🇻🇳 Fine-Tuned PhoBERT for Vietnamese Sentiment Analysis | Automated Training Pipeline
Автор: Quân Nguyễn
Загружено: 2025-02-23
Просмотров: 279
Experience a fully automated fine-tuning pipeline for Vietnamese sentiment analysis using PhoBERT! From data preprocessing to model evaluation, everything is streamlined for efficiency.
✅ Automated Data Collection: Built a Scrapy pipeline to crawl sentiment data from Booking.com.
✅ End-to-end automation: Data cleaning, word segmentation & class balancing
✅ Dockerized for Efficiency: Optimized with NVIDIA CUDA, OpenJDK, VnCoreNLP and cached dependency installation for seamless deployment and accelerated NLP processing.
✅ Auto-visualization: Interactive interpreter with histograms for class imbalance & heatmap for confusion matrix.
✅ Optimized fine-tuning: [CLS] embeddings + fully connected classifier (LayerNorm, GELU activation, and Dropout)
✅ Real-time training updates via WebSockets
✅ Automated evaluation: Precision, recall, F1-score & confusion matrix
Watch the demo to see how my streamlined pipeline accelerates sentiment analysis! ⚡ Like & subscribe! 👍🔔
#AI #DeepLearning #PhoBERT #SentimentAnalysis
Доступные форматы для скачивания:
Скачать видео mp4
-
Информация по загрузке:
