Pinned Repositories
Best-Egg-NLP-Task
Build a Bidirectional GRU sentiment analysis model to predict the score column.
Compter-Vision-OpenCV-Image-Processing
Three small OpenCV image processing projects: Histogram match two images, Remove noise and edge detection, Mean shift segmentation etc.
Fast-Neural-Style-Transfer-Using-TF-Hub
This Notebook take the fast style transfer using TF-Hub part from the original source and extended to video style transfer. It is a lot of fun to play style transfer.
Huggingface-transformer-pipelines-for-15-NLP-tasks
This Notebook implemented pretrained transformer pipelines for 15 NLP tasks. Framework uses Tensorflow except table question answering which is only available for PyTorch.
Machine-Learning-Classifications-for-Heart-Disease-Prediction
Implemented Logistic Regression, SVM, KNN, Decision Tree, Random Forest, Gradient Boosting, XGBoost, ANN etc. ML algorithms for heart disease analysis and prediction.
Machine-Learning-Models-for-California-Housing-Price-Prediction
Implemented Linear Regression, Lasso, Ridge, Decision Tree, Random Forest and XGBoost ML algorithms for housing price prediction.
Machine-Learning-Projects
Natural Language Processing Projects
NLP-Multi-Class-Prediction-with-Imbalance-Dataset
This is a current real world NLP problem with 11 categories from a financial company. I provided multiple solutions to solve multi-class classifications with imbalance data.
Reinforcement-Learning-DQN-for-ATARI-s-Pong-Game---TensorFlow-2.0-Keras
Reinforcement Learning DQN for ATARI’s Pong Game---TensorFlow 2.0/Keras
YOLOV3-Detection-of-Truck-Vehicle-Types
You only look once (YOLO) is a state-of-the-art, real-time object detection system. YOLOV3 is extremely fast and accurate compared with other algorithms, such as R_CNN, RetinaNet etc. It uses Darknet-53 as the backbone network and uses three scale predictions. In this Notebook a YoloV3 model was trained using Darknet by transfer learning with GPU on Colab. Collect 8 classes samples (480 images): USPS Truck, UPS Truck, Ambulance, Fire Truck, FedEx Truck, Bus, Police Car and Other Vehicle. Collect negative samples (100) with objects that we do not want to detect, such as streets, pedestrian, buildings, traffic lights, etc. Train, Validation and Test sample preparation; Sample labeling. Speeding up training by using Fine Tune technique. Install Darknet and cuDNN, compile Darknet, configuration training file. Negative samples improved model performance. Inference 8 classes Truck/Vehicle images correctly Inference videos.
yxu1168's Repositories
yxu1168/NLP-Multi-Class-Prediction-with-Imbalance-Dataset
This is a current real world NLP problem with 11 categories from a financial company. I provided multiple solutions to solve multi-class classifications with imbalance data.
yxu1168/YOLOV3-Detection-of-Truck-Vehicle-Types
You only look once (YOLO) is a state-of-the-art, real-time object detection system. YOLOV3 is extremely fast and accurate compared with other algorithms, such as R_CNN, RetinaNet etc. It uses Darknet-53 as the backbone network and uses three scale predictions. In this Notebook a YoloV3 model was trained using Darknet by transfer learning with GPU on Colab. Collect 8 classes samples (480 images): USPS Truck, UPS Truck, Ambulance, Fire Truck, FedEx Truck, Bus, Police Car and Other Vehicle. Collect negative samples (100) with objects that we do not want to detect, such as streets, pedestrian, buildings, traffic lights, etc. Train, Validation and Test sample preparation; Sample labeling. Speeding up training by using Fine Tune technique. Install Darknet and cuDNN, compile Darknet, configuration training file. Negative samples improved model performance. Inference 8 classes Truck/Vehicle images correctly Inference videos.
yxu1168/Huggingface-transformer-pipelines-for-15-NLP-tasks
This Notebook implemented pretrained transformer pipelines for 15 NLP tasks. Framework uses Tensorflow except table question answering which is only available for PyTorch.
yxu1168/Machine-Learning-Models-for-California-Housing-Price-Prediction
Implemented Linear Regression, Lasso, Ridge, Decision Tree, Random Forest and XGBoost ML algorithms for housing price prediction.
yxu1168/Reinforcement-Learning-DQN-for-ATARI-s-Pong-Game---TensorFlow-2.0-Keras
Reinforcement Learning DQN for ATARI’s Pong Game---TensorFlow 2.0/Keras
yxu1168/Best-Egg-NLP-Task
Build a Bidirectional GRU sentiment analysis model to predict the score column.
yxu1168/Compter-Vision-OpenCV-Image-Processing
Three small OpenCV image processing projects: Histogram match two images, Remove noise and edge detection, Mean shift segmentation etc.
yxu1168/Fast-Neural-Style-Transfer-Using-TF-Hub
This Notebook take the fast style transfer using TF-Hub part from the original source and extended to video style transfer. It is a lot of fun to play style transfer.
yxu1168/Machine-Learning-Classifications-for-Heart-Disease-Prediction
Implemented Logistic Regression, SVM, KNN, Decision Tree, Random Forest, Gradient Boosting, XGBoost, ANN etc. ML algorithms for heart disease analysis and prediction.
yxu1168/Machine-Learning-Projects
Natural Language Processing Projects
yxu1168/Neural-Style-Transfer
Deep Learning & Art: Neural Style Transfer
yxu1168/NLP-Named-Entity-Recognition-Using-LSTM-and-Bert
Bidirectional LSTM and Bert are implemented for Named Entity Recognition
yxu1168/NLP-Sentiment-Analysis
CNN/GRU are implemented to Sentiment Analysis on Amazon Cell Phone Reviews
yxu1168/NLP-Sentiment-Anaysis-for-Spam-Detection
Implemented BERT, LSTM neural network models for spam detection
yxu1168/Transfer-Learning
VGG16 CNN for Classification and Detection; Mask RCNN for Object Detection and Segmentation