মিডিয়া
- টুইট
- টুইট এবং উত্তর
- মিডিয়া, বর্তমান পৃষ্ঠা।
-
This illustrates one aspect discussed, a clever way of exploring many possible layouts of photos to find one that looks good: "Finding the optimum photo grid (ie the one with the lowest combined set of bad rows) is as simple as calculating the shortest path through the graph."pic.twitter.com/mvRfBM77tK
এই থ্রেডটি দেখান -
Google Cloud TPUs now offer preemptible pricing at ~70% off the reserved instance pricing. This means, for example, that you can train a ResNet-50 model for ~$7.50 instead of $25, or a Transformer neural translation model for ~$13 instead of $41. See: https://cloudplatform.googleblog.com/2018/06/Cloud-TPU-now-offers-preemptible-pricing-and-global-availability.html …pic.twitter.com/2fBmY3P6Bf
এই থ্রেডটি দেখান -
If you want to learn more about why we're so excited about TPUs and how they make machine learning research and applications so much more efficient, I highly recommend watching Zak Stone's excellent talk about TPUs from today's Google I/O session on TPUs:https://youtu.be/vm67WcLzfvc?t=1114 …
-
We just posted new DAWNBench results for ImageNet classification training time and cost using Google Cloud TPUs+AmoebaNet (architecture learned via evolutionary search). You can train a model to 93% top-5 accuracy in <7.5 hours for <$50. Results: http://dawn.cs.stanford.edu/benchmark/ pic.twitter.com/QDSicSFOHS
এই থ্রেডটি দেখান -
I enjoyed chatting with
@lmoroney at the TensorFlow Developer Summit (although I'll now chide myself for not sitting up straight).https://youtu.be/AtFZv-IuVwQ -
I'll share pictures from my own experience from a few years ago, which was not quite as amazing as this video (ours was on top of the vehicle and nearby vehicles, but an amazing experience nonetheless)
@judywawira you might like the video in the parent tweetpic.twitter.com/HGx6K1MSAH
-
Stanford's DAWNBench is a new benchmark suite measuring a variety of deep learning training and inference tasks. Google Cloud TPU results are now up in the Imagenet category and are #1 in both ImageNet Training Time and ImageNet Training Cost. http://dawn.cs.stanford.edu/benchmark/#imagenet-train-time …pic.twitter.com/vsRmc1j8G9
এই থ্রেডটি দেখান -
1/
@GoogleCloud TPUs are now in Beta for people that want access to high speed accelerators for training machine learning models. Details in blog post at: https://cloudplatform.googleblog.com/2018/02/Cloud-TPU-machine-learning-accelerators-now-available-in-beta.html …pic.twitter.com/2EcVrI9zgK
এই থ্রেডটি দেখান -
Agreed. The underlying neural architecture search models are actually better than those by world class ml/computer vision researchers, though. Black dots vs. red dots in this picture (Figure 5 from https://arxiv.org/abs/1707.07012 )pic.twitter.com/Dr1itgKLl7
-
This
@distillpub article on Feature Visualization by Chris Olah (@ch402), Alexander Mordvintsev (@zzznah), and Ludwig Schubert (@ludwigschubert) is excellent for understanding what's going inside computer vision machine learning models. https://distill.pub/2017/feature-visualization/ …pic.twitter.com/r9Yzd4vBfw
-
Congrats to Google Brain team member Chris Shallue and his collaborator Andrew Vandenburg for their work on using machine learning to find lots of new exoplanets! Blog: https://blog.google/topics/machine-learning/hunting-planets-machine-learning/ … Paper: https://www.cfa.harvard.edu/~avanderb/kepler90i.pdf … AMA: https://www.reddit.com/r/science/comments/7jrexn/science_ama_series_were_planet_hunters_from_nasa/ …pic.twitter.com/10LcWgJ0z3
-
@judywawira enjoyed my cheetah photos from 2012 when I chatted with her at the@black_in_ai workshop dinner. Others might like them as well..pic.twitter.com/dsDpSMSgvq
লোড হতে বেশ কিছুক্ষণ সময় নিচ্ছে।
টুইটার তার ক্ষমতার বাইরে চলে গেছে বা কোনো সাময়িক সমস্যার সম্মুখীন হয়েছে আবার চেষ্টা করুন বা আরও তথ্যের জন্য টুইটারের স্থিতি দেখুন।