|
gptkbp:instanceOf
|
large language model
|
|
gptkbp:architecture
|
gptkb:transformation
|
|
gptkbp:author
|
gptkb:Michael_Lewis
gptkb:Luke_Zettlemoyer
gptkb:Mikel_Artetxe
gptkb:Moya_Chen
gptkb:Shruti_Bhosale
gptkb:Stephen_Roller
gptkb:Susan_Zhang
gptkb:Ves_Stoyanov
gptkb:Y-Lan_Boureau
gptkb:Edouard_Grave
gptkb:Naman_Goyal
|
|
gptkbp:citation
|
2022
|
|
gptkbp:compatibleWith
|
commercial use
legal advice
real-time applications
financial advice
medical advice
|
|
gptkbp:contrastsWith
|
gptkb:GPT-3
|
|
gptkbp:developedBy
|
gptkb:Meta_AI
|
|
gptkbp:fullName
|
gptkb:Open_Pre-trained_Transformer
|
|
gptkbp:github
|
gptkb:facebookresearch/OPT
|
|
https://www.w3.org/2000/01/rdf-schema#label
|
OPT
|
|
gptkbp:input
|
gptkb:text
|
|
gptkbp:intendedUse
|
research
|
|
gptkbp:language
|
English
|
|
gptkbp:license
|
CC-BY-NC 4.0
|
|
gptkbp:notableFor
|
open weights
detailed release notes
large-scale language modeling
reproducibility focus
transparency in release
|
|
gptkbp:notablePublication
|
OPT: Open Pre-trained Transformer Language Models
|
|
gptkbp:openSource
|
true
|
|
gptkbp:output
|
gptkb:text
|
|
gptkbp:parameter
|
up to 175B
|
|
gptkbp:purpose
|
research
|
|
gptkbp:releaseDate
|
2022
|
|
gptkbp:size
|
13B
1.3B
125M
175B
2.7B
30B
350M
6.7B
66B
|
|
gptkbp:tokenizer
|
gptkb:Byte-Pair_Encoding
|
|
gptkbp:trainer
|
publicly available datasets
|
|
gptkbp:trainingObjective
|
causal language modeling
|
|
gptkbp:bfsParent
|
gptkb:GPT-3
gptkb:Meta_AI
gptkb:Large_Language_Models
|
|
gptkbp:bfsLayer
|
5
|