-
Notifications
You must be signed in to change notification settings - Fork 1.6k
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
[LITE][NPU][XPU] Refine subgraph pass, and support NPU/XPU model generation at execution time #2576
[LITE][NPU][XPU] Refine subgraph pass, and support NPU/XPU model generation at execution time #2576
Conversation
…ubgraph pass to support build NPU/XPU model during kernel execution phase test=develop
…into hongming/support_runtime
…into hongming/support_runtime
…into hongming/support_runtime
…mize_tool test=develop
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
执行时做图优化,总体上与lite的设计不一致,因为会带来未知的动态性
但npu接口不一致的话也没有办法,也算是npu sdk本身的设计问题
需要注意部署库的大小
目前这个方案取了平衡,图优化(即子图生成)仍然在分析阶段做,只是将子图转为目标设备的模型这个步骤放在执行阶段,目的是根据设备特点动态生成目标设备的模型以保证兼容性,同时由于目标设备模型的生成和输入尺寸强绑定,因此执行阶段生成目标设备的模型是合理的。 对于不支持NPU/XPU的部署库libpaddle_lite_jni.so大小无影响。针对支持NPU的部署库libpaddle_lite_jni.so增量约为60KB。 |
…npu to fix the compiling error on tiny publish mode test=develop
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
…place test=develop
…raph op test=develop
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
背景问题:
解决办法:
其它改动: