bouracha Goto Github PK

autoencoders

catanai

deeplearning

dijkstrasshortestpath

elo

gatsby

gen_motion

generative_imputation

graphs_min_cut

hisrepitself_fork

image_processing

learning-trends

This repository is for the detection of learning trends in commodity and stock trading

maskrcnn_rodent

Here I use transfer learning on the MaskRCNN code (using the coco weights) to learn object detection and segmentation for rodents. This was for a specific project where I needed isolate single rodents from a video of a cage with multiple rodents. This was achieved by using the learned rodent (object) detection's bounding box as a single frame in a new cropped video for which bounding boxes were linked together if an aggregated metric of their confidence of detection and separation of centroids was below some hyperparameter-like tuned value. Importantly, we also end videos when two bounding boxes collide. The result of this is a collection of videos that follow only one rodent. The code for this is in extract_cideos.py.

mergesort_inversions

A merge sort algorithm which also counts the number of inversions to sort array.

mergesortinversions

A merge sort algorithm which also counts the number of inversions to sort array.

oodmotion

code for learning trajectory dependencies for human motion prediction

pdp-assignment

Squirrels

photoswitches

pix2pix_oxia

Synthesizing and manipulating 2048x1024 images with conditional GANs

quicksort

Quicksort algorithm with multiple choices of pivot (first last and median)

reinforcement_learning

stroke_outcome_dl_v_lr

strongyconnectedcomponents

supervised

supervised2

unsupervisedmt

Phrase-Based & Neural Unsupervised Machine Translation