You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mxnet.apache.org by Chaitanya Bapat <ch...@gmail.com> on 2018/12/17 08:42:53 UTC

API change discussion to resolve inconsistencies in Gluon Model Zoo

Hello everyone,

As a contributor to Apache MXNet project, I wanted to ask the developer
community a few questions pertaining to Gluon Model Zoo API.

1. With regards to APIs for networks like - mobile nets, densenets,
resnets, squeezenets (note the case sensitivity of these APIs), they are
inconsistent compared to alexnet and inceptions.

By inconsistency I mean, **kwargs instead of mentioning all the parameters
that are required by the function. (e.g. pretrained=False, ctx=cpu(0))

Stating just 1 function here for example -

   - mxnet.gluon.model_zoo.vision.mobilenet_v2_0_25(**kwargs)
   <https://mxnet.incubator.apache.org/api/python/gluon/model_zoo.html#mxnet.gluon.model_zoo.vision.mobilenet_v2_0_25>

2. What does the community feel about this? Should we resolve this or it's
right the way it is? (Since this was an API breaking change, it was best to
ask the community before submitting a PR with changes)

3. What is the difference between APIs bearing same names but those in
Titlecase vs lowercase
e.g. mxnet.gluon.model_zoo.vision.AlexNet(classes=1000, **kwargs)
vs
mxnet.gluon.model_zoo.vision.alexnet(pretrained=False, ctx=cpu(0),
root='/home/jenkins_slave/.mxnet/models', **kwargs)

For reference and to track the above inconsistency, I have created a Github
Issue #13661 <https://github.com/apache/incubator-mxnet/issues/13661>

I would highly appreciate your replies to any or all of the above questions.
Thanks,
Chai


-- 
*Chaitanya Prakash Bapat*
*+1 (973) 953-6299*

[image: https://www.linkedin.com//in/chaibapat25]
<https://github.com/ChaiBapchya>[image: https://www.facebook.com/chaibapat]
<https://www.facebook.com/chaibapchya>[image:
https://twitter.com/ChaiBapchya] <https://twitter.com/ChaiBapchya>[image:
https://www.linkedin.com//in/chaibapat25]
<https://www.linkedin.com//in/chaibapchya/>

Re: API change discussion to resolve inconsistencies in Gluon Model Zoo

Posted by "Skalicky, Sam" <ss...@amazon.com.INVALID>.
I think it would be a good idea to do this in all the language bindings so that error messages can be appropriate and familiar in the user’s language rather than confusing exceptions coming from the c++ backend. Heres an issue I filed for tracking: https://github.com/apache/incubator-mxnet/issues/12286

Sam


On Dec 17, 2018, at 12:42 AM, Chaitanya Bapat <ch...@gmail.com>> wrote:

Hello everyone,

As a contributor to Apache MXNet project, I wanted to ask the developer
community a few questions pertaining to Gluon Model Zoo API.

1. With regards to APIs for networks like - mobile nets, densenets,
resnets, squeezenets (note the case sensitivity of these APIs), they are
inconsistent compared to alexnet and inceptions.

By inconsistency I mean, **kwargs instead of mentioning all the parameters
that are required by the function. (e.g. pretrained=False, ctx=cpu(0))

Stating just 1 function here for example -

  - mxnet.gluon.model_zoo.vision.mobilenet_v2_0_25(**kwargs)
  <https://mxnet.incubator.apache.org/api/python/gluon/model_zoo.html#mxnet.gluon.model_zoo.vision.mobilenet_v2_0_25>

2. What does the community feel about this? Should we resolve this or it's
right the way it is? (Since this was an API breaking change, it was best to
ask the community before submitting a PR with changes)

3. What is the difference between APIs bearing same names but those in
Titlecase vs lowercase
e.g. mxnet.gluon.model_zoo.vision.AlexNet(classes=1000, **kwargs)
vs
mxnet.gluon.model_zoo.vision.alexnet(pretrained=False, ctx=cpu(0),
root='/home/jenkins_slave/.mxnet/models', **kwargs)

For reference and to track the above inconsistency, I have created a Github
Issue #13661 <https://github.com/apache/incubator-mxnet/issues/13661>

I would highly appreciate your replies to any or all of the above questions.
Thanks,
Chai


--
*Chaitanya Prakash Bapat*
*+1 (973) 953-6299*

[image: https://www.linkedin.com//in/chaibapat25]
<https://github.com/ChaiBapchya>[image: https://www.facebook.com/chaibapat]
<https://www.facebook.com/chaibapchya>[image:
https://twitter.com/ChaiBapchya] <https://twitter.com/ChaiBapchya>[image:
https://www.linkedin.com//in/chaibapat25]
<https://www.linkedin.com//in/chaibapchya/>