-
Hi,
Great to see that spatio-temporal networks are also being implemented in this library. I have a few problem with the gcn_lstm layer as follows
1)Why is the input dim limited to (batch size, seq…
-
hello,i used your model to train myself captcha images,but the loss is always 14,and don't convergence。Can you show me your training images?or tell me how to see them .I want to know whether my pictur…
-
Hi, Thanks for your excellent tutorial about deep learning for Human Activity Recognition, I learned a lot about data preprocessing.
When I read your code in the [notebook](https://github.com/mariusb…
-
I try like this:
python yolov3_darknet2caffe.py mobilenet_v1_yolov3.cfg mobilenet_v1_yolov3_final.weights mobilenet_yolov3.prototxt mobilenet_yolov3.caffemodel
and get an error:
F0627 23…
-
本文提供了利用frame-level 信息的两种思路:feature-pooling 以及 recurrent neutral networks,前者没有利用temporal 信息,后者则利用了temporal 信息。
(3d-convolution尝试的必要不大,Karpathy 指出这种结构just marginally better than single frame baseline.)…
-
Like it os in the subject looks like it is something in one unit
```console
+ cd oneDNN-3.5.3
+ DESTDIR=/home/tkloczko/rpmbuild/BUILDROOT/onednn-3.5.3-2.fc37.x86_64
+ /usr/bin/cmake --install …
-
### 第六周科研总结
- 研究了CBLSTM,阅读论文《Convolutional Recurrent Neural Network for Fault Diagnosis of High-Speed Train Bogie》
- 找到了暂时没有人做的方向: LSTM-FCN 网络结构,打算通过添加双向进行改进并应用在故障检测中
- 确定了先写BLSTM方向的论文
### 第七周…
-
I could successfully build, but I am not able to run the demo.When I execute the demo(no-gpu), the error is as follows:
Check failed: registry.count(type) == 1 (0 vs. 1) Unknown layer type: Python (k…
-
`
conv = conv2d(input_, kernel_feature_size, 1, kernel_size, name="kernel_%d" % kernel_size)
def conv2d(input_, output_dim, k_h, k_w, name="conv2d"):
with tf.variable_scope(name):
…
-
# Performance Issue
- Embedding dimension을 늘리는 경우
- Model을 깊게(복잡하게) 짜는 경우.(#param가 커지는 경우)
- #params의 문제가 아니라 연산이 많아지는 경우(대표적으로 Convolution을 복잡하게 사용하는 경우가 해당됨)
## Embedding Dimension
현재 구현, 실험 …