Pinned Loading

1. Vietnamese-Entity-Recognition Vietnamese-Entity-Recognition Public

   Combinine PhoBERT, CRF, and ensemble methods to solve Vietnamese Named Entity Recognition, achieving an F1-score of approximately 90%+ on the VLSP2016 dataset.

   Jupyter Notebook 1 2

2. PersonaRAG PersonaRAG Public

   Production-grade RAG on the full OpenAI stack. Upload lecture notes, papers, and textbooks. Ask grounded questions. Get answers with inline citations — every time.

   Python 1

3. Vietnamese_Auto_Speech_Recognition_Benchamark Vietnamese_Auto_Speech_Recognition_Benchamark Public

   Fine-tuning and Performance Evaluation of QwenASR 0.6B, SqueezeFormer, Wav2Vec 2.0, and Whisper Family Models for Vietnamese Automatic Speech Recognition on the ViMD and VIVOS Corpora

   Jupyter Notebook

4. Vietnamese_Two-Speaker_Separator Vietnamese_Two-Speaker_Separator Public

   Leverages the VIVOS corpus to synthesis two-speaker scenarios and fine-tunes the Conv-TasNet architecture for Vietnamese speech separation.

   Python

5. DeepFake-Video-Detection DeepFake-Video-Detection Public

   Integrating ResNeXt-50 and LSTM to enhance robustness against GANs-based deepfake videos.

   Python 2

6. Speech-Emotion-Recognition Speech-Emotion-Recognition Public

   Building a custom CNN architecture that achieves 96% F1-Score on EmoDB corpus with speech emotion recognition.

   Jupyter Notebook 1