-
This is should be "keras.applications", but it is "keras_applications" in .densenet
This is because files in /keras_contrib/applications is wrong
ghost updated
2 years ago
-
I have run example code Densenet.py and I just changed epoch and batch size into 300 and 64.(to keep the parameters be the same with paper.)
The test accuracy is 91.72% while the original paper got t…
-
加载博主的模型后,如果一次性识别多张图片的话每次第一张图片都比较慢,到第三四张图片的时候速度才上来,保持在0.2-0.8s左右。如果每次识别一张图片然后返回结果,然后在调用ocr识别下一张,这种模式的话每张图片就到了1s多,请问如何操作才能够使识别第一张图片的时间降到正常范围呢?我的加载模型用的下面的代码:
reload(densenet)
characters = keys.alphabet…
-
After reproducing your network, I found this figure in your paper may lead to some understanding, as each EDA module didn't take all preceding feature maps as input, so it's different from DenseNet.
…
-
Based on the starter code, replace the simple model with state-of-the-art ones e.g. DenseNet, Inception etc., to evaluate both the speed and accuracy.
-
after running the code got this error. how can i fix this? plz help.
i used nvidia titan xp gpu.
----------
Traceback (most recent call last):
File "", line 1, in
runfile('C:/Users/User/…
-
Hi, @KeremTurgutlu
I'm trying to run densenet121+unet on fastai to do semantic segmentation and find your code can support it. But ,I don't know where your densenet model come from.
chx_state_dic…
-
# Intro
The bug that I'll describe is a consequence of the current coordinate systems implementation and will disappear when it will be refactored. The bug needs to be fixed upstream, but has a conse…
-
![image](https://user-images.githubusercontent.com/77728649/155135520-b28d5a52-d1ef-4b90-97df-9be4ab4a6871.png)
大佬,我用你给的densenet权重训练出现了这个问题,这是为什么呢
-
请教,我看别人只用DenseNet也能做ocr识别, 为何还要加cptn?