i-SpaSP: Structured Neural Pruning via Sparse Signal Recovery

This is a public code repository for the publication:

i-SpaSP: Structured Neural Pruning via Sparse Signal Recovery
Cameron R Wolfe, Anastasios Kyrillidis

Environment/Dependencies

Requires anaconda to be installed (python3) Anaconda can be installed at https://www.anaconda.com/products/individual

conda create -n ispasp python=3.6 anaconda
conda activate ispasp
conda install pytorch torchvision torchaudio cudatoolkit=11.3 -c pytorch
pip install -r requirements.txt

File Arrangement

Here we summarize all files present in this repo and their purpose.

+-- prune_resnet34_ispasp.py : i-SpaSP implementation for ResNet34
+-- prune_mobilenetv2_ispasp.py : i-SpaSP implementation for MobileNetV2
+-- lib/ : utility/helper functions
|   +-- data.py: helper functions for handling data
|   +-- utils.py: helper functions for computing performance metrics
+-- scripts/ : contains python scripts for running pruning experiments
|   +-- prune_rn34_ispasp.py: run an i-SpaSP pruning experiment for ResNet34  
|   +-- prune_mbnv2_ispasp.py: run an i-SpaSP pruning experiment for MobileNetV2
+-- requirements.txt : dependencies for pruning experiments

i-SpaSP: Structured Neural Pruning via Sparse Signal Recovery

Related tags

Overview

i-SpaSP: Structured Neural Pruning via Sparse Signal Recovery

Environment/Dependencies

File Arrangement

Owner

Cameron Ronald Wolfe

Official implementation of AAAI-21 paper "Label Confusion Learning to Enhance Text Classification Models"

PyTorch implementation of 1712.06087 "Zero-Shot" Super-Resolution using Deep Internal Learning

Robust Consistent Video Depth Estimation

functorch is a prototype of JAX-like composable function transforms for PyTorch.

TensorFlow Metal Backend on Apple Silicon Experiments (just for fun)

Hierarchical Clustering: O(1)-Approximation for Well-Clustered Graphs

Pytorch Implementation of "Desigining Network Design Spaces", Radosavovic et al. CVPR 2020.

A simple, fast, and efficient object detector without FPN

Keywords : Streamlit, BertTokenizer, BertForMaskedLM, Pytorch

Automatically replace ONNX's RandomNormal node with Constant node.

The Official PyTorch Implementation of "LSGM: Score-based Generative Modeling in Latent Space" (NeurIPS 2021)

details on efforts to dump the Watermelon Games Paprium cart

Ranking Models in Unlabeled New Environments （iccv21）

A tensorflow model that predicts if the image is of a cat or of a dog.

Minimisation of a negative log likelihood fit to extract the lifetime of the D^0 meson (MNLL2ELDM)

PyTorch implementation of normalizing flow models

Elevation Mapping on GPU.

Asymmetric Bilateral Motion Estimation for Video Frame Interpolation, ICCV2021

PyTorch EO aims to make Deep Learning for Earth Observation data easy and accessible to real-world cases and research alike.

This initial strategy was developed specifically for larger pools and is based on taking a moving average and deriving Bollinger Bands to create a projected active liquidity range.