Project Introduction & Development Environment Preparation & NLP 6 Sections 90:00
- Introduction and Usage of Python FastAPI10:00
- Brief Introduction to Docker and Deployment30:00
- Text Classification Preprocessing and Feature Engineering05:00
- Explanation of Jieba Segmentation Principle15:00
- Stop Words and Word Filtering15:00
- One-Hot Encoding, TF-IDF, Word Embeddings15:00
- Practical Training on Sentence Embedding15:00
News Classification (Traditional Algorithms) 9 Sections 95:00
- Introduction to News Classification Project05:00
- Logistic Regression, SVM, CNN05:00
- Gradient Descent10:00
- Ensemble of Multiple Models Based on Logistic Regression05:00
- Evaluation Metrics: Precision, Recall, F1-score, AUC20:00
- News Data Preprocessing15:00
- PyTorch Function Usage10:00
- Handling Overfitting in Deep Learning15:00
- Engineering Code Standards10:00
News Classification (Bert) 6 Sections 95:00
- Self-Attention Mechanism20:00
- Transformers Source Code Interpretation10:00
- Training Text Classification Model Based on Bert30:00
- Model Performance Evaluation05:00
- Engineering Code Standards10:00
Question-Answer Search Generation System 9 Sections 110:00
- Text Embedding Search, Cosine Similarity Techniques10:00
- Introduction to Langchain Engineering30:00
- Overview of Large Language Models and Fine-Tuning10:00
- Testing for Effective and Rapid Development20:00
- Implementing Text Search and Question-Answer with Bloom + Langchain05:00
- Implementing Search Question-Answer System with Headline News Text Data10:00
Lane Line Detection Datasets Preprocess 5 Sections 90:00
- Introduction to Lane Line Recognition Project10:00
- Interpretation of Lane Data and Labels10:00
- Data Augmentation and Dataloader Creation30:00
- Lane Label Data Processing20:00
- Explanation of Matrix Positions for Four Lane Line Labels20:00
Lane Line Detection Model Network Structure 7 Sections 110:00
- Grid Setting Methods20:00
- Analysis of Algorithm Network Structure20:00
- Calculation Module for Loss Functions20:00
- Constraints of Lane Line Rule Loss Functions20:00
- Model Training10:00
- Detection Model Evaluation10:00
Interpretation of Lane Line Paper "Ultra Fast Structure-Aware" 8 Sections 90:00
- Interpretation of Lane Line Paper "Ultra Fast Structure-Aware"05:00
- Data Augmentation Methods15:00
- Explanation of Model Structure in the Paper15:00
Lane Line Detection Interface 5 Sections 95:00
- Inference with Trained Models20:00
- Writing Engineering Inference Interfaces20:00
- Testing Trained Models, Demo, Docker Interface Deployment05:00
Subway Passenger Flow Detection Datasets Preprocess 5 Sections 95:00
- Introduction to Subway Passenger Flow Project10:00
- Common Video Processing Methods (ffmpeg, cv2)50:00
- Extracting Frames from Subway Passenger Flow Video Dataset05:00
Yolo v1-v5 5 Sections 90:00
- Introduction to Common Detection Models20:00
- Introduction to YOLOv1-v5 Series and Differences20:00
- Interpretation of YOLOv5 Training Code 20:00
- Annotation and Preparation of Subway Personnel Detection Dataset20:00
- YOLOv5 Personnel Detection Model Training10:00
- Evaluation of YOLOv5 Personnel Detection Model10:00
Object Tracking Algorithms 5 Sections 90:00
- Introduction to Common Object Tracking Algorithms20:00
- Explanation of Bytetrack Algorithm20:00
- Incorporating Personnel Recognition for Bytetrack Tracking10:00
Cross Vectors Algorithm & Docker Interface Deployment 5 Sections 90:00
- Explanation of Cross Vectors Algorithm10:00
- Development of Engineering Code for YOLOv5 + Bytetrack + Embedding20:00
- Demo and Docker Interface Deployment10:00
- Introduction to Patent Writing Techniques20:00
