Linxyhaha / GeneRec

Generative Recommendation: Towards Next-generation Recommender Paradigm
https://arxiv.org/abs/2304.03516
56 stars 6 forks source link

GeneRec

This is the pytorch implementation of our paper:

Generative Recommendation: Towards Next-generation Recommender Paradigm

Environment

Usage

AI Editor

The codes for AI Editor are in './code/AI_Editor/' folder, including three tasks 1) thumbnail selection & thumbnail generation, 2) micro-video clipping, and 3) micro-video content editing.

Thumbnail selection

The personalized selected thumbnails can be obtained by running

cd code/AI_Editor
python select_thumbnail.py

Thumbnail generation

The personalized generated thumbnails can be obtained by running

cd code/AI_Editor/Thumbnail_gen
python scripts/generate_thumbnail.py

Micro-video clipping

The personalized clipped micro-videos can be obtained by running

cd code/AI_Editor
python clip_microVid.py

Micro-video content editing

This task requires training, which can be achieved by running

cd code/AI_Editor/Content_editing/scripts
sh train.sh

At inference stage, the personalized edited micro-videos can be obtained by running

cd code/AI_Editor/Content_editing/scripts
sh edit.sh

AI Creator

The codes for AI Creator are in './code/AI_Creator/' folder, including the task of micro-video content creation.

Micro-video content creation

This task also requires training, which can be achieved by running

cd code/AI_Creator/Content_creation/scripts
sh train.sh

At inference stage, the personalized created micro-videos can be obtained by running

cd code/AI_Creator/Content_creation/scripts
sh gen.sh

Data

Please feel free to send a request email for the data. The use of the data should be for research purpose only.