editing book ...

owlbarn · Apr 12, 2022 · e3f7eda · e3f7eda
1 parent c660ec3
commit e3f7eda
Show file tree

Hide file tree

Showing 21 changed files with 1 addition and 198 deletions.
diff --git a/book/case-finance/README.md b/book/case-finance/README.md
diff --git a/book/case-finance/dune b/book/case-finance/dune
diff --git a/book/case-finance/dune.inc b/book/case-finance/dune.inc
diff --git a/book/case-finance/prelude.ml b/book/case-finance/prelude.ml
diff --git a/book/case-gpu/README.md b/book/case-gpu/README.md
diff --git a/book/case-gpu/dune b/book/case-gpu/dune
diff --git a/book/case-gpu/dune.inc b/book/case-gpu/dune.inc
diff --git a/book/case-gpu/prelude.ml b/book/case-gpu/prelude.ml
diff --git a/book/community/README.md b/book/community/README.md
diff --git a/book/community/dune b/book/community/dune
diff --git a/book/community/dune.inc b/book/community/dune.inc
diff --git a/book/community/prelude.ml b/book/community/prelude.ml
diff --git a/book/core-opt/README.md b/book/core-opt/README.md
@@ -650,6 +650,4 @@ As can be seen, my repeat operation achieves about half of that in NumPy with re
 The outer repeat operation in NumPy is implemented using the single axis version, and thus is less efficient.
 The repeat operation in Julia is much slower. One reason is that `repeat` is not a computation-intensive operation, so the optimisation techniques such as static compilation and vectorisation are of less importance than algorithm design.
 
-## Summary
-
 ## References
diff --git a/book/distributed/README.md b/book/distributed/README.md
@@ -507,8 +507,6 @@ let push kv_pairs =
     ) kv_pairs
 ```
 
-TODO:  Explain the code
-
 ![MNIST training using Actor](images/distributed/exp_accuracy_01.png){#fig:distributed:exp_accuracy_01}
 
 In [@fig:distributed:exp_accuracy_01], we conduct the real-world experiments  using 6 worker nodes with the Parameter Server framework we have implemented.

diff --git a/book/ppl/README.md b/book/ppl/README.md
diff --git a/book/ppl/dune b/book/ppl/dune
diff --git a/book/ppl/dune.inc b/book/ppl/dune.inc
diff --git a/book/ppl/prelude.ml b/book/ppl/prelude.ml
diff --git a/docs/core-opt.html b/docs/core-opt.html
@@ -333,9 +333,6 @@ <h3>Repeat Operations</h3>
 <p>The evaluation results compared with NumPy and Julia are shown in <span data-cites="fig:core-opt:opeval_repeat" class="citation">[@fig:core-opt:opeval_repeat]</span>. We also measure the peak memory usage in<span data-cites="fig:core-opt:opeval_tp_repeat_mem_00" class="citation">[@fig:core-opt:opeval_tp_repeat_mem_00]</span>. As can be seen, my repeat operation achieves about half of that in NumPy with regard to both execution speed and memory usage. The outer repeat operation in NumPy is implemented using the single axis version, and thus is less efficient. The repeat operation in Julia is much slower. One reason is that <code>repeat</code> is not a computation-intensive operation, so the optimisation techniques such as static compilation and vectorisation are of less importance than algorithm design.</p>
 </section>
 </section>
-<section class="level2" id="summary">
-<h2>Summary</h2>
-</section>
 <section class="level2" id="references">
 <h2>References</h2>
 </section>

diff --git a/docs/distributed.html b/docs/distributed.html
@@ -386,7 +386,6 @@ <h3>Accuracy</h3>
       (k, value)
     ) kv_pairs</code></pre>
 </div>
-<p>TODO: Explain the code</p>
 <figure>
 <img id="fig:distributed:exp_accuracy_01" alt="MNIST training using Actor" src="images/distributed/exp_accuracy_01.png"><figcaption>MNIST training using Actor</figcaption>
 </figure>

diff --git a/docs/toc.html b/docs/toc.html