-
### 问题描述 Please describe your issue
https://paddlelite-demo.bj.bcebos.com/models/ssd_mobilenet_v1_relu_voc_int8_300_per_layer.tar.gz
这个ssd_mobilenet_v1_relu_voc_int8_300_per_layer是用paddle-detection训…
-
Hi
I want to convert an emotion recognition model to tflite. But after conversion, the model works much worse. She almost always predicts "neutral", "angry" or "happy". The model from face api demo w…
-
Pass the layer test for relu. Make sure the it runs on top of WebNN relu implementation.
-
# Arch
`Backbone-mobilenet` + `Decoder x 5` + `Final Conv + Activation`
- 先提取 Shape/(2^5) 的 `feature map`
- 再 upsample `2^5` 到原 resolution
Input | 448x448x3 - HxWxC
-- | --
Output | `448x44…
-
```
function relu(net, input)
output=ccall((:MakeRelu, "xxx.so"), Int64, (Int64, Int64), net, input)
return output
end
function relu_grad(net, input, dout)
output=ccall((:MakeReluG…
-
### 🐛 Describe the bug
Hi there, I am newcomer to model quantization. I have some problems and hope to get some advice and help from community. Thanks in advance!
Here is a demo model:
```py…
-
5. Define Models
Cell 3
Error:
---------------------------------------------------------------------------
AttributeError Traceback (most recent call last)
in ()…
-
On a deep cnn, I'm getting the error "Incompatible combination of dilation_rate with strides."
The model works without problems within keras. Is padding not implemented? Is there a workaround?
Arc…
-
Running the program for 1 iteration per frame (128x72) leaves the card out of memory (Geforce GT 630)
Note: This does work on a 970M
```
2_dreaming_time.exe -i ../images -o ../images_out --gpu 0 -i…
-
Hello, I think your magin loss is wrong, according to the paper, after relu, there is a square you forgot