Domain-specific LLMs
Sefika Efeoglu PRO
Sefika
AI & ML interests
Natural language processing, relation extraction, entity extraction, semantic web, deep learning, NeuroSymbolic AI
Organizations
The models trained with Llama, and Mistral on FewRel dataset. The paper of the work has been submitted to possible journal
-
Sefika/KMmeans_CRE_llama_fewrel_4_task_memory_10_10
7B • Updated • 6 -
Sefika/KMmeans_CRE_llama_fewrel_4_task_memory_15_1
Text Generation • 7B • Updated • 3 -
Sefika/KMmeans_CRE_llama_fewrel_4_task_memory_15_2
Text Generation • 7B • Updated • 5 -
Sefika/KMmeans_CRE_llama_fewrel_4_task_memory_15_3
Text Generation • 7B • Updated • 6
Dynamic Approach FSCRE
Fine-tune Flan T5 Base with sentence(input) and relation(output)
This collection is prepared for my practical class at the TU Berlin.
Domain-specific LLMs
BART-base with fewrel
The models trained with Llama, and Mistral on FewRel dataset. The paper of the work has been submitted to possible journal
-
Sefika/KMmeans_CRE_llama_fewrel_4_task_memory_10_10
7B • Updated • 6 -
Sefika/KMmeans_CRE_llama_fewrel_4_task_memory_15_1
Text Generation • 7B • Updated • 3 -
Sefika/KMmeans_CRE_llama_fewrel_4_task_memory_15_2
Text Generation • 7B • Updated • 5 -
Sefika/KMmeans_CRE_llama_fewrel_4_task_memory_15_3
Text Generation • 7B • Updated • 6
no regularizer for loss
Dynamic Approach FSCRE
Fine-tune Flan T5 Base with sentence(input) and relation(output)
This collection is prepared for my practical class at the TU Berlin.
Relation Extraction Datasets
All data generated in ARDI project class. at TUB