Best encoder-only models: multi-label (shared head) and binary-ensemble (9 independent classifiers), both with generative-augmentation mitigation.
Quim Motger
quim-motger
AI & ML interests
Natural Language Processing
Organizations
None yet
T-FREX fine-tuned LLMs for feature extraction based on Named-Entity Recognition.
-
quim-motger/t-frex-bert-base-uncased
Token Classification • 0.1B • Updated • 102 • 1 -
quim-motger/t-frex-xlnet-base-cased
Token Classification • 0.1B • Updated • 42 • 2 -
quim-motger/t-frex-roberta-large
Token Classification • 0.4B • Updated • 95 • 1 -
quim-motger/t-frex-bert-large-uncased
Token Classification • 0.3B • Updated • 89 • 1
Best encoder-only models: multi-label (shared head) and binary-ensemble (9 independent classifiers), both with generative-augmentation mitigation.
T-FREX fine-tuned LLMs for feature extraction based on Named-Entity Recognition.
-
quim-motger/t-frex-bert-base-uncased
Token Classification • 0.1B • Updated • 102 • 1 -
quim-motger/t-frex-xlnet-base-cased
Token Classification • 0.1B • Updated • 42 • 2 -
quim-motger/t-frex-roberta-large
Token Classification • 0.4B • Updated • 95 • 1 -
quim-motger/t-frex-bert-large-uncased
Token Classification • 0.3B • Updated • 89 • 1
models 14
quim-motger/emotion-bert-base-binary-ensemble
Text Classification • Updated • 1
quim-motger/emotion-roberta-large-multilabel-genai-bce
Text Classification • 0.4B • Updated • 365 • 1
quim-motger/t-frex-xlnet-large-cased
Token Classification • 0.4B • Updated • 28 • 1
quim-motger/t-frex-xlnet-base-cased
Token Classification • 0.1B • Updated • 42 • 2
quim-motger/t-frex-roberta-large
Token Classification • 0.4B • Updated • 95 • 1
quim-motger/t-frex-roberta-base
Token Classification • 0.1B • Updated • 96 • 1
quim-motger/t-frex-bert-large-uncased
Token Classification • 0.3B • Updated • 89 • 1
quim-motger/t-frex-bert-base-uncased
Token Classification • 0.1B • Updated • 102 • 1
quim-motger/reviewXLNet-large
Text Generation • 0.4B • Updated • 6 • 1
quim-motger/reviewXLNet-base
Text Generation • 0.1B • Updated • 7
datasets 0
None public yet