[go: up one dir, main page]
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.

Sign up for GitHub

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

Jump to bottom

模型优化问题 #1535

Open
1 task done
arcral opened this issue Aug 17, 2020 · 2 comments
Open
1 task done

模型优化问题 #1535

arcral opened this issue Aug 17, 2020 · 2 comments
Assignees
@hankcs
Labels
feature request Suggest an idea for this project

Comments

@arcral
Copy link
arcral commented Aug 17, 2020

Describe the feature and the current behavior/state.
我想问一下是否可以考虑使用tensorflow-lite或者openvino作为推理后端，这样的话可以避免依赖整个tensorflow或pytorch库，也能提高加载速度，对于gpu资源不是很丰富的用户也更加友好，能够提升桌面cpu端的推理速度。
如果使用transformer的话，是否可以使用模型蒸馏减少堆叠层数（例如, DistilBert），同样能提升推理速度

Will this change the current api? How?
no

Who will benefit with this feature?
everyone who use this

Are you willing to contribute it (Yes/No):
可以做，但是可能没有足够的资源做

System information

  • OS Platform and Distribution (e.g., Linux Ubuntu 16.04):
  • Python version:
  • HanLP version: v2.0.0-alpha.0

Any other info
no

  • I've carefully completed this form.
@arcral arcral added the feature request Suggest an idea for this project label Aug 17, 2020
@hankcs
Copy link
Owner
hankcs commented Aug 17, 2020

Thank you for your advices. In alpha stage, I mostly focus on replicating/improving SOTA papers. I will look into the engineer part in the future. Deployment is not easy, contribution from everyone (especially industry users) is welcomed.

@hankcs
Copy link
Owner
hankcs commented Jul 9, 2021

论坛上的相关讨论：https://bbs.hankcs.com/t/topic/3834

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@hankcs hankcs

Labels
feature request Suggest an idea for this project
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@hankcs @arcral