>)
-plt.plot(t[<span style="color: #B452CD">800</span>:].numpy(), predictions, label=<span style="color: #CD5555">&#39;Predictions&#39;</span>)
-plt.title(<span style="color: #CD5555">&#39;Sine Wave Prediction&#39;</span>)
-plt.xlabel(<span style="color: #CD5555">&#39;Time&#39;</span>)
-plt.ylabel(<span style="color: #CD5555">&#39;Value&#39;</span>)
-plt.legend()
-plt.show()
 
 <span style="color: #228B22"># Plot training and validation loss</span>
 plt.figure(figsize=(<span style="color: #B452CD">10</span>, <span style="color: #B452CD">5</span>))
@@ -1740,6 +1731,7 @@ <h2 id="the-runge-kutta-4-code">The Runge-Kutta-4 code </h2>
 <section>
 <h2 id="using-the-above-data-to-train-an-rnn">Using the above data to train an RNN </h2>
 
+<p>In the code here we have reworked the previous example in order to generate data that can be handled by recurrent neural networks in order to train our model.</p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
 <div class="cell border-box-sizing code_cell rendered">
@@ -1754,18 +1746,16 @@ <h2 id="using-the-above-data-to-train-an-rnn">Using the above data to train an R
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
 
 
-<span style="color: #228B22"># ============================</span>
-<span style="color: #228B22"># 1. Physics model (from rk4.py)</span>
-<span style="color: #228B22"># ============================</span>
+<span style="color: #228B22"># Newton&#39;s equation for harmonic oscillations with external force</span>
 
-<span style="color: #228B22"># Global parameters (same idea as in rk4.py)</span>
+<span style="color: #228B22"># Global parameters</span>
 gamma = <span style="color: #B452CD">0.2</span>        <span style="color: #228B22"># damping</span>
 Omegatilde = <span style="color: #B452CD">0.5</span>   <span style="color: #228B22"># driving frequency</span>
 Ftilde = <span style="color: #B452CD">1.0</span>       <span style="color: #228B22"># driving amplitude</span>
 
 <span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">spring_force</span>(v, x, t):
     <span style="color: #CD5555">&quot;&quot;&quot;</span>
-<span style="color: #CD5555">    SpringForce from rk4.py:</span>
+<span style="color: #CD5555">    SpringForce:</span>
 <span style="color: #CD5555">    note: divided by mass =&gt; returns acceleration</span>
 <span style="color: #CD5555">    a = -2*gamma*v - x + Ftilde*cos(Omegatilde * t)</span>
 <span style="color: #CD5555">    &quot;&quot;&quot;</span>
@@ -1774,7 +1764,6 @@ <h2 id="using-the-above-data-to-train-an-rnn">Using the above data to train an R
 
 <span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">rk4_trajectory</span>(DeltaT=<span style="color: #B452CD">0.001</span>, tfinal=<span style="color: #B452CD">20.0</span>, x0=<span style="color: #B452CD">1.0</span>, v0=<span style="color: #B452CD">0.0</span>):
     <span style="color: #CD5555">&quot;&quot;&quot;</span>
-<span style="color: #CD5555">    Reimplementation of RK4 integrator from rk4.py.</span>
 <span style="color: #CD5555">    Returns t, x, v arrays.</span>
 <span style="color: #CD5555">    &quot;&quot;&quot;</span>
     n = <span style="color: #658b00">int</span>(np.ceil(tfinal / DeltaT))
@@ -1817,9 +1806,7 @@ <h2 id="using-the-above-data-to-train-an-rnn">Using the above data to train an R
     <span style="color: #8B008B; font-weight: bold">return</span> t, x, v
 
 
-<span style="color: #228B22"># =====================================</span>
-<span style="color: #228B22"># 2. Sequence generation for RNN training</span>
-<span style="color: #228B22"># =====================================</span>
+<span style="color: #228B22"># Sequence generation for RNN training</span>
 
 <span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">create_sequences</span>(x, seq_len):
     <span style="color: #CD5555">&quot;&quot;&quot;</span>
@@ -1862,9 +1849,7 @@ <h2 id="using-the-above-data-to-train-an-rnn">Using the above data to train an R
         <span style="color: #8B008B; font-weight: bold">return</span> <span style="color: #658b00">self</span>.inputs[idx], <span style="color: #658b00">self</span>.targets[idx]
 
 
-<span style="color: #228B22"># ==============================</span>
-<span style="color: #228B22"># 3. RNN model (LSTM-based)</span>
-<span style="color: #228B22"># ==============================</span>
+<span style="color: #228B22"># RNN model (LSTM-based in this example)</span>
 
 <span style="color: #8B008B; font-weight: bold">class</span> <span style="color: #008b45; font-weight: bold">RNNPredictor</span>(nn.Module):
     <span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">__init__</span>(<span style="color: #658b00">self</span>, input_size=<span style="color: #B452CD">1</span>, hidden_size=<span style="color: #B452CD">32</span>, num_layers=<span style="color: #B452CD">1</span>, output_size=<span style="color: #B452CD">1</span>):
@@ -1882,9 +1867,7 @@ <h2 id="using-the-above-data-to-train-an-rnn">Using the above data to train an R
         <span style="color: #8B008B; font-weight: bold">return</span> out
 
 
-<span style="color: #228B22"># ==============================</span>
-<span style="color: #228B22"># 4. Training loop</span>
-<span style="color: #228B22"># ==============================</span>
+<span style="color: #228B22"># Training part</span>
 
 <span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">train_model</span>(
     seq_len=<span style="color: #B452CD">50</span>,
@@ -1933,9 +1916,7 @@ <h2 id="using-the-above-data-to-train-an-rnn">Using the above data to train an R
     <span style="color: #8B008B; font-weight: bold">return</span> model, dataset
 
 
-<span style="color: #228B22"># ==============================</span>
-<span style="color: #228B22"># 5. Evaluation / visualization</span>
-<span style="color: #228B22"># ==============================</span>
+<span style="color: #228B22"># Evaluation / visualization</span>
 
 <span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">evaluate_and_plot</span>(model, dataset, seq_len=<span style="color: #B452CD">50</span>, device=<span style="color: #8B008B; font-weight: bold">None</span>):
     <span style="color: #8B008B; font-weight: bold">if</span> device <span style="color: #8B008B">is</span> <span style="color: #8B008B; font-weight: bold">None</span>:
@@ -1946,13 +1927,10 @@ <h2 id="using-the-above-data-to-train-an-rnn">Using the above data to train an R
         <span style="color: #228B22"># Take a single sequence from the dataset</span>
         x_seq, y_seq = dataset[<span style="color: #B452CD">0</span>]  <span style="color: #228B22"># shapes: (seq_len, 1), (seq_len, 1)</span>
         x_input = x_seq.unsqueeze(<span style="color: #B452CD">0</span>).to(device)  <span style="color: #228B22"># (1, seq_len, 1)</span>
-
         <span style="color: #228B22"># Model prediction (next-step for whole sequence)</span>
         y_pred = model(x_input).cpu().numpy().squeeze(-<span style="color: #B452CD">1</span>).squeeze(<span style="color: #B452CD">0</span>)  <span style="color: #228B22"># (seq_len,)</span>
-
         <span style="color: #228B22"># True target</span>
         y_true = y_seq.numpy().squeeze(-<span style="color: #B452CD">1</span>)  <span style="color: #228B22"># (seq_len,)</span>
-
         <span style="color: #228B22"># Plot comparison</span>
         plt.figure()
         plt.plot(y_true, label=<span style="color: #CD5555">&quot;True x(t+Δt)&quot;</span>, linestyle=<span style="color: #CD5555">&quot;-&quot;</span>)
@@ -1964,10 +1942,7 @@ <h2 id="using-the-above-data-to-train-an-rnn">Using the above data to train an R
         plt.tight_layout()
         plt.show()
 
-
-<span style="color: #228B22"># ==============================</span>
-<span style="color: #228B22"># 6. Main</span>
-<span style="color: #228B22"># ==============================</span>
+<span style="color: #228B22"># This is the main part of the code where we define the network</span>
 
 <span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #00688B">__name__</span> == <span style="color: #CD5555">&quot;__main__&quot;</span>:
     <span style="color: #228B22"># Hyperparameters can be tweaked as you like</span>