-
newmindai/gliner-multi-v2.1-nemotron-pii-tr
Token Classification • Updated • 18 -
newmindai/gliner2.5-mursit-kvkk-tr-v1
Token Classification • 0.2B • Updated • 15 -
newmindai/gliner2.5-kvkk-tr-v1
Token Classification • 0.3B • Updated • 18 • 1 -
newmindai/gliner2.5-kvkk-tr-v2
Token Classification • 0.3B • Updated • 1.75k
AI & ML interests
Where data finds its mind
Recent Activity
View all activity
Papers
Mecellem Models: Turkish Models Trained from Scratch and Continually Pre-trained for the Legal Domain
TurkEmbed: Turkish Embedding Model on NLI & STS Tasks
A Large-Scale Benchmark for Legal Text Embeddings
FP8 Rowwise and BF16 tensorwise models with optimized recipes for large-scale training efficiency and convergence stability.
-
newmindai/Llama-3.1-8B-Instruct-w16a16-tw
Text Generation • 8B • Updated • 32 -
newmindai/Llama-3.1-8B-Instruct-w16a8-rw
Text Generation • 8B • Updated • 39 -
newmindai/Llama-3.1-8B-Instruct_w16a8_rw_with_gw_hp
Text Generation • 8B • Updated • 30 -
newmindai/Llama-3.1-8B-Instruct-w16a8-mxtw
Text Generation • 8B • Updated • 30
-
newmindai/ModernBERT-tr-uncased-stsb-HD
Token Classification • 0.1B • Updated • 514 • 1 -
newmindai/TurkEmbed4STS-HD
Token Classification • 0.3B • Updated • 29 • 1 -
newmindai/lettucedect-210m-eurobert-tr-v1
Token Classification • 0.2B • Updated • 20 • 1 -
Turk-LettuceDetect: A Hallucination Detection Models for Turkish RAG Applications
Paper • 2509.17671 • Published • 12
Project: Turkish Embeddings from Scratch and CPT Decoders
Infrastructure: MareNostrum 5 (BSC)
-
newmindai/Mursit-Base-TR-Retrieval
Sentence Similarity • 0.2B • Updated • 1.98k • • 4 -
newmindai/Mursit-Large-TR-Retrieval
Sentence Similarity • 0.4B • Updated • 2.93k • • 8 -
newmindai/Mursit-Base
Feature Extraction • 0.2B • Updated • 147 • 3 -
newmindai/Mursit-Large
Feature Extraction • 0.4B • Updated • 145 • 4
TurkColBERT: A Benchmark of Dense and Late-Interaction Models for Turkish Information Retrieval
Funding: EuroHPC JU Benchmark Access Grant No. EHPC-BEN-2024B11-003
Infrastructure: IT4Innovations National Supercomputing Center (Karolina)
-
newmindai/gliner-multi-v2.1-nemotron-pii-tr
Token Classification • Updated • 18 -
newmindai/gliner2.5-mursit-kvkk-tr-v1
Token Classification • 0.2B • Updated • 15 -
newmindai/gliner2.5-kvkk-tr-v1
Token Classification • 0.3B • Updated • 18 • 1 -
newmindai/gliner2.5-kvkk-tr-v2
Token Classification • 0.3B • Updated • 1.75k
Project: Turkish Embeddings from Scratch and CPT Decoders
Infrastructure: MareNostrum 5 (BSC)
-
newmindai/Mursit-Base-TR-Retrieval
Sentence Similarity • 0.2B • Updated • 1.98k • • 4 -
newmindai/Mursit-Large-TR-Retrieval
Sentence Similarity • 0.4B • Updated • 2.93k • • 8 -
newmindai/Mursit-Base
Feature Extraction • 0.2B • Updated • 147 • 3 -
newmindai/Mursit-Large
Feature Extraction • 0.4B • Updated • 145 • 4
A Large-Scale Benchmark for Legal Text Embeddings
TurkColBERT: A Benchmark of Dense and Late-Interaction Models for Turkish Information Retrieval
FP8 Rowwise and BF16 tensorwise models with optimized recipes for large-scale training efficiency and convergence stability.
-
newmindai/Llama-3.1-8B-Instruct-w16a16-tw
Text Generation • 8B • Updated • 32 -
newmindai/Llama-3.1-8B-Instruct-w16a8-rw
Text Generation • 8B • Updated • 39 -
newmindai/Llama-3.1-8B-Instruct_w16a8_rw_with_gw_hp
Text Generation • 8B • Updated • 30 -
newmindai/Llama-3.1-8B-Instruct-w16a8-mxtw
Text Generation • 8B • Updated • 30
Funding: EuroHPC JU Benchmark Access Grant No. EHPC-BEN-2024B11-003
Infrastructure: IT4Innovations National Supercomputing Center (Karolina)
-
newmindai/ModernBERT-tr-uncased-stsb-HD
Token Classification • 0.1B • Updated • 514 • 1 -
newmindai/TurkEmbed4STS-HD
Token Classification • 0.3B • Updated • 29 • 1 -
newmindai/lettucedect-210m-eurobert-tr-v1
Token Classification • 0.2B • Updated • 20 • 1 -
Turk-LettuceDetect: A Hallucination Detection Models for Turkish RAG Applications
Paper • 2509.17671 • Published • 12