{"id":13958328,"url":"https://github.com/weiyinwei/MMGCN","last_synced_at":"2025-07-20T23:31:00.210Z","repository":{"id":40625833,"uuid":"200448911","full_name":"weiyinwei/MMGCN","owner":"weiyinwei","description":"MMGCN: Multi-modal Graph Convolution Network forPersonalized Recommendation of Micro-video","archived":false,"fork":false,"pushed_at":"2021-06-30T07:17:55.000Z","size":50,"stargazers_count":275,"open_issues_count":11,"forks_count":52,"subscribers_count":6,"default_branch":"master","last_synced_at":"2024-08-09T13:18:33.584Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/weiyinwei.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null}},"created_at":"2019-08-04T04:20:36.000Z","updated_at":"2024-08-07T09:07:45.000Z","dependencies_parsed_at":"2022-09-20T12:25:40.858Z","dependency_job_id":null,"html_url":"https://github.com/weiyinwei/MMGCN","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/weiyinwei%2FMMGCN","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/weiyinwei%2FMMGCN/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/weiyinwei%2FMMGCN/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/weiyinwei%2FMMGCN/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/weiyinwei","download_url":"https://codeload.github.com/weiyinwei/MMGCN/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":226844976,"owners_count":17691137,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2024-08-08T13:01:29.655Z","updated_at":"2024-11-28T01:31:42.083Z","avatar_url":"https://github.com/weiyinwei.png","language":"Python","funding_links":[],"categories":["其他_推荐系统"],"sub_categories":["网络服务_其他"],"readme":"\n\n# MMGCN: Multi-modal Graph Convolution Network for Personalized Recommendation of Micro-video\nThis is our Pytorch implementation for the paper:  \n\u003e Yinwei Wei, Xiang Wang, Liqiang Nie, Xiangnan He, Richang Hong, and Tat-Seng Chua(2019). MMGCN: Multi-modal Graph Convolution Network for Personalized Recommendation of Micro-video. In ACM MM`19, NICE, France,Oct. 21-25, 2019  \nAuthor: Dr. Yinwei Wei (weiyinwei at hotmail.com)\n\n## Introduction\nMulti-modal Graph Convolution Network is a novel multi-modal recommendation framework based on graph convolutional networks, explicitly modeling modal-specific user preferences to enhance micro-video recommendation. We update the code and use the full-ranking strategy for validation and testing.\n\n## Citation\nIf you want to use our codes and datasets in your research, please cite:\n\n``` \n@inproceedings{MMGCN,\n  title     = {MMGCN: Multi-modal graph convolution network for personalized recommendation of micro-video},\n  author    = {Wei, Yinwei and \n               Wang, Xiang and \n               Nie, Liqiang and \n               He, Xiangnan and \n               Hong, Richang and \n               Chua, Tat-Seng},\n  booktitle = {Proceedings of the 27th ACM International Conference on Multimedia},\n  pages     = {1437--1445},\n  year      = {2019}\n}\n``` \n\n\n## Environment Requirement\nThe code has been tested running under Python 3.5.2. The required packages are as follows:\n- Pytorch == 1.1.0\n- torch-cluster == 1.4.2\n- torch-geometric == 1.2.1\n- torch-scatter == 1.2.0\n- torch-sparse == 0.4.0\n- numpy == 1.16.0\n\n## Example to Run the Codes\nThe instruction of commands has been clearly stated in the codes.\n- Kwai dataset  \n```python main.py --model_name='MMGCN' --l_r=0.0005 --weight_decay=0.1 --batch_size=1024 --dim_latent=64 --num_workers=30 --aggr_mode='mean' --num_layer=2 --concat=False```\n- Tiktok dataset  \n`python main.py --model_name='MMGCN' --l_r=0.0005 --weight_decay=0.1 --batch_size=1024 --dim_latent=64 --num_workers=30 --aggr_mode='mean' --num_layer=2 --concat=False`\n- Movielens dataset  \n`python main.py --model_name='MMGCN' --l_r=0.0001 --weight_decay=0.0001 --batch_size=1024 --dim_latent=64 --num_workers=30 --aggr_mode='mean' --num_layer=2 --concat=False`  \n\nSome important arguments:  \n\n\n- `model_name`: \n  It specifies the type of model. Here we provide three options: \n  1. `MMGCN` (by default) proposed in MMGCN: Multi-modal Graph Convolution Network for Personalized Recommendation of Micro-video, ACM MM2019. Usage: `--model_name='MMGCN'`\n  2. `VBPR` proposed in [VBPR: Visual Bayesian Personalized Ranking from Implicit Feedback](https://arxiv.org/abs/1510.01784), AAAI2016. Usage: `--model_name 'VBPR'`  \n  3. `ACF` proposed in [Attentive Collaborative Filtering: Multimedia Recommendation with Item- and Component-Level Attention\n](https://dl.acm.org/citation.cfm?id=3080797), SIGIR2017. Usage: `--model_name 'ACF'`  \n  4. `GraphSAGE` proposed in [Inductive Representation Learning on Large Graphs](https://arxiv.org/abs/1706.02216), NIPS2017. Usage: `--model_name 'GraphSAGE'`\n  5. `NGCF` proposed in [Neural Graph Collaborative Filtering](https://arxiv.org/abs/1905.08108), SIGIR2019. Usage: `--model_name 'NGCF'`  \n\n\n- `aggr_mode` \n  It specifics the type of aggregation layer. Here we provide three options:  \n  1. `mean` (by default) implements the mean aggregation in aggregation layer. Usage `--aggr_mode 'mean'`\n  2. `max` implements the max aggregation in aggregation layer. Usage `--aggr_mode 'max'`\n  3. `add` implements the sum aggregation in aggregation layer. Usage `--aggr_mode 'add'`\n  \n  \n- `concat`:\n  It indicates the type of combination layer. Here we provide two options:\n  1. `concat`(by default) implements the concatenation combination in combination layer. Usage `--concat 'True'`\n  2. `ele` implements the element-wise combination in combination layer. Usage `--concat 'False'`\n## Dataset\nWe provide three processed datasets: Kwai, Tiktok, and Movielnes.  \n- You can find the full version of recommendation datasets via [Kwai](https://www.kuaishou.com/activity/uimc), [Tiktok](http://ai-lab-challenge.bytedance.com/tce/vc/), and [Movielens](https://grouplens.org/datasets/movielens/).\nSince the copyright of datasets, we cannot release them directly. \nTo facilate the line of research, we provide some toy datasets[[BaiduPan](https://pan.baidu.com/s/1BODXP7iihw8qtxpLeEv_XA)](code: zsye) or [[GoogleDriven]](https://drive.google.com/file/d/1NoisyVDFWykTszSIbHdeoBrKn0t-D0ps/view?usp=sharing). \nAnyone needs the full datasets, please contact the owner of datasets. \n\n||#Interactions|#Users|#Items|Visual|Acoustic|Textual|\n|:-|:-|:-|:-|:-|:-|:-|\n|Kwai|1,664,305|22,611|329,510|2,048|-|100|\n|Tiktok|726,065|36,656|76,085|128|128|128|\n|Movielens|1,239,508|55,485|5,986|2,048|128|100|\n\n-`train.npy`\n   Train file. Each line is a user with her/his positive interactions with items: (userID and micro-video ID)  \n-`val.npy`\n   Validation file. Each line is a user several positive interactions with items: (userID and micro-video ID)  \n-`test.npy`\n   Test file. Each line is a user with several positive interactions with items: (userID and micro-video ID)  \n\n\nCopyright (C) \u003cyear\u003e  Shandong University\n\nThis program is licensed under the GNU General Public License 3.0 (https://www.gnu.org/licenses/gpl-3.0.html). Any derivative work obtained under this license must be licensed under the GNU General Public License as published by the Free Software Foundation, either Version 3 of the License, or (at your option) any later version, if this derivative work is distributed to a third party.\n\nThe copyright for the program is owned by Shandong University. For commercial projects that require the ability to distribute the code of this program as part of a program that cannot be distributed under the GNU General Public License, please contact \u003cweiyinwei@hotmail.com\u003e to purchase a commercial license.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fweiyinwei%2FMMGCN","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fweiyinwei%2FMMGCN","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fweiyinwei%2FMMGCN/lists"}