This graph is PyTorch-related search queries, for instance. In general, all indicators are heavily correlated.pic.twitter.com/l9bUkRlArz
Show this thread
PyTorch has a lot of marketing firepower behind it, and as a result there's a common misconception that it has "momentum". Does it? I can't tell for sure, but the handful of traction indicators I monitor are showing that its user base has likely peaked around April-May 2018https://twitter.com/vincenzomanzoni/status/1047561148990730246 …
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
This Tweet is unavailable.
-
No, it would make no sense for Keras to support PyTorch (or Chainer for that matter). Also, "many people say" is not how we make design decisions
- Show replies
-
-
-
Here's one data-point: Frameworks mentioned at ICLR 2018-2019: TensorFlow 228->266, Keras: 42->56, Pytorch 87->252https://www.reddit.com/r/MachineLearning/comments/9kys38/r_frameworks_mentioned_iclr_20182019_tensorflow/ …
-
The PyTorch user base has grown significantly from last year. My point is just that it has been stalling for the past 6 months.
- Show replies
New conversation -
-
-
I have used both. PyTorch is great to learn. Tf all the way for work though.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
PyTorch has just reached the first release and http://Fast.ai chose it as the backend of its deep learning library. Probably we cannot already measure the trend, but there are weak signals that it can start. http://www.fast.ai/2018/10/02/fastai-ai/ …
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
I don't see the ads nor working for Fb. I was caffe and numpy user, stick to Keras for a year, even contributed to both and then migrated to PyTorch. Never regreted this.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Today, I have decided to stop my short gloomy experience with the eager mode. I ran away from endless issues and I found myself writing nonsense code: `tf.cond` `tf.while_loop` ` tf.Print` ` tf.gather` ` tf.scatter_update` `tf.what_the_****`.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
To my surprise, I discovered that the ****** `tf.while_loop` does not work any longer. At the end of the day, nothing works properly neither the graph mode nor the eager one. I am really disappointed. Meanwhile, Tensorflow developers brag about this ****.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.