Aman Bhansali

Using an already existing ESRGAN model, degrading our HR images to generate training data for finetuning and further checking the MSE and PSNR for the generated and the original HR images. Link to the original link of Real ESRGAN Repo - https://github.com/xinntao/Real-ESRGAN

Using RNN, LSTM for Offline Handwriting Text Recognition. Analyzing an already given paper and mentioning its Extended Abstract, Critical Analysis and some suggested Innovations

Developing a sentiment analysis model and creating an API server for the same.

Using SVD and PCA to observe the top and most influential modes present in a shock wave generated from a super-sonic wind tunnel.

Using the google stock prices dataset and build a model to predict the future prices using previous instances.

Implementation of Decision tree classifier and regressor on Drugs and concrete dataset. Includes knowledge about Information gain and Gini index

Google Summer of Code resources.

Config files for my GitHub profile.

Official main repository for LFortran

✨ Standard library for JavaScript and Node.js. ✨

No description.

No description.

No description.

Applying supervised learning approaches on the dataset and finally predicting the labels in a multi class format. Dataset and more detailed problem present in Kaggle competition link: https://www.kaggle.com/t/04a239e37f69de0040810bc615b96063

Applying MLP model on IRIS and house prices advanced regression techniques dataset and computing average mean-square error for complete test data. Computing l1 (manhattan distance) and l2 (euclidean distance) loss and comparing the results. Plotting the loss curve, and a histogram of prediction errors with appropriate bin size.

Implemented PCA algorithm from scratch on MNIST Dataset. Visualizing the reconstructed images made and comparing them with the original image. Visualizing the residual images by subtracting the reconstructed image from the original image (for different values of Principal components). Finding the reconstruction error (pixel-wise root-mean-square) for each sample and plot them for a different number of principal components.