pulkit1joshi / simgnn Goto Github PK
View Code? Open in Web Editor NEWKeras implementation of "SimGNN: A Neural Network Approach to Fast Graph Similarity Computation". Includes synthetic GED data.
License: MIT License
Keras implementation of "SimGNN: A Neural Network Approach to Fast Graph Similarity Computation". Includes synthetic GED data.
License: MIT License
Currently, the model is using the IMDB dataset in the format of JSON files. The next task is to generate a synthetic data-set for better performance checking.
Data format :
The graph needs to be connected and exported in JSON format. GED can be calculated using any standard algorithm for 16 nodes.
Currently, the network is fully established, however, few out-put dimensions (NTN layer and Dense layer) are not up to the paper given in read me. (Check SimGNN paper, experimental data). So the network needs be changed to get exactly same results as given in the paper.
ImportError: cannot import name 'parameter_parser' from 'parser' (unknown location)
how can I fix it?
hi~ I had an error following... Could you know the clue of this error?,,
and,, i wanna know your tensor version.
thank you advance!!
==============================================
Traceback (most recent call last):
File "./src/main.py", line 81, in
main()
File "./src/main.py", line 60, in main
model = simgnn(parser)
File "D:\ys_wj\SimGNN-keras\SimGNN-main\src\simgnn.py", line 28, in simgnn
x = shared_attention(x[0])
File "C:\Users\numa97\Anaconda3\envs\gnn\lib\site-packages\tensorflow_core\python\keras\engine\base_layer.py", line 721, in call
base_layer_utils.create_keras_history(inputs)
File "C:\Users\numa97\Anaconda3\envs\gnn\lib\site-packages\tensorflow_core\python\keras\engine\base_layer_utils.py", line 189, in create_keras_history
_, created_layers = _create_keras_history_helper(tensors, set(), [])
File "C:\Users\numa97\Anaconda3\envs\gnn\lib\site-packages\tensorflow_core\python\keras\engine\base_layer_utils.py", line 260, in _create_keras_history_helper
layer_inputs, op.outputs)
File "C:\Users\numa97\Anaconda3\envs\gnn\lib\site-packages\tensorflow_core\python\keras\engine\base_layer.py", line 2032, in _add_inbound_node
input_tensors)
File "C:\Users\numa97\Anaconda3\envs\gnn\lib\site-packages\tensorflow_core\python\util\nest.py", line 569, in map_structure
structure[0], [func(*x) for x in entries],
File "C:\Users\numa97\Anaconda3\envs\gnn\lib\site-packages\tensorflow_core\python\util\nest.py", line 569, in
structure[0], [func(*x) for x in entries],
File "C:\Users\numa97\Anaconda3\envs\gnn\lib\site-packages\tensorflow_core\python\keras\engine\base_layer.py", line 2031, in
inbound_layers = nest.map_structure(lambda t: t._keras_history.layer,
AttributeError: 'tuple' object has no attribute 'layer'
Issue: The error before save and after saving do not match. The problem might be with the custom layers used. This needs to be checked.
How can I use Pytorch Geometric to implement your method?
Hey,
I tried testing it with only one training data set:
{
"labels_1": ["11", "11", "9"],
"labels_2": ["8", "11", "5"],
"graph_2": [[0,1],[1,2]],
"ged": 11,
"graph_1": [[0,1],[1,3]]
}
Which results in the following error:
ValueError: Dimensions must be equal, but are 8 and 16 for '{{node functional_1/graph_conv/MatMul_1}} = MatMul[T=DT_FLOAT, transpose_a=false, transpose_b=false](functional_1/graph_conv/Reshape, functional_1/graph_conv/Reshape_1)' with input shapes: [3,8], [16,64].
So how does this work. I tried to wrap my head around this but it is not clear to me what graph pairs work and which doesn't
Codebeat gave 3 warnings regarding Attention Branch Condition being too high. This could be fixed.
https://codebeat.co/projects/github-com-pulkit1joshi-simgnn-main
Hi,
How are you generating embeddings for the graph? Which python files are used for the same . I tried to look for the files but did not succeed.
Thanks in advance.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.