I am currently working with a very simple CNN and I am trying to figure out a way to make it into a Full CNN. By Full I mean not having any Dense layer and instead implement a deconvnet. The deconvnet part will be used as a localization part of the CNN.
What I am wondering is if there is a way to implement a deconvnet part in nolearn? In detail, a Deconv2dlayer and Unpooling? Has anyone else looked more into it or even done it?
Here is a reference for more details:
https://datascience.stackexchange.com/questions/8999/deconvolutional-network-in-semantic-segmentation
Related
I'm starting to look into the Qt WaylandCompositor and I am wondering if it is possible to get access to the composed frame of a WaylandOutput, preferably rendered off-screen. So that I can implement something similar than the Weston remoting plugin for streaming the composed scene to other devices.
Looking into the topic, I found the QWaylandView that offers access to a QWaylandBufferRef, but I guess this can be only used for a surface that is input to the composer... Is there something similar for the QWaylandOutput?
About the off-screen rendering, I found QOffscreenSurface, but I found no way so far on how to tell the output to use the OffscreenSurface instead of the window binding.
Does anyone have ideas on how to approach this?
Any help is deeply appreciated.
Hi and thank you for looking into this.
(Disclaimer: I have little-to-no technical background and would like to find the least complex solution. Ideally, only "connecting" different out-of-the-box components and no coding.)
HIGH-LEVEL PROBLEM:
I have trained a model for text classification using Google AutoML. I want to make this model available on a website, ie I want to enable visitors to enter their text and to receive the model's predicted class.
CONSIDERATIONS SO FAR: AutoML allows us to deploy the model via REST API and I understand that what I want are the API's PUT and GET function (right?). Ideally, I would use some form of plug-in or script to create an input field for the user which accepts the PUT and then delivers the GET.
Are you aware of any services for this? I'm also happy to host the website in an content management system like WordPress.
I'm very open regarding other approaches to solving my problem and highly appreciate any constructive input.
Many thanks!
AutoML Documentation https://cloud.google.com/natural-language/automl/docs/predict
EDIT Jan 10 There is another question related to this and a depo is shared which supposedly provided a solution. I'm not able to access the depo but the question might help you to understand my issue. Is there a way to use Googles AutoML with JavaScript?
EDIT Jan 16 I have learned that in order to provide the input to the model the POST function could be used instead of the PUT.
I'm building a custom SSIS data flow component that, as part of its process, takes multiple inputs and rolls them into a single output (along the lines of what the UNION ALL component does). I'm having a little bit of trouble implementing the runtime methods and having been looking for source code for asynchronous transformations that use multiple inputs that I can use as a model for my own code.
Does anyone know of available source code or have their own code they'd be willing to share with me to help get me on the right track? Much appreciated!
Off the top of my head, I'd look at the Programming Samples from MS's IS team as well as the Community Transform. Some of those are open source so if you can find one that's async, you should be able to see what you're missing.
My half-remembered logic is that you'll need to look at all the InputCollections for a given IDTSComponentMetaData100 and then work with that data.
So basically what I'm trying to do is make something like this:
http://store.kapit.biz/demo/treemap/prod/index.html#
I've searched long and hard to find a site that would show me some pseudocode (at least) or an algorithm as to how to implement this type of application. But everything I find are only applications that take in data and do all the work for me to create the graph. I need to actually create it for myself, just a simple one for now. Does anyone know where I can find this? What I want it to do is take data from a database and use it to create the treemap, then express it in a GUI like in the link above. If any specifics are needed I can provide it. I'm trying to do this in Flash Builder using Flex and using ColdFusion as the backend. Thanks!
Check this out
Axiis is a open source library based on Degrafa that enables you to make complex data visualizations.
Cheers
I've inherited a asp.net project and I'm finding that the code behind pages contain a lot of business logic.
I've decided that in most cases it's better to leave the working code in place than try and do a massive refactoring. However there are pages that perform functionality that could be re-used in command line utilities for batch processing. I'd like to focus my energies on these pages, refactoring out the business logic and referencing that in other utilities.
I am currently looking to refactor this particular page which has 6200 lines of code in the code behind. What I'm finding is it's very tedious work trying to locate the dependencies between the code behind and the page specific objects.
I'm wondering if anybody knows of a tool, a VS feature, or a method that would allow me to systematically locate and attack these dependencies? Something that will allow me to identify any area of the code that references ViewState, a textbox, panel, drop downlist, etc..., so I can move these references to method parameters and ultimately move this functionality out of the page class.
I would begin by examining any method that doesn't follow the Single Responsibility Principle and break them down so that they do. Once that's done you should have a handle on what the code does and you should be able to group the code more easily and move it into dedicated classes for those groups creating the necessary objects to use as you go. I find ReSharper is a very useful tool to help do all this. Ultimately you will always need to have a solid grasp of the fundamentals within the code before you can successfully refactor.
We've all been there at some point and you have my deepest sympathy but your willingness to attempt it at all means you're already heading in the right direction. Good Luck!
Wow first of all sorry to here that. Anyone who would put 6000 lines of code in code behind deserve to be smacked :)
Now I have done this kind of refactoring before. I would approach this in few steps:
(1) Create Logical Regions #Region and #EndRegion
Like - Save Methods , Load Methods or
(2) See if you can create physical business objects based on these regions in your business
layer
(3) Once you are done just refer all your code to appropriate class.
I understand you want some tool to make this all go away but I am afraid by doing this you will dig the hole dipper. Understanding the code and moving it method by method will give you better understanding.
One way I can immediately think is Compile then into Assemblies and Analyze the assembly using NDepend
http://www.ndepend.com/Features.aspx#DependencyCycle
(source: ndepend.com)