Skip to content
This repository has been archived by the owner on Aug 5, 2022. It is now read-only.

An improvement on regression and data reading-revision #215

Open
wants to merge 8 commits into
base: master
Choose a base branch
from

Conversation

tujie-jiangye
Copy link

The improvement of this version is mainly aimed at multiple label regression problems. By improving LMDB reading and writing, the Intel caffe LMDB database can directly support multi label regression without the h5py database as an intermediary. Through this improvement, we find that some native Caffe regression loss is not convergent or slow convergence, this improvement has good adaptability, and shows excellent results on similar problems.

…els, and regression problems are applied. The improved Caffe can handle the standard image multi label regression problem and is very adaptable to non-standard image data. Compared to the support of native Caffe to the regression problem, this modification not only makes the use more simple, but also has a good performance in the slow convergence or non convergence of loss on some regression problems.
delete some useful line
Revision of code specification
@tujie-jiangye tujie-jiangye changed the title An improvement on regression and data reading An improvement on regression and data reading-revision May 19, 2018
Copy link
Contributor

@jgong5 jgong5 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please explain why you remove the OPENMP related code.

PreclcRandomNumbers precalculated_rand_numbers;
this->data_transformer_->GenerateRandNumbers(precalculated_rand_numbers);
#pragma omp task firstprivate(offset, precalculated_rand_numbers, data, item_id)
#endif
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why do you remove these code?

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants