Medijski sadržaj
- Tweetovi
- Tweetovi i odgovori
- Medijski sadržaj, trenutna stranica.
-
@googlecalendar New kind of spam being pushed directly to my schedule
how do I prevent that?pic.twitter.com/kbmTtOU1vb
-
After your comment about the accuracy, I dug in a bit. Turns out the zip compressor only uses the last 64 KB for its "statistics", which is about 82 samples. When using all training data, zip compression obtains 74% accuracy on MNIST
Not great, not terrible
pic.twitter.com/tU9ZqDYCAs
-
And the result page lacks the usual share button. This is not delightful
pic.twitter.com/jHbn42jfoH
Prikaži ovu nit -
I've implemented a short ablation study that shows that it is not better than the simplest possible approach (cosine similarity on averaged features for class examples): Google Colab at https://tinyurl.com/1905-12601-ablation …pic.twitter.com/3ikJgQsbsK
-
I would question the validity of this paper. In fact, the experiment section is meaningless. All of the performance on MNIST can be explained by using a cosine similarity on averaged class samples. See the Google Colab at https://tinyurl.com/1905-12601-ablation …pic.twitter.com/AiytETZSsL
-
@Barclays: All I want is see all my transactions since 1/1/2019 and what I get instead is a UI that tells me "that's all" but doesn't go further back than 13/02. And when you export the results to PDF you finally see the following... Seriously, who's doing your QA???pic.twitter.com/ZUUs4UPbXd
-
Impact of the Prevent policy, event at
@nwspk house tonight. TL;DR: really bad is an understatementpic.twitter.com/KOvZVByMlv
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.
