SarthakTiwari

Pinned Repositories

• molecule-property-predictor-webapp

  A molecule property prediction flask app

  Language:Python00

• Attention-and-Edge-MPNN

  Modified Pytorch Lightning implementation of paper:-https://jcheminf.biomedcentral.com/track/pdf/10.1186/s13321-019-0407-y

  Language:Jupyter Notebook102

• Rationales-extaction-from-active-molecules

  rationales are molecule substructures that drive specific properties of interest

  Language:Python20

• Paper-Page-Scanner

  Page detector and selector for skewed document images

  Language:Python00

• News-Classification-using-distilbert

  Pretrained distilbert base uncased model fine tuned on news dataset

  Language:Jupyter Notebook00

• Flower-Image-classification

  A swing at writing article

  Language:Jupyter Notebook00

• Data-Science--Cheat-Sheet

  Cheat Sheets

  11

• FOOD-DETECTION-IN-IMAGE-USING-UNIMIB2016-DATASET

  Language:Jupyter Notebook11

• Dimensionality-reduction-of-images-to-improve-time-efficiency-of-models

  CNN works as a feature extractor, theses features can be transformed into lower dimension using either linear or non linear dimension reduction techniques before using these features for classification. This approach is demonstrated in action by converting the features of images in MINST dataset into lower dimension by using PCA and then feeding these new features into a classifier(fully connected layers with softmax activation in this case)

  Language:Jupyter Notebook00

• ENGAGING-NEWS-DETECTION

  People use Social Media to share news with each other. As large amounts of news is shared online each second, it becomes important to filter out news which users find engaging. In the data set given the headline of the news, the domain it is on, whether the news item is marked over 18, the time the news was submitted and the user who submitted the news, identified if the news is engaging or not.

  Language:Jupyter Notebook00

SarthakTiwari's Repositories