3 code implementations • 22 Aug 2024 • Jamba Team, Barak Lenz, Alan Arazi, Amir Bergman, Avshalom Manevich, Barak Peleg, Ben Aviram, Chen Almagor, Clara Fridman, Dan Padnos, Daniel Gissin, Daniel Jannai, Dor Muhlgay, Dor Zimberg, Edden M Gerber, Elad Dolev, Eran Krakovsky, Erez Safahi, Erez Schwartz, Gal Cohen, Gal Shachaf, Haim Rozenblum, Hofit Bata, Ido Blass, Inbal Magar, Itay Dalmedigos, Jhonathan Osin, Julie Fadlon, Maria Rozman, Matan Danos, Michael Gokhman, Mor Zusman, Naama Gidron, Nir Ratner, Noam Gat, Noam Rozen, Oded Fried, Ohad Leshno, Omer Antverg, Omri Abend, Opher Lieber, Or Dagan, Orit Cohavi, Raz Alon, Ro'i Belson, Roi Cohen, Rom Gilad, Roman Glozman, Shahar Lev, Shaked Meirom, Tal Delbari, Tal Ness, Tomer Asida, Tom Ben Gal, Tom Braude, Uriya Pumerantz, Yehoshua Cohen, Yonatan Belinkov, Yuval Globerson, Yuval Peleg Levy, Yoav Shoham
The model weights for both sizes are publicly available under the Jamba Open Model License and we release ExpertsInt8 as open source.
1 code implementation • 15 May 2023 • Reut Apel, Tom Braude, Amir Kantor, Eyal Kolman
We present MeeQA, a dataset for natural-language question answering over meeting transcripts.
no code implementations • 4 Aug 2021 • Tom Braude, Idan Schwartz, Alexander Schwing, Ariel Shamir
OIA models interactions between the sentence-corresponding image and important regions in other images of the sequence.