<?xml version="1.0" encoding="UTF-8" ?>
<rss version="2.0">
    <channel>
      <title>notes</title>
      <link>https://waltdhill.github.io/notes</link>
      <description>Last 10 notes on notes</description>
      <generator>Quartz -- quartz.jzhao.xyz</generator>
      <item>
    <title>Implementations</title>
    <link>https://waltdhill.github.io/notes/implementations/</link>
    <guid>https://waltdhill.github.io/notes/implementations/</guid>
    <description><![CDATA[ Hands-on reimplementations with annotated notebooks. Each entry below is a self-contained notebook. ]]></description>
    <pubDate>Wed, 13 May 2026 00:11:57 GMT</pubDate>
  </item><item>
    <title>01_jpeg_compression</title>
    <link>https://waltdhill.github.io/notes/implementations/jpeg-compression/01_jpeg_compression</link>
    <guid>https://waltdhill.github.io/notes/implementations/jpeg-compression/01_jpeg_compression</guid>
    <description><![CDATA[ Related: jpeg-compression, discrete-cosine-transform, fourier-transform import numpy as np import matplotlib.pyplot as plt from PIL import Image from pathlib import Path %matplotlib inline IMAGE_PATH = &quot;path/to/image.png&quot; # override, or drop an image into this folder 1. ]]></description>
    <pubDate>Wed, 13 May 2026 00:11:57 GMT</pubDate>
  </item><item>
    <title>01_derivatives_and_autograd</title>
    <link>https://waltdhill.github.io/notes/implementations/karpathy-01-micrograd/01_derivatives_and_autograd</link>
    <guid>https://waltdhill.github.io/notes/implementations/karpathy-01-micrograd/01_derivatives_and_autograd</guid>
    <description><![CDATA[  Next: 02_nn_data_structs_and_forward_pass Related: backprop-graph-terminology # imports import math import numpy as np import matplotlib.pyplot as plt %matplotlib inline Micrograd Micrograd is scalar-valued autograd engine which implement backpropagation. ]]></description>
    <pubDate>Wed, 13 May 2026 00:11:57 GMT</pubDate>
  </item><item>
    <title>02_nn_data_structs_and_forward_pass</title>
    <link>https://waltdhill.github.io/notes/implementations/karpathy-01-micrograd/02_nn_data_structs_and_forward_pass</link>
    <guid>https://waltdhill.github.io/notes/implementations/karpathy-01-micrograd/02_nn_data_structs_and_forward_pass</guid>
    <description><![CDATA[  Prev: 01_derivatives_and_autograd Next: 03_backpropagation # imports import math import numpy as np import matplotlib.pyplot as plt %matplotlib inline A note on graph direction terminology See backprop-graph-terminology note Transclude of backprop-graph-terminology#graph-direction-terminology-is-ge... ]]></description>
    <pubDate>Wed, 13 May 2026 00:11:57 GMT</pubDate>
  </item><item>
    <title>03_backpropagation</title>
    <link>https://waltdhill.github.io/notes/implementations/karpathy-01-micrograd/03_backpropagation</link>
    <guid>https://waltdhill.github.io/notes/implementations/karpathy-01-micrograd/03_backpropagation</guid>
    <description><![CDATA[  Prev: 02_nn_data_structs_and_forward_pass Next: 04_backprop_train_a_neuron Related: Value object data structure, computation graph direction terminology # imports, `Value` class, init toy nn, graphviz: trace() &amp; draw_dot() import math import numpy as np import matplotlib.pyplot as plt %matplotl... ]]></description>
    <pubDate>Wed, 13 May 2026 00:11:57 GMT</pubDate>
  </item><item>
    <title>04_backprop_train_a_neuron</title>
    <link>https://waltdhill.github.io/notes/implementations/karpathy-01-micrograd/04_backprop_train_a_neuron</link>
    <guid>https://waltdhill.github.io/notes/implementations/karpathy-01-micrograd/04_backprop_train_a_neuron</guid>
    <description><![CDATA[  Prev: 03_backpropagation Next: 05_backprop_backward_method Related: Value object data structure, computation graph direction terminology # imports, `Value` class, graphviz: trace() &amp; draw_dot() import math import numpy as np import matplotlib.pyplot as plt %matplotlib inline # Object definition... ]]></description>
    <pubDate>Wed, 13 May 2026 00:11:57 GMT</pubDate>
  </item><item>
    <title>05_backprop_backward_method</title>
    <link>https://waltdhill.github.io/notes/implementations/karpathy-01-micrograd/05_backprop_backward_method</link>
    <guid>https://waltdhill.github.io/notes/implementations/karpathy-01-micrograd/05_backprop_backward_method</guid>
    <description><![CDATA[  Prev: 04_backprop_train_a_neuron Next: 06_backprop_backward_method_auto Related: Value object data structure, computation graph direction terminology # imports, graphviz: trace() &amp; draw_dot() import math import numpy as np import matplotlib.pyplot as plt %matplotlib inline # Object definitions ... ]]></description>
    <pubDate>Wed, 13 May 2026 00:11:57 GMT</pubDate>
  </item><item>
    <title>06_backprop_backward_method_auto</title>
    <link>https://waltdhill.github.io/notes/implementations/karpathy-01-micrograd/06_backprop_backward_method_auto</link>
    <guid>https://waltdhill.github.io/notes/implementations/karpathy-01-micrograd/06_backprop_backward_method_auto</guid>
    <description><![CDATA[  Prev: 05_backprop_backward_method Next: 07_breaking_up_tanh Related: Value object data structure, computation graph direction terminology # imports, `Value` class, reset_graph() to init nn, graphviz: trace() &amp; draw_dot() import math import numpy as np import matplotlib.pyplot as plt %matplotlib... ]]></description>
    <pubDate>Wed, 13 May 2026 00:11:57 GMT</pubDate>
  </item><item>
    <title>07_breaking_up_tanh</title>
    <link>https://waltdhill.github.io/notes/implementations/karpathy-01-micrograd/07_breaking_up_tanh</link>
    <guid>https://waltdhill.github.io/notes/implementations/karpathy-01-micrograd/07_breaking_up_tanh</guid>
    <description><![CDATA[  Prev: 06_backprop_backward_method_auto Next: 08_pytorch_backprop Related: Value object data structure, computation graph direction terminology # imports, reset_graph() to init nn, graphviz: trace() &amp; draw_dot() import math import numpy as np import matplotlib.pyplot as plt %matplotlib inline # ... ]]></description>
    <pubDate>Wed, 13 May 2026 00:11:57 GMT</pubDate>
  </item><item>
    <title>08_pytorch_backprop</title>
    <link>https://waltdhill.github.io/notes/implementations/karpathy-01-micrograd/08_pytorch_backprop</link>
    <guid>https://waltdhill.github.io/notes/implementations/karpathy-01-micrograd/08_pytorch_backprop</guid>
    <description><![CDATA[  Prev: 07_breaking_up_tanh Next: 09_pytorch_gradient_descent Related: Value object data structure, computation graph direction terminology # imports, `Value` class, reset_graph() to init nn, graphviz: trace() &amp; draw_dot() import math import numpy as np import matplotlib.pyplot as plt %matplotlib... ]]></description>
    <pubDate>Wed, 13 May 2026 00:11:57 GMT</pubDate>
  </item>
    </channel>
  </rss>