piggyback
by arunmallya
Code for Piggyback: Adapting a Single Network to Multiple Tasks by Learning to Mask Weights
AI summary
Task adaptation framework
Adapting a single network to multiple tasks by learning to mask weights
- stars
- 183
- forks
- 27
- watching
- 3
Similar projects
Found by comparing what the projects do, not just their names.
NLP Task Adaptation
An approach to transfer learning for NLP tasks using adversarial reprogramming and word-level task-specific embeddings.
Noisy Supervision Method
An implementation of Masking, a method to improve the robustness of deep neural networks under noisy supervision
Object detection framework
A PyTorch implementation of the Mask R-CNN architecture
Domain adaptation model
This project implements a deep learning-based approach to adapt semantic segmentation models from one domain to another.
ML automator
Automates machine learning model training using pre-set configurations and modular design.
ML library
This is a collection of machine learning algorithms implemented in Python 3.6.
Network reprogramming
This project enables reprogramming of pre-trained neural networks to work on new tasks by fine-tuning them on smaller datasets.
Temporal network embedder
A PyTorch implementation of a representation learning framework for dynamic temporal networks
Machine Learning Tutorials
A collection of examples and code snippets teaching machine learning concepts to security professionals through hands-on Python projects
Model alignment
This project demonstrates the effectiveness of reinforcement learning from human feedback (RLHF) in improving small language models like GPT-2.
Deep learning accelerator
A framework for accelerating PyTorch deep learning training
Machine Learning Algorithm
An implementation of an online federated learning algorithm with multiple kernels for personalized machine learning
Example weighting
Project implementing a method to improve deep learning model robustness by re-weighting examples with noisy labels
Label correction mechanism
This project explores how to adapt neural networks to noisy labels by introducing a mechanism that can learn to correct the errors.
Map-reduce framework
A map-reduce framework for Clojure that compiles to Apache Pig or Cascading without requiring prior knowledge of those systems.
