Skip to content

Commit

Permalink
Lateast
Browse files Browse the repository at this point in the history
Raman authored and Raman committed Jun 30, 2024
1 parent 7d6c75f commit b3e3051
Showing 1 changed file with 144 additions and 89 deletions.
233 changes: 144 additions & 89 deletions src/slides.svelte
View file
Original file line number Diff line number Diff line change
@@ -139,10 +139,9 @@
<li>1. Maschinelles Lernen</li>
<li>2. Rechenzentren und Energieverbrauch</li>
<li>3. Energieverbrauch in der KI</li>
<li>4. Studienergebnisse und deren Interpretation</li>
<li>5. Leitfrage</li>
<li>6. Sclussfolgerung</li>
<li>7. Referenzen</li>
<li>4. Stromverbrauch von GPT-Modellen</li>
<li>5. Schlussfolgerung</li>
<li>6. Referenzen</li>
</ul>
</mybody>
</Layout>
@@ -241,22 +240,36 @@
<Notes></Notes>
</Slide>

<Slide><Layout>
<titlebar>ML - overview</titlebar>
<mybody>
<div class="flex items-center justify-center flex-col">
<!-- svelte-ignore a11y-missing-attribute -->
<img
class="h-[70vh] align-middle mb-2"
src="Machine_Learning-Overview.jpg"
/>
<div class="text-xs mt-auto">Quelle: github:christophm</div>
</div>
</mybody>
</Layout></Slide
>

<Slide><Layout>
<titlebar >ML</titlebar>
<figure>
<video controls class="w-[70%] mx-auto my-auto p-2">
<source src="Deep-Learning-Illustration.mp4" type="video/mp4" />
</video>
kaggle.com
<div class="text-xs mt-auto">Quelle: kaggle.com</div>
</figure>
</Layout></Slide>

<Slide>
<Layout>
<titlebar> Anwendungen im Maschinellen Lernen </titlebar>
<div
class="grid mt-[5%] mx-[5%]"
class="grid mt-[5%] mx-[5f%]"
style="grid-template-columns: 1fr 1fr;"
>
<!-- svelte-ignore a11y-missing-attribute -->
@@ -516,12 +529,9 @@
<Notes></Notes></Slide
>

<script>
newChapter('SE & DI')
</script>
<Slide
><Layout>
<titlebar>Studienergebnisse und deren Interpretation</titlebar>

<Slide><Layout>
<titlebar></titlebar>
<mybody>
<div class="flex items-center justify-center flex-col">
<!-- svelte-ignore a11y-missing-attribute -->
@@ -541,106 +551,151 @@
>

<script>
newChapter('Leitfrage')
newChapter('SV GPT')
</script>
<Slide><Layout>
<titlebar>Stromverbrauch von GPT-Modellen</titlebar>
<mybody>
<div class="flex items-center justify-center flex-col">
<!-- svelte-ignore a11y-missing-attribute -->
<img
class="h-[50vh] align-middle mb-2"
src="GPT_Electricity_Consumption.jpg"
/>
</div>
</mybody>
</Layout>
<Notes>
These are some presenter notes that can help during the actual
presentation.<br />
They are displayed on the presenter's screen if "presenter mode" has been entered.
</Notes></Slide
>
<Slide><Layout>
<titlebar>Stromverbrauch von GPT-Modellen</titlebar>
<mybody>
<div class="flex items-center justify-center flex-col">
<!-- svelte-ignore a11y-missing-attribute -->
<img
class="h-[50vh] align-middle mb-2"
src="GPT_Electricity_Consumption_American-House.jpg"
/>
</div>
</mybody>
</Layout>
<Notes>
These are some presenter notes that can help during the actual
presentation.<br />
They are displayed on the presenter's screen if "presenter mode" has been entered.
</Notes></Slide
>

<script>
newChapter('Schlussfolgerung')
</script>
<Slide
><Layout>
<titlebar>Energieverbrauch in der KI</titlebar>
<titlebar>Schlussfolgerung</titlebar>
<mybody>
<div class="flex items-center justify-center">
<div class="flex-1">
<ul>
<li style="margin-bottom: 3px; font-size: 0.6em;">
Wie können wir den Energieverbrauch von maschinellem Lernen nachhaltig reduzieren, ohne die Leistungsfähigkeit zu beeinträchtigen?
</li>
</ul>
</div>
<div class="flex-1 flex items-center justify-center flex-col">
<!-- svelte-ignore a11y-missing-attribute -->
<img
class="h-[50vh] align-middle mb-2"
src="KI-Engergieverbrauch.jpg"
/>
<div class="text-xs mt-auto">Quelle: ingenieur.de</div>
</div>
<ul>
<li style="margin-bottom: 10px; font-size: 0.6em;">
• Die Entwicklung der GPT-Modelle zeigt eine signifikante <br/> Zunahme im Stromverbrauch mit jeder Generation
</li>
<li style="margin-bottom: 10px; font-size: 0.6em;">
• Hoher Energiebedarf pro Modell
</li>
<li style="margin-bottom: 10px; font-size: 0.6em;">
• Notwendigkeit für Effizienzsteigerung und Forschung
</li>
<li style="margin-bottom: 10px; font-size: 0.6em;">
• Herausforderungen für Skalierung und Nachhaltigkeit
</li>
</ul>
</mybody>
</Layout>
<Notes></Notes></Slide
>


<Slide
><Layout>
<titlebar>Durchschnittlicher Wachstumsrate GPT-3 - GPT-3.5</titlebar>
<mybody>
<div class="flex items-center justify-center flex-col">
<!-- svelte-ignore a11y-missing-attribute -->
<img
class="h-[50vh] align-middle mb-2"
src="GPT_Electricity_Consumption_GPT3_And_GPT3.5.jpg"
/>
</div>
</mybody>
</Layout>
<Notes></Notes></Slide
>
>

<Slide
><Layout>
<titlebar>Durchschnittlicher Wachstumsrate GPT-4 - GPT-5</titlebar>
<mybody>
<div class="flex items-center justify-center flex-col">
<!-- svelte-ignore a11y-missing-attribute -->
<img
class="h-[50vh] align-middle mb-2"
src="GPT_Electricity_Consumption_GPT4_And_GPT5.jpg"
/>
</div>
</mybody>
</Layout>
<Notes></Notes></Slide
>

<Slide
><Layout>
<titlebar>Durchschnittlicher Wachstumsrate GPT-3 - GPT-5</titlebar>
<mybody>
<div class="flex items-center justify-center flex-col">
<!-- svelte-ignore a11y-missing-attribute -->
<img
class="h-[50vh] align-middle mb-2"
src="GPT_Electricity_Consumption_All_Modules.jpg"
/>
</div>
</mybody>
</Layout>
<Notes></Notes></Slide
>

<script>
newChapter('Referenzen')
</script>
<Slide
><Layout>
<Slide>
<Layout>
<mybody>
<ul class="a">
<ul class="a" style="font-size: 25px;">
<li>Yu, et al., (2022)</li>
<li>
wsj.com <a
href="https://www.wsj.com/tech/ai/ai-energy-consumption-fc79d94f"
>, (2024)</a
>
</li>
<li>
Ingenieur.de<a
href="https://www.ingenieur.de/technik/fachbereiche/ittk/warum-rechenzentren-als-systemrelevant-eingestuft-werden-muessen/"
>, (2023)</a
>
</li>

<li>
bigml.com<a
href="https://blog.bigml.com/2021/11/22/introduction-to-image-processing/"
>, (2023)</a
>
</li>
<li>
researchgate.net<a
href="https://www.researchgate.net/figure/Image-pixel-generation-In-Digital-Image-Processing-the-digital-image-feature-extraction_fig1_283800375"
>, (2023)</a
>
</li>
<li>
pylessons.com<a href="https://pylessons.com/speech-recognition"
>, (2023)</a
>
</li>
<li>
kardome.com<a
href="https://www.kardome.com/blog-posts/difference-speech-and-voice-recognition"
>, (2023)</a
>
</li>
<li>
arxiv.org<a
href="https://arxiv.org/pdf/1906.02243"
>, Energy and Policy Considerations for Deep Learning in NLP</a
>
</li>

<li>
arxiv.org<a
href="https://arxiv.org/pdf/2303.01980"
>, Towards energy-efficient Deep Learning: An overview of energy-
efficient approaches along the Deep Learning Lifecycle</a
>
</li>
<li>wsj.com <a href="https://www.wsj.com/tech/ai/ai-energy-consumption-fc79d94f">, (2024)</a></li>
<li>Ingenieur.de <a href="https://www.ingenieur.de/technik/fachbereiche/ittk/warum-rechenzentren-als-systemrelevant-eingestuft-werden-muessen/">, (2023)</a></li>
<li>bigml.com <a href="https://blog.bigml.com/2021/11/22/introduction-to-image-processing/">, (2023)</a></li>
<li>researchgate.net <a href="https://www.researchgate.net/figure/Image-pixel-generation-In-Digital-Image-Processing-the-digital-image-feature-extraction_fig1_283800375">, (2023)</a></li>
<li>pylessons.com <a href="https://pylessons.com/speech-recognition">, (2023)</a></li>
<li>kardome.com <a href="https://www.kardome.com/blog-posts/difference-speech-and-voice-recognition">, (2023)</a></li>
<li>arxiv.org <a href="https://arxiv.org/pdf/1906.02243">, Energy and Policy Considerations for Deep Learning in NLP</a></li>
<li>arxiv.org <a href="https://arxiv.org/pdf/2303.01980">, Towards energy-efficient Deep Learning</a></li>
<li>medium.com <a href="https://www.springboard.com/blog/data-science/machine-learning-gpt-3-open-ai/">, Is the energy consumption of large language model is over-hyped</a></li>
<li>forbes.com <a href="https://www.forbes.com/sites/arielcohen/2024/05/23/ai-is-pushing-the-world-towards-an-energy-crisis/">, Is the energy consumption of large language model is over-hyped</a></li>
<li>theverge.com <a href="https://www.theverge.com/24066646/ai-electricity-energy-watts-generative-consumption">, Is the energy consumption of large language model is over-hyped</a></li>
</ul>
</mybody>
</Layout>
<Notes>
These are some presenter notes that can help during the actual
presentation.<br />
These are some presenter notes that can help during the actual presentation.<br />
They are displayed on the presenter's screen if "presenter mode" has been entered.
</Notes></Slide
>
</Notes>
</Slide>


<script>
newChapter('Sonstiges')
newChapter('Anhang')
</script>
<Slide
><Layout>

0 comments on commit b3e3051

Please sign in to comment.