Tweetovi
- Tweetovi, trenutna stranica.
- Tweetovi i odgovori
- Medijski sadržaj
Blokirali ste korisnika/cu @tomhosking
Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @tomhosking
-
Tom Hosking proslijedio/la je Tweet
We are sill accepting applications until 31 January for the 4-year PhD program in the Centre in Doctoral Training in NLP https://nlp-cdt.ac.uk/ and for the 3-year PhD program in ILCC http://web.inf.ed.ac.uk/ilcc/study-with-us/studentships/linguistics-speech-technology-cognitive-science …
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
In my defence,
@huggingface BERT expects a mask that is the inverse of@PyTorch built in transformers. A hangover from the TF conversion I guess?Prikaži ovu nitHvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Pro tip: try masking the padding tokens instead of the actual tokens
Prikaži ovu nitHvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
I got bored of curating
@SchedMD Slurm jobs by hand, so I built a super lightweight monitoring app: https://github.com/tomhosking/mckenzie … It's pretty basic for now, but PRs very welcome! (cc@tomsherborne@jamesowers@EdinburghNLP )pic.twitter.com/7bfyHHGN9V
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Live footage of me trying to integrate BERT into my modelhttps://twitter.com/garethgwynn/status/1206610973404073984 …
0:05Prikaži ovu nitHvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Tom Hosking proslijedio/la je Tweet
6th June 2020 is party night. Bring friends, lovers & glad rags for memories money can't buy. Team GB Tokyo Olympic trials, European 10,000m Cup & more. You are the event. https://blog.strava.com/galleries/night-of-the-10000m-pbs/ …
#Highgate10k
@tomhoskingpic.twitter.com/qL96iZWDV1
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Tom Hosking proslijedio/la je Tweet
Facebook AI is sharing MLQA, an extractive question answering (QA) evaluation benchmark aligned across Arabic, German, Hindi, Spanish, Vietnamese, and Simplified Chinese. It will help the AI community improve and extend QA in more languages. https://ai.facebook.com/blog/mlqa-evaluating-cross-lingual-extractive-question-answering …pic.twitter.com/lP65q5ceSH
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
I've been working with
@pranavrajpurkar's SQuAD for a while now, but only just noticed that this question appears 6 times in the training set: "I couldn't could up with another question. But i need to fill this space because I can't submit the hit. " Thanks, mystery turker
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Tom Hosking proslijedio/la je Tweet
QA Models should work in any language. So, we're releasing MLQA, a new cross-lingual QA evaluation dataset! Check out the paper and dataset: https://arxiv.org/pdf/1910.07475.pdf … https://github.com/facebookresearch/MLQA … With Barlas Oguz, Ruty Rinott,
@riedelcastro,@SchwenkHolger@facebookai@ucl_nlp
pic.twitter.com/IPkMjP8B4I
Prikaži ovu nitHvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Tom Hosking proslijedio/la je Tweet
Applications very welcome for the UKRI Centre for Doctoral Training in Natural Language Processing
@Edin_CDT_NLP at University of Edinburgh@InfAtEd! Details to apply: https://edin.ac/cdt-in-nlp Deadlines:
- non EU/UK applications: 29.11.19
- EU/UK applications: 31.1.20Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Tom Hosking proslijedio/la je Tweet
These are the numbers of actions that
@facebook take in their categories. We're talking millions and billions. You can see why@facebookai is very busy.#TTOconpic.twitter.com/DIVKr00vzu
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Question generation leaderboard update: UniLM from
@donglixp et al. makes it back onto the board after being evaluated on the standard split. Cool paper that shows the power of transfer learning! http://aqleaderboard.tomhosking.co.uk/squadHvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Tom Hosking proslijedio/la je Tweet
Our two-step, self-supervised approach to extractive question answering (QA) first trains a model to generate questions, then uses those questions to train a standard extractive QA model. https://ai.facebook.com/blog/research-in-brief-unsupervised-question-answering-by-cloze-translation/ …pic.twitter.com/7JQStT0915
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
...and just found an
@ACL2019_Italy paper that claims improvements over SotA, based on results from 2017. Recent scores are almost 50% (!!) higher. This is really problematic! Claiming SotA != achieving SotAPrikaži ovu nitHvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
cc
@riedelcastro - I think their use of a very small learning rate during fine-tuning also helped ;)Prikaži ovu nitHvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
This is a really nice paper - and gives an example where a really careful choice of reward when fine tuning a NLG model *can* give better outputhttps://twitter.com/byryuer/status/1173653035710418944 …
Prikaži ovu nitHvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Tom Hosking proslijedio/la je Tweet
I want to share a thread of discussion about QG task. I believe it's time to standardized the SQuAD QG task's dev-test setup. Otherwise, the claim of SOTA makes no sense. I suggest using the first QG paper's split from
@Xinya16 (https://github.com/xinyadu/nqg/tree/master/data …).https://twitter.com/byryuer/status/1173949834404851712 …Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Apologies to
@byryuer who has better attention to detail than I do - after updating the leaderboard to account for the many different splits in circulation (which are not comparable to each other), their paper currently sits at the top!https://twitter.com/tomhosking/status/1173927627230461953 …Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Two papers accepted to @emnlp2019 claim SotA in question generation, but do not outperform an approach from a EMNLP 2018 paper!
Prikaži ovu nitHvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
QPP from
@byryuer also looks strong, but they create their own test set (rather than using the standard split from@Xinya16 - https://github.com/tomhosking/squad-du-split …) so the results are sadly not directly comparablePrikaži ovu nitHvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
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.
1/2