|
gptkbp:instanceOf
|
large language model
|
|
gptkbp:architecture
|
gptkb:transformation
|
|
gptkbp:availableOn
|
gptkb:Hugging_Face
gptkb:AWS
gptkb:Microsoft_Azure
|
|
gptkbp:context
|
4096 tokens
|
|
gptkbp:developedBy
|
gptkb:Meta_AI
|
|
gptkbp:hasModel
|
decoder-only transformer
|
|
https://www.w3.org/2000/01/rdf-schema#label
|
Llama 2 34B
|
|
gptkbp:input
|
gptkb:text
|
|
gptkbp:intendedUse
|
research
commercial
|
|
gptkbp:language
|
English
|
|
gptkbp:license
|
custom commercial license
|
|
gptkbp:notableFor
|
high performance on benchmarks
open weights for research and commercial use
|
|
gptkbp:openSource
|
partially
|
|
gptkbp:output
|
gptkb:text
|
|
gptkbp:parameter
|
34 billion
|
|
gptkbp:partOf
|
Llama 2 family
|
|
gptkbp:predecessor
|
gptkb:Llama_1
|
|
gptkbp:releaseDate
|
July 2023
|
|
gptkbp:successor
|
gptkb:Llama_3
|
|
gptkbp:supports
|
code generation
reasoning tasks
instruction following
|
|
gptkbp:supportsChat
|
yes
|
|
gptkbp:supportsFineTuning
|
yes
|
|
gptkbp:tokenizer
|
Byte-Pair Encoding (BPE)
|
|
gptkbp:trainer
|
reinforcement learning from human feedback
supervised fine-tuning
publicly available data sources
|
|
gptkbp:bfsParent
|
gptkb:Llama_2_70B
|
|
gptkbp:bfsLayer
|
7
|