Comments (8)
anilsemizoglu
commented on July 18, 2024
2
changing line 5 in keras2cpp.py to the following line has removed the tuple error.
from tensorflow.keras.layers import (
from keras2cpp.
comntr
commented on July 18, 2024
Attaching the 7 MB weights file: weights.zip
from keras2cpp.
comntr
commented on July 18, 2024
I've added some logging and got this:
def export_model(model, filename):
with open(filename, 'wb') as f:
layers = [layer for layer in model.layers
if type(layer).__name__ not in ['Dropout']]
f.write(struct.pack('I', len(layers)))
for layer in layers:
print(f'Exporting layer: {type(layer)}')
f.write(struct.pack('I', LAYERS.index(type(layer)) + 1))
export(layer, f)Exporting layer: <class 'keras.layers.embeddings.Embedding'>
Exporting layer: <class 'keras.layers.recurrent.LSTM'>
Exporting layer: <class 'keras.layers.recurrent.LSTM'>
Exporting layer: <class 'keras.layers.recurrent.LSTM'>
Exporting layer: <class 'keras.layers.wrappers.TimeDistributed'>
Traceback (most recent call last):
File "h5-parser.py", line 14, in <module>
export_model(model, 'example.model')
File "/home/user1/data/src/keras2cpp/keras2cpp.py", line 214, in export_model
f.write(struct.pack('I', LAYERS.index(type(layer)) + 1))
ValueError: tuple.index(x): x not in tuple
from keras2cpp.
comntr
commented on July 18, 2024
I guess it's just this lib doesn't implement the Keras TimeDistributed layer.
"Fixing" it this way in the original keras-char-rnn model:
# model.add(TimeDistributed(Dense(vocab_size)))
model.add(Dense(vocab_size))Training still converges, the model still works, though it's probably weaker now.
Regarding the cmake problem. I have no idea what causes it, but running just cmake . produces quite a bit of output without errors. I don't know how to get it compile the files, but running g++ directly works:
g++ -std=gnu++17 cpp_model.cc src/*.cc src/layers/*.cc
Now I could run ./a.out and it indeed printed a long vector - supposedly that softmax output with the 256 probabilities.
from keras2cpp.
oleg3630
commented on July 18, 2024
I'm having the same problem:
Exporting layer: <class 'keras.engine.input_layer.InputLayer'>
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
<ipython-input-12-f273440699f5> in <module>
1 print("saving keras2cpp model")
----> 2 export_model(model, 'segmentation.model')
3 print("done!")
4 #model.save('segmentation_16classes_'+str(imageSize[0])+'_final.h5')
5 #!ls
~\GoogleDrive/NN\keras2cpp.py in export_model(model, filename)
212 for layer in layers:
213 print(f'Exporting layer: {type(layer)}')
--> 214 f.write(struct.pack('I', LAYERS.index(type(layer)) + 1))
215 export(layer, f)
ValueError: tuple.index(x): x not in tuple
My model is Unet from segmentation_models with variable input size (None, None):
model = Unet('resnet50', input_shape=(None, None, 3), intencoder_weights='imagenet', encoder_freeze = True, classes=classesCount, decoder_filters=(384, 256, 192, 128, 64), activation='softmax')
Is this library incompatible with None input size or what?
Thank you in advance!
from keras2cpp.
furkank75
commented on July 18, 2024
@oleg3630 have you solve this problem? I would like convert the unet from Keras to cpp.
from keras2cpp.
furkank75
commented on July 18, 2024
Model: "model_1"
__________________________________________________________________________________________________
Layer (type) Output Shape Param # Connected to
==================================================================================================
input_1 (InputLayer) (None, 256, 256, 1) 0
__________________________________________________________________________________________________
conv2d_1 (Conv2D) (None, 256, 256, 24) 240 input_1[0][0]
__________________________________________________________________________________________________
batch_normalization_1 (BatchNor (None, 256, 256, 24) 96 conv2d_1[0][0]
__________________________________________________________________________________________________
activation_1 (Activation) (None, 256, 256, 24) 0 batch_normalization_1[0][0]
__________________________________________________________________________________________________
conv2d_2 (Conv2D) (None, 256, 256, 24) 5208 activation_1[0][0]
__________________________________________________________________________________________________
batch_normalization_2 (BatchNor (None, 256, 256, 24) 96 conv2d_2[0][0]
__________________________________________________________________________________________________
activation_2 (Activation) (None, 256, 256, 24) 0 batch_normalization_2[0][0]
__________________________________________________________________________________________________
max_pooling2d_1 (MaxPooling2D) (None, 128, 128, 24) 0 activation_2[0][0]
__________________________________________________________________________________________________
dropout_1 (Dropout) (None, 128, 128, 24) 0 max_pooling2d_1[0][0]
__________________________________________________________________________________________________
conv2d_3 (Conv2D) (None, 128, 128, 48) 10416 dropout_1[0][0]
__________________________________________________________________________________________________
batch_normalization_3 (BatchNor (None, 128, 128, 48) 192 conv2d_3[0][0]
__________________________________________________________________________________________________
activation_3 (Activation) (None, 128, 128, 48) 0 batch_normalization_3[0][0]
__________________________________________________________________________________________________
conv2d_4 (Conv2D) (None, 128, 128, 48) 20784 activation_3[0][0]
__________________________________________________________________________________________________
batch_normalization_4 (BatchNor (None, 128, 128, 48) 192 conv2d_4[0][0]
__________________________________________________________________________________________________
activation_4 (Activation) (None, 128, 128, 48) 0 batch_normalization_4[0][0]
__________________________________________________________________________________________________
max_pooling2d_2 (MaxPooling2D) (None, 64, 64, 48) 0 activation_4[0][0]
__________________________________________________________________________________________________
dropout_2 (Dropout) (None, 64, 64, 48) 0 max_pooling2d_2[0][0]
__________________________________________________________________________________________________
conv2d_5 (Conv2D) (None, 64, 64, 96) 41568 dropout_2[0][0]
__________________________________________________________________________________________________
batch_normalization_5 (BatchNor (None, 64, 64, 96) 384 conv2d_5[0][0]
__________________________________________________________________________________________________
activation_5 (Activation) (None, 64, 64, 96) 0 batch_normalization_5[0][0]
__________________________________________________________________________________________________
conv2d_6 (Conv2D) (None, 64, 64, 96) 83040 activation_5[0][0]
__________________________________________________________________________________________________
batch_normalization_6 (BatchNor (None, 64, 64, 96) 384 conv2d_6[0][0]
__________________________________________________________________________________________________
activation_6 (Activation) (None, 64, 64, 96) 0 batch_normalization_6[0][0]
__________________________________________________________________________________________________
max_pooling2d_3 (MaxPooling2D) (None, 32, 32, 96) 0 activation_6[0][0]
__________________________________________________________________________________________________
dropout_3 (Dropout) (None, 32, 32, 96) 0 max_pooling2d_3[0][0]
__________________________________________________________________________________________________
conv2d_7 (Conv2D) (None, 32, 32, 192) 166080 dropout_3[0][0]
__________________________________________________________________________________________________
batch_normalization_7 (BatchNor (None, 32, 32, 192) 768 conv2d_7[0][0]
__________________________________________________________________________________________________
activation_7 (Activation) (None, 32, 32, 192) 0 batch_normalization_7[0][0]
__________________________________________________________________________________________________
conv2d_8 (Conv2D) (None, 32, 32, 192) 331968 activation_7[0][0]
__________________________________________________________________________________________________
batch_normalization_8 (BatchNor (None, 32, 32, 192) 768 conv2d_8[0][0]
__________________________________________________________________________________________________
activation_8 (Activation) (None, 32, 32, 192) 0 batch_normalization_8[0][0]
__________________________________________________________________________________________________
max_pooling2d_4 (MaxPooling2D) (None, 16, 16, 192) 0 activation_8[0][0]
__________________________________________________________________________________________________
dropout_4 (Dropout) (None, 16, 16, 192) 0 max_pooling2d_4[0][0]
__________________________________________________________________________________________________
conv2d_9 (Conv2D) (None, 16, 16, 384) 663936 dropout_4[0][0]
__________________________________________________________________________________________________
batch_normalization_9 (BatchNor (None, 16, 16, 384) 1536 conv2d_9[0][0]
__________________________________________________________________________________________________
activation_9 (Activation) (None, 16, 16, 384) 0 batch_normalization_9[0][0]
__________________________________________________________________________________________________
conv2d_10 (Conv2D) (None, 16, 16, 384) 1327488 activation_9[0][0]
__________________________________________________________________________________________________
batch_normalization_10 (BatchNo (None, 16, 16, 384) 1536 conv2d_10[0][0]
__________________________________________________________________________________________________
activation_10 (Activation) (None, 16, 16, 384) 0 batch_normalization_10[0][0]
__________________________________________________________________________________________________
conv2d_transpose_1 (Conv2DTrans (None, 32, 32, 192) 663744 activation_10[0][0]
__________________________________________________________________________________________________
concatenate_1 (Concatenate) (None, 32, 32, 384) 0 conv2d_transpose_1[0][0]
activation_8[0][0]
__________________________________________________________________________________________________
dropout_5 (Dropout) (None, 32, 32, 384) 0 concatenate_1[0][0]
__________________________________________________________________________________________________
conv2d_11 (Conv2D) (None, 32, 32, 192) 663744 dropout_5[0][0]
__________________________________________________________________________________________________
batch_normalization_11 (BatchNo (None, 32, 32, 192) 768 conv2d_11[0][0]
__________________________________________________________________________________________________
activation_11 (Activation) (None, 32, 32, 192) 0 batch_normalization_11[0][0]
__________________________________________________________________________________________________
conv2d_12 (Conv2D) (None, 32, 32, 192) 331968 activation_11[0][0]
__________________________________________________________________________________________________
batch_normalization_12 (BatchNo (None, 32, 32, 192) 768 conv2d_12[0][0]
__________________________________________________________________________________________________
activation_12 (Activation) (None, 32, 32, 192) 0 batch_normalization_12[0][0]
__________________________________________________________________________________________________
conv2d_transpose_2 (Conv2DTrans (None, 64, 64, 96) 165984 activation_12[0][0]
__________________________________________________________________________________________________
concatenate_2 (Concatenate) (None, 64, 64, 192) 0 conv2d_transpose_2[0][0]
activation_6[0][0]
__________________________________________________________________________________________________
dropout_6 (Dropout) (None, 64, 64, 192) 0 concatenate_2[0][0]
__________________________________________________________________________________________________
conv2d_13 (Conv2D) (None, 64, 64, 96) 165984 dropout_6[0][0]
__________________________________________________________________________________________________
batch_normalization_13 (BatchNo (None, 64, 64, 96) 384 conv2d_13[0][0]
__________________________________________________________________________________________________
activation_13 (Activation) (None, 64, 64, 96) 0 batch_normalization_13[0][0]
__________________________________________________________________________________________________
conv2d_14 (Conv2D) (None, 64, 64, 96) 83040 activation_13[0][0]
__________________________________________________________________________________________________
batch_normalization_14 (BatchNo (None, 64, 64, 96) 384 conv2d_14[0][0]
__________________________________________________________________________________________________
activation_14 (Activation) (None, 64, 64, 96) 0 batch_normalization_14[0][0]
__________________________________________________________________________________________________
conv2d_transpose_3 (Conv2DTrans (None, 128, 128, 48) 41520 activation_14[0][0]
__________________________________________________________________________________________________
concatenate_3 (Concatenate) (None, 128, 128, 96) 0 conv2d_transpose_3[0][0]
activation_4[0][0]
__________________________________________________________________________________________________
dropout_7 (Dropout) (None, 128, 128, 96) 0 concatenate_3[0][0]
__________________________________________________________________________________________________
conv2d_15 (Conv2D) (None, 128, 128, 48) 41520 dropout_7[0][0]
__________________________________________________________________________________________________
batch_normalization_15 (BatchNo (None, 128, 128, 48) 192 conv2d_15[0][0]
__________________________________________________________________________________________________
activation_15 (Activation) (None, 128, 128, 48) 0 batch_normalization_15[0][0]
__________________________________________________________________________________________________
conv2d_16 (Conv2D) (None, 128, 128, 48) 20784 activation_15[0][0]
__________________________________________________________________________________________________
batch_normalization_16 (BatchNo (None, 128, 128, 48) 192 conv2d_16[0][0]
__________________________________________________________________________________________________
activation_16 (Activation) (None, 128, 128, 48) 0 batch_normalization_16[0][0]
__________________________________________________________________________________________________
conv2d_transpose_4 (Conv2DTrans (None, 256, 256, 24) 10392 activation_16[0][0]
__________________________________________________________________________________________________
concatenate_4 (Concatenate) (None, 256, 256, 48) 0 conv2d_transpose_4[0][0]
activation_2[0][0]
__________________________________________________________________________________________________
dropout_8 (Dropout) (None, 256, 256, 48) 0 concatenate_4[0][0]
__________________________________________________________________________________________________
conv2d_17 (Conv2D) (None, 256, 256, 24) 10392 dropout_8[0][0]
__________________________________________________________________________________________________
batch_normalization_17 (BatchNo (None, 256, 256, 24) 96 conv2d_17[0][0]
__________________________________________________________________________________________________
activation_17 (Activation) (None, 256, 256, 24) 0 batch_normalization_17[0][0]
__________________________________________________________________________________________________
conv2d_18 (Conv2D) (None, 256, 256, 24) 5208 activation_17[0][0]
__________________________________________________________________________________________________
batch_normalization_18 (BatchNo (None, 256, 256, 24) 96 conv2d_18[0][0]
__________________________________________________________________________________________________
activation_18 (Activation) (None, 256, 256, 24) 0 batch_normalization_18[0][0]
__________________________________________________________________________________________________
conv2d_19 (Conv2D) (None, 256, 256, 1) 25 activation_18[0][0]
==================================================================================================
Total params: 4,863,865
Trainable params: 4,859,449
Non-trainable params: 4,416
__________________________________________________________________________________________________
Exporting layer: <class 'keras.engine.input_layer.InputLayer'>
Traceback (most recent call last):
File "D:/Projects/1_uNet/test.py", line 6, in <module>
export_model(model1, 'example.model')
File "D:\Projects\1_uNet\keras2cpp.py", line 214, in export_model
f.write(struct.pack('I', LAYERS.index(type(layer)) + 1))
ValueError: tuple.index(x): x not in tuple
The error came by conv2d..
from keras2cpp.
bevancollins
commented on July 18, 2024
changing line 5 in keras2cpp.py to the following line has removed the tuple error.
from tensorflow.keras.layers import (
the above didn't work for me but this did:
from tensorflow.python.keras.layers import (
from keras2cpp.
Related Issues (20)
- How to convert image to keras2cpp::Tensor? (Not really an Issue but a question)
- This project only support C
- This project only support std = c++17. Is there any other version? HOT 6
- Multi-class classification
- Assert error conv2d in dims != weight dims HOT 4
- Using it with qmake HOT 1
- Не собирается библиотека под UE4 HOT 3
- Не понимаю где косячу) or assert on BatchNormalization HOT 1
- VC++ dll HOT 1
- Convolutional networks HOT 1
- Error C2440 'initializing': cannot convert from 'initializer list' to 'keras2cpp::Tensor' HOT 2
- Inference result is different between keras model and c++
- Inference on c++ model fails: file cannot be read
- Cannot find test directory and files in that directory HOT 1
- embedding layer with mask_zero = True produces incorrect predictions
- VC++ Conv2D runtime error
- /
- Initialize a 4-dim Tensor for inference HOT 4
- Using from cpp project in VS2022 (Windows)
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from keras2cpp.