Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[🐛BUG] DCN在mvlen100k下uni100与full的AUC差距过大 #2117

Open
cywuuuu opened this issue Nov 27, 2024 · 2 comments
Open

[🐛BUG] DCN在mvlen100k下uni100与full的AUC差距过大 #2117

cywuuuu opened this issue Nov 27, 2024 · 2 comments
Assignees
Labels
bug Something isn't working

Comments

@cywuuuu
Copy link

cywuuuu commented Nov 27, 2024

描述这个 bug
对 bug 作一个清晰简明的描述。
DCN在mvlen100k下uni100与full的AUC差距过大, uni100时auc仅0.58左右,而full时auc为0.78左右,请问原因?谢谢
如何复现
复现这个 bug 的步骤:

  1. 您引入的额外 yaml 文件
load_col:
    inter: ['user_id', 'item_id', 'rating', 'timestamp']
    user: ['user_id', 'age', 'gender', 'occupation']
    item: ['item_id', 'release_year', 'class']
threshold: {'rating': 4}
normalize_all: True
# model config
embedding_size: 10
# Training and evaluation config
epochs: 500
train_batch_size: 4096
eval_batch_size: 4096


topk: [1,2,3, 10, 100]
train_neg_sample_args: ~
metrics: ['AUC']
valid_metric: AUC
# eval_args:
#   # mode: uni100
device: 'cuda:0'
  1. 您的代码
from recbole.quick_start import run_recbole

config_dict = {
    'device': 'cuda:0'  # 或 'cpu',根据您的需求
}

run_recbole(model='DCN', dataset='ml-100k', config_file_list=['test.yaml'], config_dict=config_dict)

  1. 您的运行脚本
    直接运行即可
@cywuuuu cywuuuu added the bug Something isn't working label Nov 27, 2024
@cywuuuu
Copy link
Author

cywuuuu commented Nov 27, 2024

请问这是否正常?还请问recbole支持对lightgcn求auc吗?谢谢 @Fotiligner

@BishopLiu
Copy link
Collaborator

BishopLiu commented Dec 24, 2024

@cywuuuu 你好,我们经过检查发现uni100对于数据和标签进行了修改,后续我们会加入assert避免错误的使用,full评估下的auc结果是正确的,recbole不支持对lightgcn求auc。

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working
Projects
None yet
Development

No branches or pull requests

2 participants