MMdnn | help users inter-operate among different deep learning | Machine Learning library

by microsoft Python Version: 0.3.1 License: MIT

X-Ray Key Features Code Snippets(3)Community Discussions(5)Vulnerabilities Install Support

kandi X-RAY | MMdnn Summary

MMdnn is a Python library typically used in Artificial Intelligence, Machine Learning, Deep Learning, Pytorch, Tensorflow, Keras applications. MMdnn has no bugs, it has no vulnerabilities, it has build file available, it has a Permissive License and it has medium support. You can download it from GitHub.

MMdnn is a comprehensive and cross-framework tool to convert, visualize and diagnose deep learning (DL) models. The "MM" stands for model management, and "dnn" is the acronym of deep neural network.

Support

Quality

Security

License

Reuse

Support

MMdnn has a medium active ecosystem.

It has 5688 star(s) with 977 fork(s). There are 185 watchers for this library.

It had no major release in the last 12 months.

There are 324 open issues and 289 have been closed. On average issues are closed in 138 days. There are 11 open pull requests and 0 closed requests.

It has a neutral sentiment in the developer community.

The latest version of MMdnn is 0.3.1

Quality

MMdnn has 0 bugs and 0 code smells.

Security

MMdnn has no vulnerabilities reported, and its dependent libraries have no vulnerabilities reported.

MMdnn code analysis shows 0 unresolved vulnerabilities.

There are 0 security hotspots that need review.

License

MMdnn is licensed under the MIT License. This license is Permissive.

Permissive licenses have the least restrictions, and you can use them in most projects.

Reuse

MMdnn releases are available to install and integrate.

Build file is available. You can build the component from source.

Installation instructions, examples and code snippets are available.

It has 29830 lines of code, 1721 functions and 214 files.

It has high code complexity. Code complexity directly impacts maintainability of the code.

Top functions reviewed by kandi - BETA

kandi has reviewed MMdnn and discovered the below as its top functions. This is intended to give you an instant insight into MMdnn implemented functionality, and help decide if they suit your requirements.

Convert a reset network to its final endpoints .
Performs an expanded convolution .
Gets the mobilenet version .
This is a helper function for debugging .
Performs an Inception resnet operation .
Convert a batch normalization layer to a batch norm .
Builds the nas net .
Renames the convolution
Emits a pooling layer .
Performs a normal mobilization network .

Get all kandi verified functions for this library.

MMdnn Key Features

No Key Features are available at this moment for MMdnn.

MMdnn Examples and Code Snippets

Train_Caffe_model

C++

Lines of Code : 21

License : No License

Copy

cd $SIMCAM_SDK/train/data/Images_xmls/
python3 video2img.py

python3 video2img.py -h
usage: video2img.py [-h] [--input INPUT] [--output OUTPUT]
                    [--num NUMFRAMEPERSECOND]
optional arguments:
  -h, --help            show this help m

Sketch Pytorch,How to convert a keras model to a pytorch model

Python

Lines of Code : 11

License : Strong Copyleft (GPL-3.0)

Copy

Parse file [ir.pb] with binary format successfully.
Pytorch Emitter has not supported operator [UpSampling2D].
up_sampling2d_1
Pytorch Emitter has not supported operator [UpSampling2D].
up_sampling2d_2
Pytorch Emitter has not supported operator [UpSa

Are Safer Looking Neighborhoods More Lively? A Multimodal Investigation into Urban Life,Disclaimer

Jupyter Notebook

Lines of Code : 5

License : Permissive (MIT)

Copy

python -m mmdnn.conversion._script.convertToIR -f caffe -d kit_imagenet -n places205CNN_finetune.prototxt -w places205CNN_finetune_snap_iter_10000.caffemodel_save

python -m mmdnn.conversion._script.IRToCode --dstFramework pytorch --IRModelPath pytor

Community Discussions

Trending Discussions on MMdnn

Incomparable weight shape between caffe and tensorflow / keras

global variable (although defined) errors out as not defined in python

different result between Keras predict and Opencv DNN module

How to extract output of torch model in c++?

Using Tensorflow 2.X model on OpenCV

QUESTION

Incomparable weight shape between caffe and tensorflow / keras

Asked 2022-Feb-09 at 18:45

I am trying to convert a caffe model to keras, I have successfully been able to use both MMdnn and even caffe-tensorflow. The output I have are .npy files and .pb files. I have not had much luck with the .pb files, so I stuck to .npy files which contain the weights and biases. I have reconstructed an mAlexNet network as follows:

...

ANSWER

Answered 2022-Feb-09 at 18:45

The problem is the bias vector. It is shaped as a 4D tensor but Keras assumes it is a 1D tensor. Just flatten the bias vector:

Source https://stackoverflow.com/questions/71054511

QUESTION

global variable (although defined) errors out as not defined in python

Asked 2020-Jul-09 at 11:12

I have a rather weird problem in python: I am trying to run this script (it is auto generated using MMddn, which converts one neural network model to another - but this background is irrelevant to this question - just an FYI):

https://zerobin.net/?a8436f2ae6791499#dhZsFWXc91YpvlHajIqLY74MdeP8pE98E3IELiAD3bw=

but when I execute (using another script which calls this script) it I get:

...

ANSWER

Answered 2020-Jul-09 at 08:51

Python has name mangling inside class scopes, which means that names starting with two underscores are renamed. Inside the class scope, __weights_dict actually refers to _KitModel__weights_dict, i.e. not the name of the global variable.

As a fix, rename all occurrences of __weights_dict to _weights_dict.

Source https://stackoverflow.com/questions/62810436

QUESTION

different result between Keras predict and Opencv DNN module

Asked 2020-Jun-17 at 08:49

I get different results with this two snippets and cannot figure out why. I tried changing interpolation method on image loading for the keras version, without any luck (this should be correct because OpenCV uses bilinear to resize by default). The model used is the same, but has been converted to caffe because OpenCV DNN don't support H5 weights, conversion has been made with this tool. Results seem to good in both cases (predicted class, is a classification problem), but values are very different (i.e. 0.92 Keras vs 0.66 OpenCV for the top class in an image). Any idea ?

Keras version:

...

ANSWER

Answered 2020-Jun-17 at 08:49

It turns out that key point stands in the resize function, as described here Pillow (Keras image is just a wrapper for that) behaves differently than OpenCV, hence the different result. Answer is .. never mix different image processing library if it's not necessary.. :)

Source https://stackoverflow.com/questions/62413262

QUESTION

How to extract output of torch model in c++?

Asked 2020-Jun-06 at 12:16

I have got trained keras model and converted it using mmdnn. Then I try use it in c++ code:

...

ANSWER

Answered 2020-Jun-06 at 12:16

In this case the answer of model is tuple of 2 images. We can extract them by such way:

Source https://stackoverflow.com/questions/62158785

QUESTION

Using Tensorflow 2.X model on OpenCV

Asked 2020-Mar-11 at 23:08

I have to use a Tensorflow 2.X model with the OpenCV framework (v.4.X with C++).

To do this, I need a single .pb file or a .pb and a .pbtxt file, instead of a Tensorflow Saved Model like the one I have.

So my question is: Is there a way to convert a Saved Model in a format that OpenCV could read? Like, maybe, a caffe model?

I tried with MMdnn but it gives me a strange error:

...

ANSWER

Answered 2020-Mar-11 at 23:08

If you have the .h5 file, you can try this approach instead of MMdnn, using TensorFlow. The function converts the current session into a static computation graph to capture current states. Then you can write the graph in .pb format using tf.train.write_graph.

You can load the pretrained model with model = load_model('./model/keras_model.h5') before you freeze the graph. There is also a blog post for further explanation.

Source https://stackoverflow.com/questions/60562295

Community Discussions, Code Snippets contain sources that include Stack Exchange Network

Vulnerabilities

No vulnerabilities reported

Install MMdnn

You can get a stable version of MMdnn by. And make sure to have Python installed or you can try the newest version by.
MMdnn provides a docker image, which packages MMdnn and Deep Learning frameworks that we support as well as other dependencies. You can easily try the image with the following steps:.
Install Docker Community Edition(CE) Learn more about how to install docker
Pull MMdnn docker image docker pull mmdnn/mmdnn:cpu.small
Run image in an interactive mode docker run -it mmdnn/mmdnn:cpu.small

Support

Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to and actually do, grant us the rights to use your contribution. For details, visit https://cla.microsoft.com. When you submit a pull request, a CLA-bot will automatically determine whether you need to provide a CLA and decorate the PR appropriately (e.g., label, comment). Simply follow the instructions provided by the bot. You will only need to do this once across all repos using our CLA. This project has adopted the Microsoft Open Source Code of Conduct. For more information see the Code of Conduct FAQ or contact opencode@microsoft.com with any additional questions or comments.

Find more information at: