-
Notifications
You must be signed in to change notification settings - Fork 413
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
fp16测试正常,量化成int8后没有预测框? #99
Comments
table文件: Conv_0_param_0 37.113487 42.639313 50.678776 36.664253 79.230194 32.896748 134.279312 141.645355 38.193413 51.205990 36.698421 31.202383 21.400766 64.979218 50.875202 30.667765 64.606087 41.935471 18.435116 38.045803 24.518932 189.656387 47.672409 32.356129 33.455742 164.158035 44.772491 78.754814 149.723984 38.045803 31.326496 147.067810 |
量化后的命名更改成以下试试:
|
测试了下,命名和你的完全一样,还是什么都检测不到.... |
请尝试一下该博客: |
量化后的输出节点也会跟着改变,是否已经进行了相应的更改? |
fp16测试正常,但是量化成int8后没有预测框(测试了很多图片都没有框),您帮忙看看我结果是正常的量化现象,还是我哪个步骤出错了呢???
步骤1:生成imagelist.txt 200张
步骤2:生成.table文件
步骤3:量化:fp16->int8
预测效果(左边fp16 右边int8):
The text was updated successfully, but these errors were encountered: