Skip to content

Commit 8d92222

Browse files
fix spacing issues, invisible hr
1 parent 7149f93 commit 8d92222

File tree

1 file changed

+14
-13
lines changed

1 file changed

+14
-13
lines changed

index.html

+14-13
Original file line numberDiff line numberDiff line change
@@ -44,7 +44,7 @@
4444
}
4545
#main_window {
4646
width: 80%;
47-
min-width: 280px;
47+
min-width: 350px;
4848
margin: 0 auto;
4949
text-align: center;
5050
}
@@ -183,7 +183,7 @@ <h1 class="faded title" style="margin-top:-3%;">
183183
<button onclick="window.open('https://arxiv.org/abs/2002.04013');"
184184
class="arxiv_button">Read the paper</button>
185185
</div>
186-
<a href="https://twitter.com/intent/tweet?url=learning-at-home.github.io&button_hashtag=learningathome"
186+
<a href="https://twitter.com/intent/tweet?button_hashtag=learningathome&hashtags=joinhivemind&text=Join%20the%20deep%20learning%20hivemind!%0Alearning-at-home.github.io"
187187
class="twitter-hashtag-button" data-show-count="true" data-size="large">Tweet #learningathome</a>
188188
<script async src="https://platform.twitter.com/widgets.js" charset="utf-8"></script>
189189

@@ -204,7 +204,8 @@ <h1 class="faded title" style="margin-top:-3%;">
204204
Pretrained transformers <a href=https://w4ngatang.github.io/static/papers/superglue.pdf target="_blank" rel="noopener noreferrer">dominate</a> most NLP tasks;</li>
205205
<li style="margin-top: 12px;">
206206
Bigger CNNs <a href="https://arxiv.org/abs/1912.11370" target="_blank" rel="noopener noreferrer">perform better</a> at computer vision;</li>
207-
<li style="margin-top: 12px;">GPT-3 has <a href="https://arxiv.org/abs/2005.14165" target="_blank" rel="noopener noreferrer">175B</a> parameters and <a target="_blank" rel="noopener noreferrer" href="https://arxiv.org/abs/2006.16668">still counting!</a></li>
207+
<li style="margin-top: 12px;">
208+
GPT-3 has <a href="https://arxiv.org/abs/2005.14165" target="_blank" rel="noopener noreferrer">175B</a> parameters and <a target="_blank" rel="noopener noreferrer" href="https://arxiv.org/abs/2006.16668">still counting!</a></li>
208209
</ul>
209210

210211
With transfer learning, these large models can harness nearly unlimited raw data to improve performance on both <a href=https://paperswithcode.com/task/language-modelling target="_blank" rel="noopener noreferrer">academic benchmarks</a> and solve <a href=https://medium.com/towards-artificial-intelligence/crazy-gpt-3-use-cases-232c22142044 target="_blank" rel="noopener noreferrer">new unexpected</a> tasks.
@@ -221,23 +222,23 @@ <h1 class="faded title" style="margin-top:-3%;">
221222
<div class style="width: 100%; max-width: 440px; vertical-align: middle; display: inline-block;">
222223
<div class="carousel"
223224
data-flickity='{ "draggable": ">1", "pageDots": false}'>
224-
<div class="carousel-cell" style="padding-left:20px; padding-right:10px;">
225-
<img width=360px height=250px src="kaplan1.png"></div>
225+
<div class="carousel-cell" style="padding-left:10px; padding-right:10px;">
226+
<img width=340px height=250px src="kaplan1.png"></div>
226227
<div class="carousel-cell" style="padding-left:10px; padding-right:10px;">
227228
<img width=320px height=250px src="gpipe1.png"></div>
228-
<div class="carousel-cell" style="padding-left:10px; padding-right:20px;">
229+
<div class="carousel-cell" style="padding-left:10px; padding-right:10px;">
229230
<img width=320px height=250px src="gpipe2.png"></div>
230231
</div>
231232
</div>
232233

233234
<p style="width: 100%; max-width: 900px; margin-top:16px; text-align: left">
234-
That said, training large neural networks isn't cheap. The hardware used for the <a href="https://arxiv.org/abs/1909.08053" target="_blank" rel="noopener noreferrer">previous largest</a> language model costs over $25 million. A single training run for GPT-3 will set you back <a href="https://lambdalabs.com/blog/demystifying-gpt-3/" target="_blank" rel="noopener noreferrer">at least $4.6M</a> in cloud GPUs. As a result, researchers can't contribute to state-of-the-art deep learning models and practitioners can't build applications without <a href=https://blogs.microsoft.com/ai/openai-azure-supercomputer target="_blank" rel="noopener noreferrer">being supported</a> by a megacorporation. If&nbsp;we&nbsp;want AI to benefit everyone, it can't be privately owned.
235+
That said, training large neural networks isn't cheap. The hardware used for the <a href="https://arxiv.org/abs/1909.08053" target="_blank" rel="noopener noreferrer">previous largest</a> language model costs over $25 million. A single training run for GPT-3 will set you back <a href="https://lambdalabs.com/blog/demystifying-gpt-3/" target="_blank" rel="noopener noreferrer">at least $4.6M</a> in cloud GPUs. As a result, researchers can't contribute to state-of-the-art deep learning models and practitioners can't build applications without <a href=https://blogs.microsoft.com/ai/openai-azure-supercomputer target="_blank" rel="noopener noreferrer">being supported</a> by a megacorporation. If&nbsp;we&nbsp;want the future of AI to be open, it can't be privately owned.
235236
</p>
236237

237238
</div>
238239
<br>
239240

240-
<div class="faded" style="margin-top:35px; width: 100%; max-width: 900px; display: inline-block; text-align: left;">
241+
<div class="faded" style="margin-top:32px; width: 100%; max-width: 900px; display: inline-block; text-align: left;">
241242
<p class="faded title" style="font-size:28px;">
242243
What is hivemind?</p>
243244
<br>
@@ -248,7 +249,7 @@ <h1 class="faded title" style="margin-top:-3%;">
248249
<img src="https://w7.pngwing.com/pngs/634/575/png-transparent-youtube-logo-youtube-computer-icons-font-awesome-logo-video-icon-angle-rectangle-triangle.png" style="width:100%; max-width:880px; min-width:320px; margin-top:12px; float:center;">
249250
<br>
250251
<div style="text-align: center">
251-
<div style="width: 100%; max-width: 500px; vertical-align: top; display: inline-block; padding-right: 5px; text-align: left; float: left;">
252+
<div style="width: 100%; max-width: 500px; vertical-align: top; display: inline-block; padding-right: 5px; text-align: left;">
252253
<span class="faded text" style="margin-top:16px">
253254
Here's how a hivemind experiment is organized:
254255
<ul style="text-align: left; list-style-position: inside; margin-top: 16px; margin-left: -32px;">
@@ -287,7 +288,7 @@ <h1 class="faded title" style="margin-top:-3%;">
287288
<li style="margin-top: 12px;">
288289
Training neural networks on multiple servers with varying compute, bandwidth and reliability; </li>
289290
<li style="margin-top: 12px;">
290-
<b>[TBA]</b> Training outrageously large neural networks across multiple compute clusters.</li>
291+
<b>[TBA]</b> Training huge neural networks across multiple compute clusters.</li>
291292
</ul>
292293
<br>
293294

@@ -314,19 +315,19 @@ <h1 class="faded title" style="margin-top:-3%;">
314315
</div>
315316

316317

317-
<div style="margin: 0 auto; margin-top: 16px;">
318+
<div style="margin: 0 auto; margin-top: 0px;">
318319
<a class="github-button" href="https://github.com/learning-at-home/hivemind" data-size="large" data-show-count="true" aria-label="Star learning-at-home/hivemind on GitHub">Browse the code</a>
319320
<div style="overflow: hidden; white-space: nowrap; margin: 0 auto; display: inline-block;">
320321
<button onclick="window.open('https://arxiv.org/abs/2002.04013');"
321322
class="arxiv_button">Read the paper</button>
322323
</div>
323-
<a href="https://twitter.com/intent/tweet?url=learning-at-home.github.io&button_hashtag=learningathome"
324+
<a href="https://twitter.com/intent/tweet?button_hashtag=learningathome&hashtags=joinhivemind&text=Join%20the%20deep%20learning%20hivemind!%0Alearning-at-home.github.io"
324325
class="twitter-hashtag-button" data-show-count="true" data-size="large">Tweet #learningathome</a>
325326
<script async src="https://platform.twitter.com/widgets.js" charset="utf-8"></script>
326327

327328
<script async defer src="https://buttons.github.io/buttons.js"></script>
328329
</div>
329-
<hr style="margin-top:32px; margin-bottom: 64px; width:50%">
330+
<hr style="margin-bottom: 64px; width:0%; border: 0 solid white">
330331
</div>
331332
</div>
332333
</div>

0 commit comments

Comments
 (0)