Search Results

Search found 28303 results on 1133 pages for 'multi site'.

Page 130/1133 | < Previous Page | 126 127 128 129 130 131 132 133 134 135 136 137  | Next Page >

  • Understanding a multilayer perceptron network

    - by Jonas Nielsen
    Hi all, I'm trying to understand how to train a multilayer, however, I'm having some trouble figuring out on choosing the amount of neurons in my network. For a specific task, I have four input sources that can each input one of three states. I guess that would mean four input neurons firing either 0, 1 or 2, but as far as I'm told, input should be kept binary? Furthermore am I having some issues choosing on the amount of neurons in the hidden layer. Any comments would be great. Thanks.

    Read the article

  • ASP.NET Menu: how to build custom menu

    - by Turi
    hi all, i have this situation: i have to build a menu with datasource from the db, because i have to filter them among the user that is loged in. i have tried building a menu using repeater, and all this works fine till i realized that i wanted the menu with more than two levels. i have also tried with treeview, but i really don't want this solution (problem with the style and the treeview doesn't have a pretty view). Could anyone give me just the idea ? thnx in advance.

    Read the article

  • multi_index composite_key replace with iterator

    - by Rohit
    Is there anyway to loop through an index in a boost::multi_index and perform a replace? #include <iostream> #include <string> #include <boost/multi_index_container.hpp> #include <boost/multi_index/composite_key.hpp> #include <boost/multi_index/member.hpp> #include <boost/multi_index/ordered_index.hpp> using namespace boost::multi_index; using namespace std; struct name_record { public: name_record(string given_name_,string family_name_,string other_name_) { given_name=given_name_; family_name=family_name_; other_name=other_name_; } string given_name; string family_name; string other_name; string get_name() const { return given_name + " " + family_name + " " + other_name; } void setnew(string chg) { given_name = given_name + chg; family_name = family_name + chg; } }; struct NameIndex{}; typedef multi_index_container< name_record, indexed_by< ordered_non_unique< tag<NameIndex>, composite_key < name_record, BOOST_MULTI_INDEX_MEMBER(name_record,string, name_record::given_name), BOOST_MULTI_INDEX_MEMBER(name_record,string, name_record::family_name) > > > > name_record_set; typedef boost::multi_index::index<name_record_set,NameIndex>::type::iterator IteratorType; typedef boost::multi_index::index<name_record_set,NameIndex>::type NameIndexType; void printContainer(name_record_set & ns) { cout << endl << "PrintContainer" << endl << "-------------" << endl; IteratorType it1 = ns.begin(); IteratorType it2 = ns.end (); while (it1 != it2) { cout<<it1->get_name()<<endl; it1++; } cout << "--------------" << endl << endl; } void modifyContainer(name_record_set & ns) { cout << endl << "ModifyContainer" << endl << "-------------" << endl; IteratorType it3; IteratorType it4; NameIndexType & idx1 = ns.get<NameIndex>(); IteratorType it1 = idx1.begin(); IteratorType it2 = idx1.end(); while (it1 != it2) { cout<<it1->get_name()<<endl; name_record nr = *it1; nr.setnew("_CHG"); bool res = idx1.replace(it1,nr); cout << "result is: " << res << endl; it1++; } cout << "--------------" << endl << endl; } int main() { name_record_set ns; ns.insert( name_record("Joe","Smith","ENTRY1") ); ns.insert( name_record("Robert","Brown","ENTRY2") ); ns.insert( name_record("Robert","Nightingale","ENTRY3") ); ns.insert( name_record("Marc","Tuxedo","ENTRY4") ); printContainer (ns); modifyContainer (ns); printContainer (ns); return 0; } PrintContainer ------------- Joe Smith ENTRY1 Marc Tuxedo ENTRY4 Robert Brown ENTRY2 Robert Nightingale ENTRY3 -------------- ModifyContainer ------------- Joe Smith ENTRY1 result is: 1 Marc Tuxedo ENTRY4 result is: 1 Robert Brown ENTRY2 result is: 1 -------------- PrintContainer ------------- Joe_CHG Smith_CHG ENTRY1 Marc_CHG Tuxedo_CHG ENTRY4 Robert Nightingale ENTRY3 Robert_CHG Brown_CHG ENTRY2 --------------

    Read the article

  • Announcing: Great Improvements to Windows Azure Web Sites

    - by ScottGu
    I’m excited to announce some great improvements to the Windows Azure Web Sites capability we first introduced earlier this summer.  Today’s improvements include: a new low-cost shared mode scaling option, support for custom domains with shared and reserved mode web-sites using both CNAME and A-Records (the later enabling naked domains), continuous deployment support using both CodePlex and GitHub, and FastCGI extensibility.  All of these improvements are now live in production and available to start using immediately. New “Shared” Scaling Tier Windows Azure allows you to deploy and host up to 10 web-sites in a free, shared/multi-tenant hosting environment. You can start out developing and testing web sites at no cost using this free shared mode, and it supports the ability to run web sites that serve up to 165MB/day of content (5GB/month).  All of the capabilities we introduced in June with this free tier remain the same with today’s update. Starting with today’s release, you can now elastically scale up your web-site beyond this capability using a new low-cost “shared” option (which we are introducing today) as well as using a “reserved instance” option (which we’ve supported since June).  Scaling to either of these modes is easy.  Simply click on the “scale” tab of your web-site within the Windows Azure Portal, choose the scaling option you want to use with it, and then click the “save” button.  Changes take only seconds to apply and do not require any code to be changed, nor the app to be redeployed: Below are some more details on the new “shared” option, as well as the existing “reserved” option: Shared Mode With today’s release we are introducing a new low-cost “shared” scaling mode for Windows Azure Web Sites.  A web-site running in shared mode is deployed in a shared/multi-tenant hosting environment.  Unlike the free tier, though, a web-site in shared mode has no quotas/upper-limit around the amount of bandwidth it can serve.  The first 5 GB/month of bandwidth you serve with a shared web-site is free, and then you pay the standard “pay as you go” Windows Azure outbound bandwidth rate for outbound bandwidth above 5 GB. A web-site running in shared mode also now supports the ability to map multiple custom DNS domain names, using both CNAMEs and A-records, to it.  The new A-record support we are introducing with today’s release provides the ability for you to support “naked domains” with your web-sites (e.g. http://microsoft.com in addition to http://www.microsoft.com).  We will also in the future enable SNI based SSL as a built-in feature with shared mode web-sites (this functionality isn’t supported with today’s release – but will be coming later this year to both the shared and reserved tiers). You pay for a shared mode web-site using the standard “pay as you go” model that we support with other features of Windows Azure (meaning no up-front costs, and you pay only for the hours that the feature is enabled).  A web-site running in shared mode costs only 1.3 cents/hr during the preview (so on average $9.36/month). Reserved Instance Mode In addition to running sites in shared mode, we also support scaling them to run within a reserved instance mode.  When running in reserved instance mode your sites are guaranteed to run isolated within your own Small, Medium or Large VM (meaning no other customers run within it).  You can run any number of web-sites within a VM, and there are no quotas on CPU or memory limits. You can run your sites using either a single reserved instance VM, or scale up to have multiple instances of them (e.g. 2 medium sized VMs, etc).  Scaling up or down is easy – just select the “reserved” instance VM within the “scale” tab of the Windows Azure Portal, choose the VM size you want, the number of instances of it you want to run, and then click save.  Changes take effect in seconds: Unlike shared mode, there is no per-site cost when running in reserved mode.  Instead you pay only for the reserved instance VMs you use – and you can run any number of web-sites you want within them at no extra cost (e.g. you could run a single site within a reserved instance VM or 100 web-sites within it for the same cost).  Reserved instance VMs start at 8 cents/hr for a small reserved VM.  Elastic Scale-up/down Windows Azure Web Sites allows you to scale-up or down your capacity within seconds.  This allows you to deploy a site using the shared mode option to begin with, and then dynamically scale up to the reserved mode option only when you need to – without you having to change any code or redeploy your application. If your site traffic starts to drop off, you can scale back down the number of reserved instances you are using, or scale down to the shared mode tier – all within seconds and without having to change code, redeploy, or adjust DNS mappings.  You can also use the “Dashboard” view within the Windows Azure Portal to easily monitor your site’s load in real-time (it shows not only requests/sec and bandwidth but also stats like CPU and memory usage). Because of Windows Azure’s “pay as you go” pricing model, you only pay for the compute capacity you use in a given hour.  So if your site is running most of the month in shared mode (at 1.3 cents/hr), but there is a weekend when it gets really popular and you decide to scale it up into reserved mode to have it run in your own dedicated VM (at 8 cents/hr), you only have to pay the additional pennies/hr for the hours it is running in the reserved mode.  There is no upfront cost you need to pay to enable this, and once you scale back down to shared mode you return to the 1.3 cents/hr rate.  This makes it super flexible and cost effective. Improved Custom Domain Support Web sites running in either “shared” or “reserved” mode support the ability to associate custom host names to them (e.g. www.mysitename.com).  You can associate multiple custom domains to each Windows Azure Web Site.  With today’s release we are introducing support for A-Records (a big ask by many users). With the A-Record support, you can now associate ‘naked’ domains to your Windows Azure Web Sites – meaning instead of having to use www.mysitename.com you can instead just have mysitename.com (with no sub-name prefix).  Because you can map multiple domains to a single site, you can optionally enable both a www and naked domain for a site (and then use a URL rewrite rule/redirect to avoid SEO problems). We’ve also enhanced the UI for managing custom domains within the Windows Azure Portal as part of today’s release.  Clicking the “Manage Domains” button in the tray at the bottom of the portal now brings up custom UI that makes it easy to manage/configure them: As part of this update we’ve also made it significantly smoother/easier to validate ownership of custom domains, and made it easier to switch existing sites/domains to Windows Azure Web Sites with no downtime. Continuous Deployment Support with Git and CodePlex or GitHub One of the more popular features we released earlier this summer was support for publishing web sites directly to Windows Azure using source control systems like TFS and Git.  This provides a really powerful way to manage your application deployments using source control.  It is really easy to enable this from a website’s dashboard page: The TFS option we shipped earlier this summer provides a very rich continuous deployment solution that enables you to automate builds and run unit tests every time you check in your web-site, and then if they are successful automatically publish to Azure. With today’s release we are expanding our Git support to also enable continuous deployment scenarios and integrate with projects hosted on CodePlex and GitHub.  This support is enabled with all web-sites (including those using the “free” scaling mode). Starting today, when you choose the “Set up Git publishing” link on a website’s “Dashboard” page you’ll see two additional options show up when Git based publishing is enabled for the web-site: You can click on either the “Deploy from my CodePlex project” link or “Deploy from my GitHub project” link to walkthrough a simple workflow to configure a connection between your website and a source repository you host on CodePlex or GitHub.  Once this connection is established, CodePlex or GitHub will automatically notify Windows Azure every time a checkin occurs.  This will then cause Windows Azure to pull the source and compile/deploy the new version of your app automatically.  The below two videos walkthrough how easy this is to enable this workflow and deploy both an initial app and then make a change to it: Enabling Continuous Deployment with Windows Azure Websites and CodePlex (2 minutes) Enabling Continuous Deployment with Windows Azure Websites and GitHub (2 minutes) This approach enables a really clean continuous deployment workflow, and makes it much easier to support a team development environment using Git: Note: today’s release supports establishing connections with public GitHub/CodePlex repositories.  Support for private repositories will be enabled in a few weeks. Support for multiple branches Previously, we only supported deploying from the git ‘master’ branch.  Often, though, developers want to deploy from alternate branches (e.g. a staging or future branch). This is now a supported scenario – both with standalone git based projects, as well as ones linked to CodePlex or GitHub.  This enables a variety of useful scenarios.  For example, you can now have two web-sites - a “live” and “staging” version – both linked to the same repository on CodePlex or GitHub.  You can configure one of the web-sites to always pull whatever is in the master branch, and the other to pull what is in the staging branch.  This enables a really clean way to enable final testing of your site before it goes live. This 1 minute video demonstrates how to configure which branch to use with a web-site. Summary The above features are all now live in production and available to use immediately.  If you don’t already have a Windows Azure account, you can sign-up for a free trial and start using them today.  Visit the Windows Azure Developer Center to learn more about how to build apps with it. We’ll have even more new features and enhancements coming in the weeks ahead – including support for the recent Windows Server 2012 and .NET 4.5 releases (we will enable new web and worker role images with Windows Server 2012 and .NET 4.5 next month).  Keep an eye out on my blog for details as these new features become available. Hope this helps, Scott P.S. In addition to blogging, I am also now using Twitter for quick updates and to share links. Follow me at: twitter.com/scottgu

    Read the article

  • touchesBegan / Ended incorrectly identifying second, third, etc. touch

    - by Rob
    I have an issue where touchesBegan and touchesEnded are incorrectly identifying my second, third, etc touch if I continue to hold down my first touch. If I lift my finger up off the first touch, then it will recognize the next touch just fine. It's only when I hold my first touch down continuously and then try and touch a different area with a different finger at the same time. It will then incorrectly register that second touch as being from the first touch again. Any insights into how I can fix this? - (void)touchesBegan:(NSSet *)touches withEvent:(UIEvent *)event { UITouch* touch = [touches anyObject]; NSString* filename = [listOfStuff objectAtIndex:[touch view].tag]; // do something with the filename now } - (void)touchesEnded:(NSSet *)touches withEvent:(UIEvent *)event { ITouch* touch = [touches anyObject]; NSString* buttonPressed = [listOfStuff objectAtIndex:[touch view].tag]; // do something with this info now }

    Read the article

  • Is it true that quickly closing a webpage opened from a search engine result can hurt the site's ranking?

    - by Austin ''Danger'' Powers
    My web designer recently told me that I need to be careful not to Google for my business' website, click on its search result link, then quickly close the page (or click back) too many times. He says "Google knows" that I didn't stay on the page, and could penalize my site for having a high click-through rate if it happens too much (it was something along those lines- I forget the exact wording). Apparently, it could look like the behavior of a visitor who was not interested in what they found (hence the alleged detrimental effect on the site's search ranking). This sounds hard to believe to me because I would not have thought any information is transmitted which tells Google (or anyone, for that matter) whether or not a website is still open in a browser (in my case Firefox v25.0). Could there possibly be any truth to this? If not, why might he have come to this conclusion? Is there some click-tracking or similar technology employed by search engines which does something similar? Looking forward to hearing everyone's thoughts.

    Read the article

  • Art of Touch : Le nouveau site de Microsoft pour la création et le partage de l'art numérique en HTML5 avec ses souris tactiles

    Art of Touch : Le nouveau site de Microsoft pour la création et le partage de l'art numérique avec ses souris tactiles Microsoft vient de procéder au lancement du programme « Art of Touch », qui permet aux utilisateurs ayant un navigateur compatible HTML5 de créer et partager de l'art numérique. Le projet « Art of Touch » est une campagne marketing pour la famille de souris tactiles (Touch Mouse, Arc Touch Mouse et Explorer Touch Mouse) commercialisées par l'éditeur. Le site dédié au programme permet aux internautes de créer des oeuvres d'arts originales. Ils peuvent utiliser trois pinceaux différents et p...

    Read the article

  • SharePoint 2010 Hosting :: Error Message - Your search cannot be completed because this site is not assigned to an indexer

    - by mbridge
    Error when we are trying to access search in sharepoint site: "Your search cannot be completed because this site is not assigned to an indexer. Contact your administrator for more information." Solving Problem: 1. Go to SharePoint Central Administration > Application Management > Content Databases (Underneath SharePoint Web Application Management). 2. Select the correct SharePoint web application – click on the name of the Content databases  - this will open the  “Manage Content Database Settings” page. 3. Make sure that the Search Server is set on the “Manage Content Database Settings” page. Hope it helps!!

    Read the article

  • Java Meta Search Engine API

    - by Loki
    I'm currently researching Java libraries to help in building a meta type search engine in the sense of being able to replace any given search engine in the back-end of the application or to simultaneously search using multiple search engines. I'm not interested in the GUI part here, just the generalization of search engine APIs and usage. I'd like to know about the common libraries used to achieve this task and if there are any common patterns used in this case. I imagined that this problem is common enough to be able to find plenty of stuff on Google, but it seems like search is a very proprietary domain and not much information is fed back to the community.

    Read the article

  • 3 or 4 monitors with Nvidia and Ubuntu

    - by Jason
    I saw that you are (were?) running 4 monitors with Ubuntu 8.10 and two Nvidia cards (http://stackoverflow.com/questions/27113/how-to-use-3-monitors). I was curious if you were doing this with Xinerama, a hacked up TwinView config, or multiple X screens, or some other method? Does it work with compiz? I intend to run my Dell 30" in the middle with two 1280x1024 on the sides and continue to use one X screen, and run compiz, on Ubuntu 9.04. Currently, I am using 2 monitors with twinview and compiz, which runs fantastic. I just can't get the third monitor running (unless I enable it in its own X screen, and then enable Xinerama to enable windows to be dragged as if all one X screen, but this breaks compiz, and I don't care much for having separate X screen). I am very interested in knowing how you set up 4 monitors with 2 GPU's. Thanks!

    Read the article

  • Execute multiple command lines with the same process using C#

    - by rima
    Hi according to my last question here I try to write a sql Editor or some thing like this,in this way I try to connect to CMD from C# and execute my command. now my problem is for example I connect to SQLPLUS after that I cant get SQLPLUS command,and the other resource I review don't satisfy me.Please help me how after I connected to Sqlplus ,I can a live my process to run my sql command? right now I use this code: //Create process System.Diagnostics.Process pProcess = new System.Diagnostics.Process(); //strCommand is path and file name of command to run pProcess.StartInfo.FileName = strCommand; //strCommandParameters are parameters to pass to program pProcess.StartInfo.Arguments = strCommandParameters; pProcess.StartInfo.UseShellExecute = false; //Set output of program to be written to process output stream pProcess.StartInfo.RedirectStandardOutput = true; //Optional pProcess.StartInfo.WorkingDirectory = strWorkingDirectory; //Start the process pProcess.Start(); //Get program output string strOutput = pProcess.StandardOutput.ReadToEnd(); //Wait for process to finish pProcess.WaitForExit(); but i customize it.I separate the initialize, i mean i just create process object one time,but I still have problem. to run the second command I use these codes for second time calling: pProcess.StartInfo.FileName = strCommand; //strCommandParameters are parameters to pass to program pProcess.StartInfo.Arguments = strCommandParameters; //Start the process pProcess.Start(); //Get program output string strOutput = pProcess.StandardOutput.ReadToEnd(); //Wait for process to finish pProcess.WaitForExit(); Thanks in advance

    Read the article

  • boost::multi_array resize exception?

    - by Glen
    I'm trying to figure out if the boost::multi_array constructor or resize method can throw a bad_alloc exception (or some other exception indicating the allocation or resize failed). I can't find this information in the documentation anywhere.

    Read the article

  • Template with constant expression: error C2975 with VC++2008

    - by Arman
    Hello, I am trying to use elements of meta programming, but hit the wall with the first trial. I would like to have a comparator structure which can be used as following: intersect_by<ID>(L1.data, L2.data, "By ID: "); intersect_by<IDf>(L1.data, L2.data, "By IDf: "); Where: struct ID{};// Tag used for original IDs struct IDf{};// Tag used for the file position //following Boost.MultiIndex examples template<typename Tag,typename MultiIndexContainer> void intersect_by( const MultiIndexContainer& L1,const MultiIndexContainer& L2,std::string msg, Tag* =0 /* fixes a MSVC++ 6.0 bug with implicit template function parms / ) { / obtain a reference to the index tagged by Tag */ const typename boost::multi_index::index<MultiIndexContainer,Tag>::type& L1_ID_index= get<Tag>(L1); const typename boost::multi_index::index<MultiIndexContainer,Tag>::type& L2_ID_index= get<Tag>(L2); std::set_intersection( L1_ID_index.begin(), L1_ID_index.end(), L2_ID_index.begin(), L2_ID_index.end(), std::inserter(s, s.begin()), strComparator() // Here I get the C2975 error ); } template<int N> struct strComparator; template<> struct strComparator<0>{ bool operator () (const particleID& id1, const particleID& id2) const { return id1.ID struct strComparator<1{ bool operator () (const particleID& id1, const particleID& id2) const { return id1.IDf }; What I am missing? kind regards Arman.

    Read the article

  • Multithreading and Interrupts

    - by Nicholas Flynt
    I'm doing some work on the input buffers for my kernel, and I had some questions. On Dual Core machines, I know that more than one "process" can be running simultaneously. What I don't know is how the OS and the individual programs work to protect collisions in data. There are two things I'd like to know on this topic: (1) Where do interrupts occur? Are they guaranteed to occur on one core and not the other, and could this be used to make sure that real-time operations on one core were not interrupted by, say, file IO which could be handled on the other core? (I'd logically assume that the interrupts would happen on the 1st core, but is that always true, and how would you tell? Or perhaps does each core have its own settings for interrupts? Wouldn't that lead to a scenario where each core could react simultaneously to the same interrupt, possibly in different ways?) (2) How does the dual core processor handle opcode memory collision? If one core is reading an address in memory at exactly the same time that another core is writing to that same address in memory, what happens? Is an exception thrown, or is a value read? (I'd assume the write would work either way.) If a value is read, is it guaranteed to be either the old or new value at the time of the collision? I understand that programs should ideally be written to avoid these kinds of complications, but the OS certainly can't expect that, and will need to be able to handle such events without choking on itself.

    Read the article

  • Linux's thread local storage implementation

    - by anon
    __thread Foo foo; How is "foo" actually resolved? Does the compiler silently replace every instance of "foo" with a function call? Is "foo" stored somewhere relative to the bottom of the stack, and the compiler stores this as "hey, for each thread, have this space near the bottom of the stack, and foo is stored as 'offset x from bottom of stack'"?

    Read the article

  • Using Jquery.Form Plugin + MultiFile to automatically upload a single file

    - by Alan Neal
    I wanted to find a way to upload a single file*, in the background, have it start automatically after file selection, and not require a flash uploader, so I am trying to use two great mechanisms (jQuery.Form and JQuery MultiFile) together. I haven't succeeded, but I'm pretty sure it's because I'm missing something fundamental. Just using MultiFile, I define the form as follows... <form id="photoForm" action="image.php" method="post" enctype="multipart/form-data"> The file input button is defined as... <input id="photoButton" "name="sourceFile" class="photoButton max-1 accept-jpg" type="file"> And the Javascript is... $('#photoButton').MultiFile({ afterFileSelect: function(){ document.getElementById("photoForm").submit(); } }); This works perfectly. As soon as the user selects a single file, MultiFile submits the form to the server. If instead of using MultiFile, as shown above, let's say I include a Submit button along with the JQuery Form plugin defined as follows... var options = { success: respondToUpload }; $('#photoForm').ajaxForm(options); ... this also works perfectly. When the Submit button is clicked, the form is uploaded in the background. What I don't know how to do is get these two to work together. If I use Javascript to submit the form (as shown in the MultiFile example above), the form is submitted but the JQuery.Form function is not called, so the form does not get submitted in the background. I thought that maybe I needed to change the form registration as follows... $('#photoForm').submit(function() { $('#photoForm').ajaxForm(options); }); ...but that didn't solve the problem. The same is true when I tried .ajaxSubmit instead of .ajaxForm. What am I missing? BTW: I know it might sound strange to use MultiFile for single-file uploads, but the idea is that the number of files will be dynamic based on the user's account. So, I'm starting with one but the number changes depending on conditions.

    Read the article

  • How to implement "tab drop in window" and "tab drop out of window" effect

    - by stanleyxu2005
    How to implement "tab drop in window" and "tab drop out of window" effect, just like the google chrome browser? My imagination is that: It is not a single application. The main browser frame is an application, the tabs are several applications without a window frame. When the main browser frame is resized or moved, all tabs will be notified to be resized or moved. Is there any existing component or code example can do this?

    Read the article

  • Should I block bots from my site and why?

    - by Frank E
    My logs are full of bot visitors, often from Eastern Europe and China. The bots are identified as Ahrefs, Seznam, LSSRocketCrawler, Yandex, Sogou and so on. Should I block these bots from my site and why? Which ones have a legitimate purpose in increasing traffic to my site? Many of them are SEO. I have to say I see less traffic if anything since the bots have arrived in large numbers. It would not be too hard to block these since they all admit in their User Agent that they are bots.

    Read the article

  • Neural Network Always Produces Same/Similar Outputs for Any Input

    - by l33tnerd
    I have a problem where I am trying to create a neural network for Tic-Tac-Toe. However, for some reason, training the neural network causes it to produce nearly the same output for any given input. I did take a look at Artificial neural networks benchmark, but my network implementation is built for neurons with the same activation function for each neuron, i.e. no constant neurons. To make sure the problem wasn't just due to my choice of training set (1218 board states and moves generated by a genetic algorithm), I tried to train the network to reproduce XOR. The logistic activation function was used. Instead of using the derivative, I multiplied the error by output*(1-output) as some sources suggested that this was equivalent to using the derivative. I can put the Haskell source on HPaste, but it's a little embarrassing to look at. The network has 3 layers: the first layer has 2 inputs and 4 outputs, the second has 4 inputs and 1 output, and the third has 1 output. Increasing to 4 neurons in the second layer didn't help, and neither did increasing to 8 outputs in the first layer. I then calculated errors, network output, bias updates, and the weight updates by hand based on http://hebb.mit.edu/courses/9.641/2002/lectures/lecture04.pdf to make sure there wasn't an error in those parts of the code (there wasn't, but I will probably do it again just to make sure). Because I am using batch training, I did not multiply by x in equation (4) there. I am adding the weight change, though http://www.faqs.org/faqs/ai-faq/neural-nets/part2/section-2.html suggests to subtract it instead. The problem persisted, even in this simplified network. For example, these are the results after 500 epochs of batch training and of incremental training. Input |Target|Output (Batch) |Output(Incremental) [1.0,1.0]|[0.0] |[0.5003781562785173]|[0.5009731800870864] [1.0,0.0]|[1.0] |[0.5003740346965251]|[0.5006347214672715] [0.0,1.0]|[1.0] |[0.5003734471544522]|[0.500589332376345] [0.0,0.0]|[0.0] |[0.5003674110937019]|[0.500095157458231] Subtracting instead of adding produces the same problem, except everything is 0.99 something instead of 0.50 something. 5000 epochs produces the same result, except the batch-trained network returns exactly 0.5 for each case. (Heck, even 10,000 epochs didn't work for batch training.) Is there anything in general that could produce this behavior? Also, I looked at the intermediate errors for incremental training, and the although the inputs of the hidden/input layers varied, the error for the output neuron was always +/-0.12. For batch training, the errors were increasing, but extremely slowly and the errors were all extremely small (x10^-7). Different initial random weights and biases made no difference, either. Note that this is a school project, so hints/guides would be more helpful. Although reinventing the wheel and making my own network (in a language I don't know well!) was a horrible idea, I felt it would be more appropriate for a school project (so I know what's going on...in theory, at least. There doesn't seem to be a computer science teacher at my school). EDIT: Two layers, an input layer of 2 inputs to 8 outputs, and an output layer of 8 inputs to 1 output, produces much the same results: 0.5+/-0.2 (or so) for each training case. I'm also playing around with pyBrain, seeing if any network structure there will work. Edit 2: I am using a learning rate of 0.1. Sorry for forgetting about that. Edit 3: Pybrain's "trainUntilConvergence" doesn't get me a fully trained network, either, but 20000 epochs does, with 16 neurons in the hidden layer. 10000 epochs and 4 neurons, not so much, but close. So, in Haskell, with the input layer having 2 inputs & 2 outputs, hidden layer with 2 inputs and 8 outputs, and output layer with 8 inputs and 1 output...I get the same problem with 10000 epochs. And with 20000 epochs. Edit 4: I ran the network by hand again based on the MIT PDF above, and the values match, so the code should be correct unless I am misunderstanding those equations. Some of my source code is at http://hpaste.org/42453/neural_network__not_working; I'm working on cleaning my code somewhat and putting it in a Github (rather than a private Bitbucket) repository. All of the relevant source code is now at https://github.com/l33tnerd/hsann.

    Read the article

  • Multiple NT service owned by the same program in Delphi

    - by Claudio
    I'm looking for Delphi sample code to develope a Win32 Windows service which can be installed many times (with different Name). The idea is to have 1 exe and 1 registry key with 1 subkey for every service to be installed. I use the exe to install/run many service, every service take his parameter from his registry subkey. Does anyone have a sample code? Many thanks Claudio

    Read the article

  • Why is it so hard to find anything on MS site?

    - by Amir Rezaei
    I have always had this question in my mind and I would really be happy to get an explanation for this. Is it only me or do you also feel the same way that it's hard to find anything on MS site. For example, every time I need to download .NET framework I have to Google it. You never know what you can download, no category for downloads. You are simply left to a search field. You never know if you downloaded the latest version of the file. The tragically true is that you have to rely on their competitor Google to find anything on their site. I know that they are a big company. But is it really that hard to have an organized way to publish information?

    Read the article

  • How can I identify unknown query string fragments that are coming to my site?

    - by Jon
    In the Google Analytics content overview for a site that I work on, the home page is getting many pageviews with some unfamiliar query string fragments, example: /?jkId=1234567890abcdef1234567890abcdef&jt=1&jadid=1234567890&js=1&jk=key words&jsid=12345&jmt=1 (potentially identifiable IDs have been changed) It clearly looks like some kind of ad tracking info, but noone who works on the site knows where it comes from, and I haven't been able to find any useful information from searching. Is there some listing of common query string keys available anywhere? Alternatively, does anyone happen to know where these keys (jkId, jt, jadid, js, jk, jsid and jmt) might come from?

    Read the article

  • Performance degrades for more than 2 threads on Xeon X5355

    - by zoolii
    Hi All, I am writing an application using boost threads and using boost barriers to synchronize the threads. I have two machines to test the application. Machine 1 is a core2 duo (T8300) cpu machine (windows XP professional - 4GB RAM) where I am getting following performance figures : Number of threads :1 , TPS :21 Number of threads :2 , TPS :35 (66 % improvement) further increase in number of threads decreases the TPS but that is understandable as the machine has only two cores. Machine 2 is a 2 quad core ( Xeon X5355) cpu machine (windows 2003 server with 4GB RAM) and has 8 effective cores. Number of threads :1 , TPS :21 Number of threads :2 , TPS :27 (28 % improvement) Number of threads :4 , TPS :25 Number of threads :8 , TPS :24 As you can see, performance is degrading after 2 threads (though it has 8 cores). If the program has some bottle neck , then for 2 thread also it should have degraded. Any idea? , Explanations ? , Does the OS has some role in performance ? - It seems like the Core2duo (2.4GHz) scales better than Xeon X5355 (2.66GHz) though it has better clock speed. Thank you -Zoolii

    Read the article

  • VB.net Reading from ListViews with Multiple Columns

    - by xzerox
    Alright. So I was able to find out how to read from the first column but I need to read from both of them. I am using full row select which I need on there. Here is the code I am using to get it for the first column. Dim I As Integer For I = 0 To ListView1.SelectedItems.Count - 1 MsgBox(ListView1.SelectedItems(I).Text) Next MessageBox.Show(ListView1.Columns(0).Text)

    Read the article

  • Are there any good reasons to intentionally serve a new web site in Quirks mode?

    - by wsanville
    I was a little surprised that Amazon's site doesn't specify a doctype, and is rendered in quirks mode. What could possibly be the reason for this? I understand what quirks mode is and why doctypes were introduced, but I can't understand why this would be intentionally left off. I guess it might simplify markup if they're trying to support ancient browsers, but isn't that like shooting yourself in the foot when it comes to modern browsers, especially when their site is so Javascript rich? Does this level the playing field when it comes to supporting really old browsers? Is there something else I'm missing?

    Read the article

< Previous Page | 126 127 128 129 130 131 132 133 134 135 136 137  | Next Page >