|
gptkbp:instance_of
|
gptkb:language
|
|
gptkbp:application
|
natural language processing
|
|
gptkbp:architecture
|
gptkb:Transformers
|
|
gptkbp:can_be_used_for
|
machine translation
dialog systems
question answering
text summarization
|
|
gptkbp:developed_by
|
gptkb:Microsoft_Research
improving AI communication
|
|
gptkbp:enables
|
transfer learning
|
|
gptkbp:has
|
multi-modal capabilities
high parameter count
|
|
gptkbp:has_achieved
|
state-of-the-art results
|
|
https://www.w3.org/2000/01/rdf-schema#label
|
Uni LMv3
|
|
gptkbp:improves
|
text generation
|
|
gptkbp:is_available_on
|
gptkb:Hugging_Face_Model_Hub
|
|
gptkbp:is_based_on
|
gptkb:GPT_architecture
BERT architecture
|
|
gptkbp:is_compatible_with
|
gptkb:Tensor_Flow
gptkb:Py_Torch
|
|
gptkbp:is_considered_as
|
a leading model
|
|
gptkbp:is_designed_for
|
scalability
|
|
gptkbp:is_documented_in
|
research papers
|
|
gptkbp:is_evaluated_by
|
gptkb:SQu_AD_benchmark
gptkb:MS_MARCO
gptkb:GLUE_benchmark
gptkb:Super_GLUE_benchmark
efficiency
robustness
generalization
|
|
gptkbp:is_influenced_by
|
previous Uni LM versions
|
|
gptkbp:is_known_for
|
flexibility in tasks
|
|
gptkbp:is_optimized_for
|
performance and efficiency
|
|
gptkbp:is_part_of
|
AI advancements
Uni LM family
|
|
gptkbp:is_related_to
|
deep learning
transformer models
NLP research
|
|
gptkbp:is_supported_by
|
community contributions
|
|
gptkbp:is_tested_for
|
real-world datasets
|
|
gptkbp:is_trained_in
|
large text corpora
|
|
gptkbp:is_used_for
|
entity recognition
sentiment analysis
text classification
text entailment
|
|
gptkbp:is_used_in
|
research and industry
|
|
gptkbp:performance
|
other language models
|
|
gptkbp:pretraining_objective
|
masked language modeling
|
|
gptkbp:provides
|
contextual embeddings
|
|
gptkbp:release_year
|
gptkb:2021
|
|
gptkbp:successor
|
gptkb:Uni_LMv2
|
|
gptkbp:supports
|
multi-task learning
zero-shot learning
|
|
gptkbp:tuning
|
sequence-to-sequence tasks
|
|
gptkbp:uses
|
self-attention mechanism
|
|
gptkbp:bfsParent
|
gptkb:Uni_LM
|
|
gptkbp:bfsLayer
|
6
|