🔄 In a Training Loop
Minh-Thien Nguyen
minhnguyent546
AI & ML interests
Research interests: Embeddings Models, Image-Text retrieval for Vietnamese, Optimal transport, RAG, and Image classification. Some pet projects on distributed training, training model on TPU, RAG for complex document multiple-choice QA.
Recent Activity
liked a dataset 1 day ago
allenai/WildChat-4.8M liked a dataset 1 day ago
nvidia/Nemotron-RL-Ultra-Training-Blends liked a model 1 day ago
tugot17/LFM2.5-1.2B-Instruct-DSpark-5LOrganizations
cotu-legal-retriever is a family of models optimized for Vietnamese legal retrieval tasks.
-
minhnguyent546/cotu-legal-retriever-Octen-Embedding-4B-stage1
Sentence Similarity • 4B • Updated • 57 -
minhnguyent546/cotu-legal-retriever-Qwen3-Embedding-4B-stage1
Sentence Similarity • 4B • Updated • 52 -
minhnguyent546/cotu-legal-retriever-Qwen3-Embedding-8B-stage1
Sentence Similarity • 8B • Updated • 52 -
minhnguyent546/KaLM-Embedding-Gemma3-12B-2511-tokenizer-for-transformers-v5
Updated
ViCLIP-OT: The First Foundation Vision-Language Model for Vietnamese Image–Text Retrieval with Optimal Transport
-
minhnguyent546/ViCLIP-OT
Feature Extraction • 0.2B • Updated • 47 • 3 -
minhnguyent546/ViSigLIP-OT
Feature Extraction • 0.2B • Updated • 12 • 1 -
minhnguyent546/ViCLIP-OT-checkpoints
Feature Extraction • Updated -
ViCLIP-OT: The First Foundation Vision-Language Model for Vietnamese Image-Text Retrieval with Optimal Transport
Paper • 2602.22678 • Published
cotu-legal-retriever is a family of models optimized for Vietnamese legal retrieval tasks.
-
minhnguyent546/cotu-legal-retriever-Octen-Embedding-4B-stage1
Sentence Similarity • 4B • Updated • 57 -
minhnguyent546/cotu-legal-retriever-Qwen3-Embedding-4B-stage1
Sentence Similarity • 4B • Updated • 52 -
minhnguyent546/cotu-legal-retriever-Qwen3-Embedding-8B-stage1
Sentence Similarity • 8B • Updated • 52 -
minhnguyent546/KaLM-Embedding-Gemma3-12B-2511-tokenizer-for-transformers-v5
Updated
Datasets for training embeddings models (and fine-tuning for retrieval tasks)
ViCLIP-OT: The First Foundation Vision-Language Model for Vietnamese Image–Text Retrieval with Optimal Transport
-
minhnguyent546/ViCLIP-OT
Feature Extraction • 0.2B • Updated • 47 • 3 -
minhnguyent546/ViSigLIP-OT
Feature Extraction • 0.2B • Updated • 12 • 1 -
minhnguyent546/ViCLIP-OT-checkpoints
Feature Extraction • Updated -
ViCLIP-OT: The First Foundation Vision-Language Model for Vietnamese Image-Text Retrieval with Optimal Transport
Paper • 2602.22678 • Published