ML_DL_with_Python...

This repository will contain both code and additional links to code / reading material refrences for ML and DL with Python ML_DL_with_Python

Table of Contents

PyTorch

• 1 PyTorch-SimpleNet

• PyTorch_UNet

• 1.1-PyTorch-Layers-Conv2d-Architecture

• 1.2-PyTorch-transforms.Compose

• 1.3-PyTorch-freezingLayers-torch.no_grad()

• 1.4-PyTorch-Initial-ObjectDetection-YouTube

• 1.4a-PyTorch-Initial-ObjectDetection-Git

• 1.5-PyTorch--SimpleNet-Airplane-Classifier

• 1.5a-PyTorch_normalization_setting_from_Imagenet_Explained

• 1.5b-PyTorch_normalization_SOUMITH_CHINTALA_reply_for_transform.Normalize

• SOUMITH_CHINTALA

• pytorch_CycleGAN_and_pix2pix_Test

Absolutely excellent code and documentation , could get the Test - Inference running within an hour on own local system Own Video initial experiment -- YouTube Link - pytorch_CycleGAN_and_pix2pix_Test

• [2] Diffusers - Text to Image

• 2.1-Create-own-pipeline

• [3] detectron2 from Facebook AI Research- for Object Detection

• 3.1-detectron2-FAIR-Panoptic-Segmentation-YouTube

• 3.1a-detectron2-FAIR-Panoptic-Segmentation-GitRepo

• (https://github.com/RohitDhankar/detectron2_coco_panoptic_seg/blob/main/output_dir/__11_04_2022_23_50__/_dir_img_gt_bbox_/_11_04_2022_23_50_/000000000139_0.50_.png)

• 3.2-detectron2-faster_rcnn_R_50_FPN_3x

• 3.3-detectron2-config_explained -self.get_config.merge_from_file(model_zoo.get_config_file("COCO-Detection/faster_rcnn_R_50_FPN_3x.yaml"))

• 3.3a-detectron2-LAUNCH_explained LAUNCH-detectron2

• 3.3b-detectron2-model_zoo_explained

• 3.3c-detectron2 , COCO API explained , pycocotools , training data set , validation set , categories of the COCO Dataset

• [pyCOCO-get-annotation-ID's]

• annotation_ids = coco_obj.getAnnIds(imgIds=ls_image_ids_bikes[iter_image], catIds=[2])(https://github.com/RohitDhankar/Obj_Detect_Detectron2/blob/950503a6f64783d8d54e50af49e3e9888a14b428/src/det2_1.py#L101C5-L101C91)

• 3.4-detectron2 from FAIR

• 3.5-kaggle_sartorius-cell-instance-segmentation-GIT-Repo 3.2-kaggle_sartorius-YOuTube-Explainer

• 4-OpenPCDet_Initial_Demo_MAYAVI_KITTI_LIDAR

• 4.1-OpenPCDet_Git-Experiments_KITTI_LIDAR

[Neural_Networks-Basics]

• [1] Neural Networks - mostly CNN for Image tasks

• CNN-Layers_architecture_Explained-GITRepo

• 1-Book-1-Practical-Deep-Learning(Anirudh-Koul)

• 1.1-Book-1-->Image_Similarity_Search_18March22-YouTube

• 1.1a-Book-1-->Image_Similarity_Search_18March22-GitRepo

• 1.1b-Book-1-->Image-Feature-Extraction-GitRepo

• 1.1c-Book-1-->Image-Feature-Extraction_for_ACTIVE_Learning_LIGHTLY-YouTube

• BOOK-2--Programming PyTorch for DeepLearning by Ian Pointer (O’Reilly). Copyright 2019 Ian Pointer, 978-1-492-04535-9

• 2-Book-2-->PyTorch-SimpleNet-AirplaneClassifier

-TRAIN----torch.cuda.memory_allocated---> 0.02 GB
Epoch:10,TRAIN_Loss:0.00,VAL_Loss:0.00, Accuracy = 1.00
---labels[prediction]----
 not_plane
---labels[prediction]----
 not_plane

Important Papers and Readings

Deep Residual Learning for Image Recognition

Deep Residual Learning for Image Recognition

AUTHORS -- Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun
Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. 

PAPER CITED SOURCE -- 
Subjects: 	Computer Vision and Pattern Recognition (cs.CV)
Cite as: 	arXiv:1512.03385 [cs.CV]
  	(or arXiv:1512.03385v1 [cs.CV] for this version)
  	
https://doi.org/10.48550/arXiv.1512.03385

• Selected_ARXIV_Papers

Active Learning

• [1] Active Learning - mostly Lightly
• testing Lightly
• 1.1-Lightly Active Learning -YouTube
• 1.2-Lightly Active Learning-SimSiam -YouTube
• 1.3-Lightly Active Learning- -YouTube
• 1.4-FeatureExtraction-Resnet50

Object Detection -- Yolov4

• [1] Yolov4
• 1.1-Initial-Experiments-Yolov4
• 1.2-Yolov4-Git(DEPRECATED-theAIGuysCode)

GEO-GIS-DeepLearning-for-Geography

• [1] raster-vision
• raster-vision
• 1.1-raster-vision
• 1.2-QGIS-SpaceNet
• 1.2a-QGIS-SpaceNet
• 1.3-QGIS-LAStools_LIDAR_1
• 1.3-QGIS-LAStools_LIDAR

OpenCV - OpenCV Projects

• [1] OpenCV Projects
• 1.1-MediaPipe_OpenCV_Hand_Pose_Tracking
• 1.2-Credit Card Digit Detection
• 1.2a-Credit Card Digit Detection
• 1.3-OpenCV_mask_blue-cv2.bitwise_and)
• 1.4-OpenCV_SIFT_ORB_KeyPoints_Detection)
• OCR_OpenCv_PyTesseract
• Old_Code__needs_refactoring__CIFAR_kNN_Stanford
• Old_Code__needs_refactoring__OCR_Keras_TensorFlow

Generic Core Python

• [1] Python-3 basics , data transforms etc.

• PySpark

• pandas

• [2] Download and Preprocess various kinds of Data

• data_aggregations

• [3] Visualizations of Data for EDA and others

• 3.1-EDA-Exploratory-Data-Analysis

• 3.2-Bokeh-plots-for-Tukeys-Five-Point-Summary

• 3.3-Bokeh-Box-plots-for-Tukeys-Five-Point-Summary

• [3.4-statsmodels] --> TODO (https://github.com/RohitDhankar/time_series/blob/main/ts_1.ipynb)

• [3.5-matplotlib-TimeSeries-Plots] --> TODO (https://github.com/RohitDhankar/time_series/blob/main/ts_1.ipynb)

• 3.6-Levenshtein-distance

• DataTables-DigitalCognition

• Code-walkthrough-datatables.net-datatables.js

• PCA-->INTRO-->cPCA-ContrastivePrincipleComponentsAnalysis

Not Core Machine Learning [Full Stack Machine Learning]

• [1] Software Development - Data Focused App Development
• 1.1-ChatGPT-FlaskApp
• 1.1a-ChatGPT-FlaskApp-VariationAPI
• 1.2-Django-jQuery-Bokeh-Exploratory-Data-Analysis
• 1.2-uWSGI
• Django-wsgi - https://github.com/RohitDhankar/digitalCognition/blob/master/dc_dash_proj/wsgi.py
• 1.3-NGINX
• 1.3-Nginx_uWsgi_Django_Registration_Redux
• [2] Tornado - async framework
• [2.1-Tornado] (https://github.com/RohitDhankar/Machine-Learning-with-Python_ML_Py/tree/master/src/dev_tornado)

Classic Legacy NLP -- NLTK , Spacy , BERT

• FaceBook-WebHooks-ChatBot
• spacy_init
• BERT--TODO
• BERT-Transformers-TODO
• Back_Propagation_ChainRule
• ResNet-ShortCutConnection-Prevents-VanishingGradients

LLM Usage -- LangChain , ScikitLLM

• [1] Using Large Language Models - LangChain , ScikitLLM , Gradio
• [TODO] -ScikitLLM- https://github.com/iryna-kondr/scikit-llm
• [TODO] - Gradio web UI for Large Language Models. - https://github.com/oobabooga/text-generation-webui
• [TODO]

[TODO--List]

• todo_list_readme.md
• Education_Certificates