We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
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
CpuVector vec(*bufs_[PARAMETER_VALUE].get()); Header header; header.version = kFormatVersion; header.valueSize = sizeof(real); header.size = getSize(); CHECK_EQ(header.size, vec.getSize()); CHECK(s.write(reinterpret_cast<char*>(&header), sizeof(header))) << "Fail to write parameter " << getName(); CHECK(s.write(reinterpret_cast<char*>(vec.getData()), header.size * sizeof(real))) << "Fail to write parameter " << getName(); if (config_.is_sparse()) { CpuIVector rows(*intBufs_[PARAMETER_ROWS].get()); CpuIVector cols(*intBufs_[PARAMETER_COLS].get()); CHECK(s.write(reinterpret_cast<char*>(rows.getData()), rows.getSize() * sizeof(int))) << "Fail to write parameter " << getName(); CHECK(s.write(reinterpret_cast<char*>(cols.getData()), cols.getSize() * sizeof(int))) << "Fail to write parameter " << getName(); } return true; }
Parameter稀疏存储(特指训练完毕后,存储并用户预测的情况)的使用场景 ?
请给出sparse的全面解释。
The text was updated successfully, but these errors were encountered:
weight是sparse一种场合是Pruning模型压缩方法: https://arxiv.org/abs/1510.00149 。另外fc_layer本身也支持weight是稀疏存储。
Sorry, something went wrong.
#890
fleet api doc (PaddlePaddle#948)
6df389c
* fleet api * fleet api * fleet api
Docs and Clear-api Modify (PaddlePaddle#948)
93cdd64
* modify docs * modify docs * del inheritance * add inheri
backyes
Successfully merging a pull request may close this issue.
Parameter稀疏存储(特指训练完毕后,存储并用户预测的情况)的使用场景 ?
请给出sparse的全面解释。
The text was updated successfully, but these errors were encountered: