Very minimal pytorch boilerplate with wandb logging and multi gpu support
-
Updated
Sep 21, 2020 - Python
Very minimal pytorch boilerplate with wandb logging and multi gpu support
Multiclass Classification Model to predict the effective writings of essays written by students ranging from 6th-12th grade. The data was take from the Kaggle Competition.
This repository is used to keep track of the code I have written for the recommendation system.
A template repository to quickly get started on a new Machine Learning project.
Fall 2021 Introduction to Deep Learning - Homework 1 Part 2 (Frame Level Classification of Speech)
Showcase an MLOps project using Prefect, Wandb and GithubActions
Contrastive learning for multilingual complex named entity recognition. Bert + CRF model.
Using a supercar and a common car datasets, we analyzed car variables with prices. Also created two machine learning models to create more relationships between variables and predict prices
Recommendation Engine powered by Matrix Factorization.
All Assignments of the course, Statistical Methods in AI at IIITH, Monsoon 2024
In this project we have developed a Deep Autoencoder using Dense Neural Networks to perform dimensionality reduction on MNIST and FMNIST datasets. The project includes training, saving, and evaluating models using PyTorch. Utilized the Weight & Biases library for monitoring and comparison of model performance
🐋 Template for ML projects using VSCode Development Containers
First project to implement from data 2 deployment on streamlit
TinyML tools for and with WandB
Vectorized CNN implementation from scratch using only numpy
Fine tuning Mistral-7b with PEFT(Parameter Efficient Fine-Tuning) and LoRA(Low-Rank Adaptation) on Puffin Dataset(multi-turn conversations between GPT-4 and real humans)
Julia macros for logging to Weights & Biases (wandb.ai).
Add a description, image, and links to the wandb topic page so that developers can more easily learn about it.
To associate your repository with the wandb topic, visit your repo's landing page and select "manage topics."