url
stringlengths 13
4.35k
| tag
stringclasses 1
value | text
stringlengths 109
628k
| file_path
stringlengths 109
155
| dump
stringclasses 96
values | file_size_in_byte
int64 112
630k
| line_count
int64 1
3.76k
|
|---|---|---|---|---|---|---|
https://i.publiclab.org/wiki/leaflet-environmental-layers
|
code
|
Leaflet Environmental Layers
Leaflet Environmental Layers is a plugin for the Leaflet web mapping library, which makes it possible to view many different sources of environmental data on a single unified map.
The code on GitHub, including how to install it: https://github.com/publiclab/leaflet-environmental-layers/
There are so many different sources of environmental data, and it's not always easy to relate or contrast them -- how does the government hosted Toxics Release Inventory stack up against data from air sensors, or relate to wind patterns? How do alerts on SkyTruth relate to demographic information like income or race and ethnicity?
We hope this interface will help make this kind of comparison/cross-referencing easier.
- Purple Air sensor map
- MapKnitter.org maps
- Toxic Release Inventory (TRI) data
- Frac Tracker data
- Water Reporter data
- OpenInfraMap - petroleum, water, power infrastructure
- Weather/wind data
- JusticeMap demographic (race/ethnicity, income) data
- OdorLog.com community odor reporting data
- SkyTruth Alerts environmental issue data
To request a new layer, please ask a question below!
Existing layer requests are here: https://github.com/publiclab/leaflet-environmental-layers/labels/new-layer
|What do we know about EPA National Emissions Inventory (NEI) data?||@warren||over 4 years ago||1||0|
|What more data-layers can we show on map ?||@sagarpreet||almost 5 years ago||1||19|
|Nothing yet on the topic "leaflet-environmental-layers" -- be the first to post something!|
Activities should include a materials list, costs and a step-by-step guide to construction with photos. Learn what makes a good activity here.
|GSoC 2021: Geographic Features Refinement - Final Report||@barun1024||almost 2 years ago||5||4|
|Introducing new inline LEL power tags||@sagarpreet||over 3 years ago||3||2|
|New Features in Leaflet-Environmental Layers Library||@sagarpreet||over 4 years ago||3||5|
|Leaflet Environmental Layer Library - Part 2||@sagarpreet||almost 5 years ago||2||8|
|Leaflet Environmental Layers Library||@sagarpreet||about 5 years ago||9||21|
|
s3://commoncrawl/crawl-data/CC-MAIN-2023-23/segments/1685224653631.71/warc/CC-MAIN-20230607074914-20230607104914-00598.warc.gz
|
CC-MAIN-2023-23
| 2,095 | 26 |
https://politics.meta.stackexchange.com/questions/6662/deleting-comment-violating-se-policies
|
code
|
This concerns the comments following my response to question Does Hamas have any leverage against Israel (other than the hostages)? I flagged a comment that referred to some of my assertion as ludicrous, and moderators removed the comment and the subsequent discussion. There is however another comment which makes the same point, but in a stronger language, which the moderators have probably simply missed:
Downvoted for the idiotic rants about "over-mediatization" - not only a value judgement on how much coverage you think an ongoing massacre should attract, but qualified without any evidence whatsoever for "hardly made a topic". You must have a good memory if you can remember off the top of your head each news story that broke on each of those conflicts. Not to mention this current "war" really isn't one so much as it is a massacre borne out of revenge, which the casualty figures cited are evidence of.
(emphasis is mine)
Just wanted to attract attention of the moderators. Please abstain of using the Meta for discussing validity of either the answer or the comment.
|
s3://commoncrawl/crawl-data/CC-MAIN-2024-18/segments/1712296817112.71/warc/CC-MAIN-20240416222403-20240417012403-00070.warc.gz
|
CC-MAIN-2024-18
| 1,080 | 4 |
http://bkhome.org/blog/?viewDetailed=00841
|
code
|
This morning, Phil, the Economic Development Officer for the Shire of Perenjori, saw me carrying my laptop into the Telecentre, and enquired into the status of my application for a satellite connection.
I explained that there seems to be some problem with the Federal Government subsidy approval, and Skymesh had ignored my last two email enquiries. Phil said leave it to me, and a little while later he popped his head into the computer room and told me it was all fixed. Phil deals with govt people all the time, knows how to untangle red tape. The approval will be in the mail to me and I have to forward it to Skymesh.
An incredibly helpful guy. He knows all about the Puppy project and wants to help out in whatever way possible.
Today I also posted a cheque for 3 months dialup Internet with Netbay Internet. Incredibly cheap, see my recent post. If it works okay, I'll probably sign up for a year, as I have realised that I should not abandon dialup just because I have broadband satellite available. I do need to test Puppy with dialup. Rerwin is carrying the ball here, helping Puppy users to get their dialup working and also now going into 3G modems. Rerwin sends me patches, but I really do need to be able to test after I have applied the patches, as a confirmation that I have implemented it right.
For those of you who are not Aussies, you will not know about the subsidy that I have mentioned above. Before the great economic catastrophy that has struck the world (and just after, in an attempt to stave off a recession), the Aussie govt was splashing money around, subsidising this and that. One of the things that the rural politicians managed to wrangle was a subsidy known as the Australian Broadband Guarantee, intended to enable anyone in rural Australia to have broadband Internet access at a comparable price to city people. Where broadband is not available via the telephone line, or a local wireless network, the govt provides an up-front payment for a satellite installation.
I do have such an installation, but with technology moving on apparently to get reconnected I need to upgrade my transceiver unit and the dish. That's where the red tape came in, and that's what Phil has fixed.
The Federal Govt now finds itself in the red, billions of dollars in the red, so is desperately trying to rein in expenditure. There's one subsidy that I would have liked to avail myself of, which expires on 30th of June 2009, that's their photovoltaic system installation subsidy -- another up-front payment that covers most of the cost. Vendors were offering 1Kw installations, grid connected, the whole thing including installation, for about $2500 after the subsidy.
But, I didn't know what vendor would install in my rural area, nor do I have the money, so let it go.
Comments:Posted on 26 Jun 2009, 14:04 by Lobster
Thanks to Phil
It used to upset that a developed nation could not provide broadband. Many thanks to Phil. Bravo. I hope he will continue to support Perenjori, Ozland and Barry as Puppy moves towards top 5 distribution status and open source becomes the modus oparandi for governments and citizens worldwide. Maybe the Telecentre and other local facilities will become Puppy powered? Being modem friendly is also a great Puppy strength. Good news.
|
s3://commoncrawl/crawl-data/CC-MAIN-2015-48/segments/1448398462686.42/warc/CC-MAIN-20151124205422-00026-ip-10-71-132-137.ec2.internal.warc.gz
|
CC-MAIN-2015-48
| 3,282 | 11 |
https://www.physicsforums.com/threads/syntax-in-t-sql.868092/
|
code
|
Hi, I wonder if someone remembers their T-SQL. We have this statement: Just curious: I remember using SET where SELECT appears. Is this SELECT used in order to assign values to many variables simultaneously? Also, Is there a reason to use two separate DECLARE statements , or could we just use a single one? EDIT : I am also confused about the OBJECT_ID function. I tried it on my SQL Server and got an error message when entering the name of the database I was using, i.e., OBJECT_ID( 'Database Name') generated an error message. Thanks.
|
s3://commoncrawl/crawl-data/CC-MAIN-2018-22/segments/1526794867092.48/warc/CC-MAIN-20180525121739-20180525141739-00057.warc.gz
|
CC-MAIN-2018-22
| 538 | 1 |
http://docs.afterworks.com/FumeFX5maya/Tutorial%2007%20Effectors.htm
|
code
|
Tutorial 07: Effectors
Effectors allow a user to control almost any parameter on a per - voxel basis. For example, by using Effectors it is possible to have localized gravity, or even Vorticity that depends on the voxel velocity. A Master Effector can directly control amount of smoke, fire, fuel, temperature and color in a voxel.
In this tutorial, we will teach you how to use Effectors.
In Maya, select File - >Open, and from your [MAYA]/FumeFX/Tutorials folder, select the file Tut_07_start.ma.
This is a simple scene and contains a FumeFX grid, one Simple Source, an Object Source and one Effector. We will use Effectors to create smoke based on the velocity in the grid.
Open FumeFX relationship manager and select Simple source and Object Source.
Select the ffxSimpleSource1 and disable Fuel and Smoke channel.
Create an FumeFX Effector and position it directly above ffxSimpleSource1. Set effector shape to Sphere; its effect will now be localized only to voxels within the spherical shape. Set Diameter to 11 and Translate y coordinate to 30.
Each Effector has a Channel Name parameter which is used to select the FumeFX parameter(s) you want to control. Set the Effector's Channel name to "eff1".
Select the FumeFX Grid, open the relationship manager, and select ffxEffectorShape1.
Scroll down to Simulation - >Smoke rollout. You'll notice that Effectors can be applyed not only to FumeFX parameters, but to FumeFX channels as well. Since we want to create smoke based on the velocity in the grid, that's exactly what we need. Under the Smoke rollout, click on the Master Effector icon - Master Effectors are used to assign effectors to channels. The Effector Connect dialog will open. Type "eff1" into the field, which will assign the entered channel name to the selected parameter (in this case, the smoke channel) and press OK button to close the dialog.
Select ffxEffectorShape1 and under the Input rollout, change Type to Velocity. Our Effector's output will now be controlled by the Velocity field. Set Range Min. to 2.0 and Max. to 3.0. This means that velocities between 2.0 and 3.0 will be converted to 0.0 - 1.0 range. The resulting value is passed to the Output module.
For now we will have the Output Value curve disabled, but we will use it later to perform additional tweaks of the Output value (and even invert it).
Under Output rollout, set Output Operator to "Add" which means that the Effector will add to the Smoke channel.
As this Effector will be used to add smoke to the grid, the maximum amount that will be added is defined by the Scalar parameter; set it to 0.1. A higher value will add more smoke to the grid.
Let's analyse how our effector will be applied. With it's Input Type set to Velocity, channel name connected to the Smoke Master Effector and it's shape set to sphere, it will read velocity values from the voxels located inside the sphere and convert them (based on current Range Min/Max values) to 0.0 - 1.0 range. Finally, our scalar valueof 0.1 will adjust that 0.0 - 1.0 range to 0.0 - 0.1. This new range will be used to add to the smoke channel.
Now create another effector and position it above pSphere01. Open FumeFX attribute editor and from its relationship manager select ffxEfector2.
Select the ffxEffector2 node and set the Effector's Shape to Box and set it's Width/ Length/Heightto 22/15/44.X coordinate should be around 30. Set Channel name to "eff2", Input Type to Velocity, Output Type to Add and Scalar to 0.1, as before. Since Object Source is emitting only temperature, velocities will result only from buoyancy and will achieve much lower values. Set Range Min/Max to 0.3 - 1.0.
Go to the Simulation - >Smoke rollout as before. Open the Smoke Master Effector Connect dialog and click on the "..." button; all channels of all effectors added to the selected FumeFX Grid will be listed here. Click on "eff2"This will append ???eff2" to the list of channels that control this parameter.
Before you start simulating this scene, make sure to adjust your output paths.
Once you start simulation, you will notice that the Effectors are creating smoke based on the velocities in the grid. For our first effector, a smoke value of 0.1 will be added to voxels that have velocity of 3.0 and higher, while in the areas of velocity below 2.0 no smoke will be added.
But what if you want to add the smoke the other way around?
Fortunately, there is a quick and effective way to do that by using Output Value curve.
You need to keep in mind that input is on horizontal axis and output on vertical axis. To invert how smoke is added all we need to do is to enable the Use Output Value and invert the curve like on the image below.
You can read it in the following way. When input is 0.0, output will be maximum (0.1 in our case) and when input is 1.0, output will be 0.0.
When you run Simulation it will become obvious that most of the smoke is now added to areas with low velocity, the opposite of what we had in the first step.
|
s3://commoncrawl/crawl-data/CC-MAIN-2023-23/segments/1685224654606.93/warc/CC-MAIN-20230608071820-20230608101820-00144.warc.gz
|
CC-MAIN-2023-23
| 4,981 | 26 |
https://flowfuse.com/blog/2023/04/securing-node-red-in-production/
|
code
|
Step-by-step guide for securing your Node-RED projects.
Node-RED is very easy to get up and running. Whether you run it locally, in Docker, on a Raspberry Pi, or on a service such as FlowFuse Cloud you can have a project up and running in minutes.
One thing that can get overlooked is the security of Node-RED. From personal experience, the first few times I installed Node-RED I was more focussed on the possibilities of what I could do with this new tool than I was keeping my projects secure.
In this article I’m going to look at some easy ways to make your Node-RED project more secure, even when first learning about it in a hobby environment.
# Protecting access from your LAN to Node-RED
Once you have an instance of Node-RED running it can usually be accessed from anywhere on your LAN (local area network) by pointing a web browser to the relevant IP address and port.
With a URL similar to the one above, depending on your specific network and Node-RED configuration, anyone on your LAN can view but more importantly edit your flows. This can be really useful when you are first learning about Node-RED but it’s always a good idea to get into the habit of locking down access to the editor, even if you trust everyone who can access your LAN.
One of the easiest ways to protect your flows is to add a username and password to your Node-RED instance.
The first step is to find your Node-RED settings.js file. It's not always in the same place but on a default Debian Linux installation it can be found in this directory.
If you list that directory you should now see something like this:
We now need to edit settings.js, I'm going to use my favourite text editor, Nano to do that.
We now need to find and edit the following section of the settings file:
For this example, I'm going to add a password and uncomment the relevant section of the settings file, you could also change the username for additional security.
To create the password we'll need to use a command line tool which is included in Node-RED. Open a second terminal then run this command:
node-red admin hash-pw
Put in your new password, I'll use the password 'flowforge' in this example. The tool returns your password in a hashed format:
We can now return to the other terminal window, uncomment the section then paste in the new password, this is how it looks for me:
We can now save and exit out of the settings file.
The last step is to restart Node-RED, I'm using Debian so the command is:
Now, when we try to access Node-RED I will need to provide a username and password.
You might also want to consider turning off the editor interface once you are happy with your flows. This can make it a little harder to make changes to your project but it also gives you peace of mind that nobody has accidentally or deliberately changed your flows. You can turn off the editor interface as follows.
Edit your settings.js file as explained above, look for the following section:
All you need to do is uncomment the bottom line then change the value from false to true, once done it should look something like this:
Now restart Node-RED as covered above, then try accessing your Node-RED instance again. You will no longer be able to edit or view your flows.
Using these two features, we now have much better control over who can access the design interface for Node-RED.
# Traffic to your Node-RED instance is unencrypted
Hopefully, we all know the importance of encrypting your connections between devices to stop people intercepting your traffic. This isn't a huge concern when working on your home LAN but what if you want to access your Node-RED instance from a remote location?
There are two obvious options, HTTPS, and a VPN (Virtual Private Network).
We could setup your Node-RED traffic to run over HTTPS, this solution ensures that all traffic to and from your Node-RED is encrypted. The downside to this approach is it's quite complex to set up. We will need to have a domain name, open up ports on our LAN's firewall, use a HTTPS certificate provider and then make sure we remember to renew the certificates as needed. It's doable if you are comfortable with those concepts (I covered how to do this as part of my blog hosting FlowFuse on Google Cloud) but there is an easier way to get started, using a VPN.
A VPN provides a lot of security advantages depending on which you are using and how it is configured. To secure my traffic I'm going to use a great service call Tailscale which is free for personal projects.
I'm going to install Tailscale on the Raspberry Pi I'm running Node-RED on as well as any other devices I want to access my project from. Once that's done I can access Node-RED from anywhere with internet access but more importantly the traffic to and from my devices is also encrypted.
Before we start, it's important to remember that a VPN is only as secure as the company who runs it. You should always consider if you trust the VPN provider as they could potentially access your devices. I trust Tailscale but please do your own research before using a VPN provider.
The first step we need to take is creating a Tailscale account, you can sign up for free here. We next need to add our devices to our VPN using their software, I'm installing Tailscale on my Apple laptop, Google phone as well as the Raspberry Pi I'm running Node-RED on.
The install process is really easy, even on the Pi running Raspbian the steps you need to take are well explained in the Tailscale docs.
For the Pi, these are the commands we need to run.
- Add Tailscale to the Apt package manager.
curl -fsSL https://pkgs.tailscale.com/stable/debian/bullseye.noarmor.gpg | sudo tee /usr/share/keyrings/tailscale-archive-keyring.gpg >/dev/null
curl -fsSL https://pkgs.tailscale.com/stable/debian/bullseye.tailscale-keyring.list | sudo tee /etc/apt/sources.list.d/tailscale.list
sudo apt-get update
- Install Tailscale
sudo apt-get install tailscale
- Start Tailscale and connect your device
sudo tailscale up
After running the last command, we need to follow the on screen prompts to link our devices to your VPN.
One last thing which you might want to consider doing, every few months you will need to reconnect your devices to your VPN, if you are only going to be accessing your Node-RED device over the VPN you should consider disabling your Tailscale key expiry.
OK, now we've got our devices all connected you should see something like this in the Tailscale dashboard.
I can now access Node-RED on my Pi from my laptop and phone by pointing a browser to the correct IP address (as shown in the image above) with the port for Node-RED:
I’ve now secured all traffic between my devices and Node-RED project, I can access Node-RED from anywhere on the internet.
If you follow these steps you should be on the right path to running a more secure Node-RED instance. There is a lot more you can do and I recommend you read the relevant docs on the Node-RED website to gain some more ideas.
# What about hosting Node-RED on a cloud solution such as FlowFuse?
In this article, I've focussed on hosting Node-RED on a Pi on your own LAN but if you use FlowFuse Cloud to host Node-RED the solutions discussed above are either ready out of the box or are not needed.
By default, the editor for Node-RED is protected using your FlowFuse user credentials. You can also use SSO to further protect your user accounts and give access to Node-RED to your team members.
All traffic to FlowFuse and your Node-RED instances is protected by HTTPS. FlowFuse has set up the domain name and manages the certificates so you can spend time on your flows rather than configuring security.
FlowFuse has a free trial if you'd like to see how we've made secure hosting of Node-RED easy.
How ever you host Node-RED, it's a great idea to get into good security practices as early as possible to ensure that no unsecured Node-RED instances are exposed to the internet. I hope some of the tips above help you get started down the path to creating more secure Node-RED projects.
- Storing Data: Getting Started with Node-RED
- Capture Data from edge devices with Node-RED
- Install the FlowFuse Edge Agent on the Raspberry Pi 5
- Custom Vuetify components for Dashboard 2.0
- Charting REST API Data in a Dashboard
|
s3://commoncrawl/crawl-data/CC-MAIN-2024-18/segments/1712296819089.82/warc/CC-MAIN-20240424080812-20240424110812-00111.warc.gz
|
CC-MAIN-2024-18
| 8,264 | 60 |
http://dmx.sourceforge.net/html/dmxmap_8c.html
|
code
|
dmxmap.c File Reference
This file implements a mapping from remote XInput event types to Xdmx XInput event types.
The exglobals.h file defines global server-side variables with names Device* to be integers that hold the value of the type of the server-side XInput extension event.
The client-side X11/extensions/XInput.h file defines macros with THE EXACT SAME Device* names!
Using those macros to extract remote server event type values from the (opaque) XDevice structure is appropriate, but makes a direct mapping to the Device* integers impossible. So we use the normalized XI_Device* names for these routines.
Create a mapping from remoteEvent to serverEvent. The remoteEvent is the type returned from the remote server. The serverEvent is from the XI_* list of events in include/extensions/XIproto.h.
Lookup a mapping for remoteEvent. The remoteEvent is the type returned from the remote server. The return value is that which was passed into dmxMapInsert (i.e., a value from the XI_* list in include/extensions/XIproto.h). If a mapping is not available, -1 is returned.
Generated June 13, 2004 for Distributed Multihead X by
|
s3://commoncrawl/crawl-data/CC-MAIN-2018-05/segments/1516084890991.69/warc/CC-MAIN-20180122034327-20180122054327-00316.warc.gz
|
CC-MAIN-2018-05
| 1,131 | 8 |
https://wpfilebase.com/ideabase/page/3/?comments-4179
|
code
|
PayPal links broken
I am wanting to renew my license, however the ”buy now” button results in an error from PayPal stating:
Things don’t appear to be working at the moment. Please try again later. Of course, I can’t create a new support ticket since my license has expired. :/
chris (ccsanford at gmail)
Sharepoint Cloud Sync
Feature Request: the ability to sync with Office 365 \ SharePoint.
Notify users you are out of business
When you are no longer answering bulletin boards and support tickets remain unanswered with no response for three weeks, and prior support tickets removed, one assumes that you are out of business.
You should not be accepting payments from PayPal or other source if you do not intend to provide support that is paid for. In most countries, this is known as fraud
Amazon Prime Video – www.amazon.com/mytv enter code
Amazon Prime is a paid subscription to https://www.amazonmytvcodes.com/, that gives you access to a large range of solutions: free speedy shipping, infinite video streaming, exclusive access to deals, amongst other people. . It helps its readers unlimited streaming of movies and TV shows.
Delete Link for Front End
Allows users to delte files from the front end (without the context menu)
how can I limit a file exentension to for example: mp3 only
How can force the plugin to limit a file exentension to for example mp3 only?
As far as I have seen in the seeting it is only possible to exclude some extensions.
Woudn’t it make more sense the other way round?
Thanks for your help in advanced,
|
s3://commoncrawl/crawl-data/CC-MAIN-2023-40/segments/1695233508959.20/warc/CC-MAIN-20230925083430-20230925113430-00028.warc.gz
|
CC-MAIN-2023-40
| 1,551 | 18 |
http://linux.softpedia.com/get/GAMES-ENTERTAINMENT/Puzzle/Circuits-Fischer-53069.shtml
|
code
|
Circuits is a PyGame port of a game wrote for Android phones, by the same author.
Create closed circuits by using random tiles. Before placing them with the mouse. You can rotate them using the mouse wheel, the cursor keys (left/right) or S and D
There are three difficulty levels, three different gameplay mode (which require different tactics), there are also eleven awards the player can earn for completing different tasks. Those range from the very easy (make more than 500 points) to the incredibly difficult (”make more than 2000 points” or “make more than 600 points while using less than 3 seconds per tile”).
What's New in This Release: [ read full changelog ]
· Fixed a scrolling/grabbing bug in the Award window All Awards should now be working (sorry about circuits_1.0.4.tar.gz which actually lacked the game.)
|
s3://commoncrawl/crawl-data/CC-MAIN-2013-20/segments/1368702718570/warc/CC-MAIN-20130516111158-00074-ip-10-60-113-184.ec2.internal.warc.gz
|
CC-MAIN-2013-20
| 833 | 5 |
https://sourceforge.net/p/yii/news/2009/11/yii-php-framework-11-beta-is-released/
|
code
|
We are very pleased to announce the beta release of Yii Framework v1.1! You can download this new release from the Yii download site:
In this beta release, we fixed several minor bugs that were found in the previous alpha release. We also included some minor feature enhancements. In particular, we added support for using table prefix in Yii DAO and ActiveRecord. We refactored the sorting support so that it is more flexible. We also added support for using customized locale data when developing multi-language applications.
Besides the changes made in the core framework, we also made significiant progress in the zii project (the official extension library for Yii). We have completed the encapsulation of the jQuery UI (JUI) library and introduced a full set of widgets to facilitate using JUI plugins.
For a list of all features included in this release, please visit:
Do not use this release for production use. This is still a beta release and thus may be subject to changes in future releases without prior notice.
Thank you for your support!
The Yii Developer Team
ABOUT YII FRAMEWORK
Yii is a high-performance component-based PHP framework best suited for developing large-scale Web 2.0 applications. Yii comes with a full stack of features, including MVC, DAO/ActiveRecord, I18N/L10N, caching, jQuery-based AJAX support, authentication and role-based access control (RBAC), scaffolding, input validation, widgets, events, theming, XSS protection, Web services, and so on. With its thorough documentation and a friendly user community, Yii is easy to learn and can help you cut down your application development time significantly.
Log in to post a comment.
|
s3://commoncrawl/crawl-data/CC-MAIN-2017-34/segments/1502886106358.80/warc/CC-MAIN-20170820073631-20170820093631-00497.warc.gz
|
CC-MAIN-2017-34
| 1,669 | 10 |
https://www.goheritagerun.com/heritage-run-routes-city/
|
code
|
Do you run in your city? Have you visited the heritage sites in your city?
Here’s a chance to discover the heritage in your own and get rewarded for it too :).
Email us your favorite heritage run routes – in your city or anywhere else in the world. Explain with a brief note and list the heritage sites along the route, the best time of the day to plan a solo run there. Create the route on a map (mapometer.com / gmap-pedometer.com are easy tools you can use).
All complete entries will receive exciting Go Heritage Run goodies!
|
s3://commoncrawl/crawl-data/CC-MAIN-2024-18/segments/1712296818067.32/warc/CC-MAIN-20240421225303-20240422015303-00466.warc.gz
|
CC-MAIN-2024-18
| 533 | 4 |
https://lists.gnu.org/archive/html/fluid-dev/2011-09/msg00111.html
|
code
|
I have responded to your individual questions below:
On 09/19/2011 04:20 PM, Michael Geis wrote:
As a SoundFont designer, I personally make use of the majority of
the features in the SoundFont spec to create my instruments.
Playing just the exported samples from most of my SoundFonts will
usually sound quite different than playing the instrument designed
within the SoundFont bank. I can't speak for other SoundFont
designers, but I would imagine the same to be more-or-less true for
others as well.
We were under the (probably naive) impression that all a
sampler needs to do is loop over wave tables and apply
envelopes. Seeing that the soundfont specification actually
allows for greater complexity makes us wonder whether in order
to play soundfonts, the sampler needs to be able to do all the
things in the synthesis model. Unless there is a discrepancy
between what the specification allows for and what most
soundfonts look like in the wild. If a substantial fraction of
soundfonts just loop over wave tables and apply envelopes, the
sampler might still be useful for that subset of soundfonts if
it just grabbed their wave tables and envelope parameters. The
answer must be trivial for someone who has used soundfonts for
a bit, I must admit it is not clear to me.
Otherwise (i.e. if soundfont generally make full use of all
the parts of the synthesis engine laid out in the spec), I see
1.Implement the entirety of the synthesis model and use the
parsed soundfont parameters as input (That sounds like
reimplementing a lot of what fluidsynth already does).
2.Play the soundfonts via fluidsynth and record the output.
The sampler then loops over that output and applies envelopes.
Does 2. even make sense or is it likely to mangle the sounds?
If it is reasonable, how many notes should I have for each
pitch? One per pitch or one per envelope phase (i.e. 5 for the
DAHDSR envelopes since delay doesn't make a sound) per pitch?
IMO, #2 is not a very good idea for the following reasons:
- Most sustained samples have loop points which define a section
of the waveform that continues to repeat as a note is held down
indefinitely. Not only will you lose these helpful loop points
when recording the output from a sampler, but your recorded
waveform will often use up more memory than the original sample
if you record some amount of what is actually looped in the
original SoundFont. As an example: the Piano samples in
GeneralUser GS are looped, but the instrument envelope causes
each note to gradually fade out as the sound loops. If you
record the FluidSynth output of one of these notes, you will not
be able to recreate this loop due to the fade-out, and your
sample will have to be very long to capture the entire note (20
seconds to capture middle-C vs. approx. 4 seconds of original
- Sample stretching is used within SoundFonts to play the same
sample over a range of keys. Most instruments will use multiple
samples throughout the range of the keyboard (for example a
piano that has a new sample every 2-3 keys). If you record
FluidSynth's output, you will have to use your ear to determine
where the instrument switches from one sample to another (not to
even mention instruments that have different samples for each
velocity layer). If you get this wrong, you may end up
unnecessarily taking multiple recordings of the same sample
(just at different pitches). This is a very easy mistake to
make when some of the SoundFont's more advanced features such as
filters and modulators are active. Also, the sample sounds most
natural at its "root key", which is the note that the actual
sample was taken at. Recording any pitch other than the root
key will result in a small loss in audio quality from the
original due to the interpolation of sample points when changing
the sample's pitch and other factors. Unfortunately, discerning
the root pitch is pretty much impossible to do by ear.
I guess this might be related to how many wave tables are
usually used for a given instrument in the soundfont format.
One per pitch? One for every envelope phase of every pitch?
This varies from instrument to instrument. Some instruments even
use multiple samples or envelope regions that vary depending on how
hard you hit the key.
No problem. You really should consider using Swami or another
SoundFont editor to learn how SoundFonts are built. From there you
will come to better understand what a sampler does and what you will
need to do for your own project. You can also export the waveforms
directly from these editors, which would be much, much better than
trying to record them through FluidSynth.
My apologies if I am somewhat lacking coherence here, I am
still trying to get a decent grasp on the subject matter.
Good luck, and God Bless!
|
s3://commoncrawl/crawl-data/CC-MAIN-2023-50/segments/1700679102612.80/warc/CC-MAIN-20231210155147-20231210185147-00150.warc.gz
|
CC-MAIN-2023-50
| 4,755 | 80 |
https://equation-solver.com/equation-solver/logarithms/fraction-simplifiersolver.html
|
code
|
Below are some phrases that our visitors typed in today in order to reach site.
How can this be of help ?
identify the search phrase that you are interested in (i.e. fraction simplifiersolver) in the table below
Click on the appropriate program demo found in the same row as your search phrase
If you find the software demonstration useful click on the purchase button to purchase the software at a special low price extended only to equation-solver.com visitors
|
s3://commoncrawl/crawl-data/CC-MAIN-2024-18/segments/1712297295329.99/warc/CC-MAIN-20240425130216-20240425160216-00219.warc.gz
|
CC-MAIN-2024-18
| 462 | 5 |
https://gitlab.tails.boum.org/tails/tails/-/commit/3ba01ef42f1b673830e82db133954325a3932c06
|
code
|
Release process: mention another option to handle manual test results.
|...||...||@@ -1067,7 +1067,7 @@ Testing|
|1. Triage test results, reproduce bugs as needed, decide what the next|
|step is and make sure it happens: add to known issues? file ticket?|
|release blocker? improve the test description (steps, expected outcome)?|
|Update the website and Git repository|
|
s3://commoncrawl/crawl-data/CC-MAIN-2023-06/segments/1674764494986.94/warc/CC-MAIN-20230127132641-20230127162641-00069.warc.gz
|
CC-MAIN-2023-06
| 370 | 6 |
http://www.verycomputer.com/178_f7b5198637da390d_1.htm
|
code
|
Should be a simple problem. I have a 'C' program that responds to POSTs. Program
does some stuff then prints a new screen. However, with Netscape Commerce
Server (AIX), the results screen does not finish loading until the handler program
terminates. I need to have the page display (and complete loading), while the
handler remains for a few minutes. I have tried flushing and closing stout, adding
extra CRLF and a bunch of other things but no success. Many browsers just paint
about 3/4 of the screen and still show loading until the handler finally dies.
I have also tried 'forking' a separate program to handle the section of code that
has to process AFTER the screen is shown, and although the parent terminates,
the screen still 'loads' until the CHILD process terminates. This action is different
from the NCSA server, which seems to provide a different communications path
between the browser and your program.
Any ideas? The real problem here is ensuring the person has seen the results page,
so if there is any other way (with Netscape Commerce Server) to determine the
user has not abandoned the request, that would suffice. (NCSA returns an error
on printf while outputing the page if the client has gone away, Netscape does not).
What I am doing now is having the original handler wait on a semaphore and the
result page includes an acknowledge button, which when submitted, clears the
semaphore and completes the original transaction. I know this messy, but can't
see any other way to do it.
(P.S) If there is a more appropriate group for this question, please let me know...
|
s3://commoncrawl/crawl-data/CC-MAIN-2022-27/segments/1656103645173.39/warc/CC-MAIN-20220629211420-20220630001420-00026.warc.gz
|
CC-MAIN-2022-27
| 1,589 | 21 |
https://www.poweronemedia.com/heres-yammer-sharepoints-lipstick/
|
code
|
In the beginning, Microsoft created Sharepoint 2013. Yammer came along and saw that Sharepoint needed a partner to reveal its inner personality and beauty. The match was made and it was good. This is because even Though SharePoint Server 2013 offers standard enterprise social features, Yammer offers a more genuine enterprise social experience. Yammer is a top-notch enterprise social network that builds bridges between employees, discussions, content, and enterprise data all to one sole location. Yammer enables employees to increase productivity by empowering them to collude in real-time in spite of differences in department, location, branch, and enterprise applications. You can even integrate Yammer functionality to your SharePoint sites by including Yammer on your navigation bar or by utilising Yammer’s app to implant a Yammer feed into a site.
Yammer is available in two different models: Yammer Enterprise and Yammer Basic. Yammer Basic is free of charge and is readily available to all users. It offers an opportunity for employees to collude with colleagues in their company. Yammer Enterprise is a premier model that expands a business’s basic Yammer network. It is provided both as a stand-alone item and with many different SharePoint Online and Office 365 plans. Yammer Enterprise offers more tools and resources to assist organizations design the most efficient possible enterprise social network.
You can select if you want to use Yammer or SharePoint’s social features. If you want to utilise the effectiveness of Yammer with SharePoint, you can select any of the following means.
One way is to add Yammer to the navigation bar on Sharepoint 2013. You can execute this task via a toggle switch to switch your Newsfeed link with a Yammer link in the top navigation bar for SharePoint. This feature is offered in Service Pack 1 (SP1) for SharePoint Server.
You can utilize Yammer newsfeed features instead of Sharepoint. To understand the full potential of Yammer, it certainly helps to switch the default SharePoint Server 2013 enterprise social features with equal Yammer features. You can take the SharePoint Server social web parts away from My Sites as well as team sites, and you can conceal the user interface controls that offer social functionality.
Try the Yammer app for Sharepoint. The recent Yammer app for SharePoint allows you to integrate Yammer feeds (like a group feed, the home feed, or comment feeds) into on-premises sites effectively making them more attractive and socially engaging.
Was this post helpful? Feel free to leave your feedback in the comments below.
|
s3://commoncrawl/crawl-data/CC-MAIN-2023-40/segments/1695233506045.12/warc/CC-MAIN-20230921210007-20230922000007-00465.warc.gz
|
CC-MAIN-2023-40
| 2,614 | 7 |
https://www.databricks.com/session/hive-bucketing-in-apache-spark
|
code
|
Bucketing is a partitioning technique that can improve performance in certain data transformations by avoiding data shuffling and sorting. The general idea of bucketing is to partition, and optionally sort, the data based on a subset of columns while it is written out (a one-time cost), while making successive reads of the data more performant for downstream jobs if the SQL operators can make use of this property. Bucketing can enable faster joins (i.e. single stage sort merge join), the ability to short circuit in FILTER operation if the file is pre-sorted over the column in a filter predicate, and it supports quick data sampling.
In this session, you’ll learn how bucketing is implemented in both Hive and Spark. In particular, Patil will describe the changes in the Catalyst optimizer that enable these optimizations in Spark for various bucketing scenarios. Facebook’s performance tests have shown bucketing to improve Spark performance from 3-5x faster when the optimization is enabled. Many tables at Facebook are sorted and bucketed, and migrating these workloads to Spark have resulted in a 2-3x savings when compared to Hive. You’ll also hear about real-world applications of bucketing, like loading of cumulative tables with daily delta, and the characteristics that can help identify suitable candidate jobs that can benefit from bucketing.
Session hashtag: #SFdev10
Tejas Patil is a Spark Committer and Tech Lead in the Spark team at Facebook. For past 7 years, he has worked on several projects related to building large scale distributed data processing systems responsible for handling Facebook's batch workloads. He is currently a PMC member and committer of Apache Nutch and has contributed to several open source projects. Tejas obtained a Master’s Degree in Computer Science from University Of California, Irvine.
|
s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030338244.64/warc/CC-MAIN-20221007175237-20221007205237-00646.warc.gz
|
CC-MAIN-2022-40
| 1,848 | 4 |
https://www.techtarget.com/searchvirtualdesktop/opinion/VMware-Workspace-ONE-for-Microsoft-Endpoint-Manager-What-does-it-mean
|
code
|
We’ve been following VMware’s work on Windows 10 management for years, but the last few months have been especially active and interesting.
At VMworld Europe, VMware announced a new offering called VMware Workspace ONE for Microsoft Endpoint Manager.
Microsoft Endpoint Manager—which represents the integrated future of Microsoft Intune and SCCM—had just been announced the day before at Microsoft Ignite 2019. This was big news, and Microsoft also declared that SCCM isn’t going away, and that co-management (using Intune and SCCM together on the same device) isn’t just a bridge, but rather, it can be a destination.
On top of all this, earlier this year, Microsoft made a change to SCCM that made it more difficult to use third-party MDM connections (e.g., MDM enrollment in Workspace ONE) on devices managed by SCCM.
I already wrote about Microsoft Endpoint Manager (or MEM, as folks are starting to call it); today we’ll look at Workspace ONE for Microsoft Endpoint Manager and what it means.
The background on Workspace ONE and SCCM
Digging back in the archives, you can review VMware’s work on Windows management over the last five years. In 2014, they offered MDM plus an agent on Windows 8.1; in 2015, they were talking about doing AppVolumes on physical devices; and 2016, they were partnered with Tanium for VMware Trustpoint.
It was after VMworld 2017 when we felt that things were getting quite serious. Gabe Knuth (then an independent blogger here at BrianMadden.com) wrote this about VMware and Windows management:
“The general sense that I have coming out of VMworld is that VMware, perhaps more than anyone else, understands exactly what has to happen to get enterprises to switch to UEM. Without a doubt, there is more to do to make this appealing to all enterprises, but with their foot on the accelerator, VMware can get their pretty quickly.”
By this time, AirWatch (as it was still called) could deploy GPOs and join machines to a domain; and VMware used Adaptiva for software distribution. AirWatch could even connect to machines that were managed via SCCM (before the co-management days).
On Microsoft’s side, at Ignite 2017, they announced a change in Windows 10 1709 that would allow devices to be enrolled in both SCCM and MDM at the same time.
Microsoft built an integration between SCCM and Intune to coordinate the handoff of management tasks from one side to another, and called this “co-management.”
Other vendors, including VMware, could take also take advantage of the changes in Windows 10 1709, and enroll SCCM-managed devices into their own MDM servers. Microsoft called this “coexistence,” though the rest of the world calls this co-management.
A few months later, VMware introduced AirLift, their product for pulling configuration data out of SCCM, and coordinating management with Workspace ONE. At the time, I wrote: “I don’t think anybody is ready to turn off SCCM any time soon, but the battle lines on this front have been drawn.” VMware was off and running on a super ambitious path.
At VMworld 2019, we got a progress report: VMware had enrolled 1 million Windows 10 devices in the first half of 2019. It’s almost the end of the year, and I haven’t seen any more recent numbers, but they did say that they were enrolling a Windows 10 device every 16 seconds. So, they should be closing in on 2 million devices for 2019.
Co-management gets harder
Now, there was one wrinkle in all of this. The SCCM 1902 update rollup made an interesting change for co-management and coexistence: If an SCCM-managed device is enrolled into an MDM server other than Intune (i.e., Workspace ONE and any other third-parties), the SCCM agent reverts to a read-only mode.
So, this essentially kills a whole swath of co-management (or coexistence in Microsoft’s terms) scenarios. This did cause problems for some customers, and it sounds like it wasn’t communicated very well.
Microsoft’s justification is that they did this to avoid potential conflicts, but this seems a little odd to me. VMware had AirLift to keep things coordinated; and also I thought that the DeviceManageability CSP was supposed to help with this, too. So, the changes to the SCCM agent seem a bit old-school Microsoft to me, but that’s business.
Enter Microsoft Endpoint Manager and Workspace ONE for MEM
As I covered in my article about Microsoft Endpoint Manager, Microsoft is now all about enabling customers to use SCCM and co-management as long as they want, and they’re connecting SCCM to Intune in a new cloud-based management plane.
Now we can finally get to the whole point of this article: Workspace ONE for Microsoft Endpoint Manager will be VMware’s recommendation for customers that are using both products.
As Shanker Iyer wrote, “Workspace ONE will build value on top of Microsoft Endpoint Manager for management of Windows 10 devices.”
What does this mean? The first round of features all appear to be about using Workspace ONE Intelligent Hub on top of MEM, and Shanker highlighted newer features like the employee onboarding experience, and Digital Employee Experience Management. Apps managed by MEM can show up in Intelligent Hub, and customers can use Workspace ONE Tunnel. VMware expects a preview to be out in early 2020.
There were a few more clarifying details in the “Definitely Not Official VMware EUC Podcast” with Brian Madden (the person) and Gabe Knuth. (So, remember that this is not official, and while both Brian and Gabe used to work at BrianMadden.com, they’re both VMware employees now.)
Under Workspace ONE for MEM, MEM will be doing the actual device management, via whatever combination of SCCM and Intune you want. (I’ve heard other sources describe it this way, too.) VMware will take their Windows agent, which can both act as the management authority and do a bunch of other things (like analytics, security, user experience, crash reporting, baseline configuration reporting, etc.) and make it so that it can be just do all of the latter tasks. So when it comes to device management, the VMware agent gets out of the way, and lets MEM do the management.
Again, this is early, not completely official information; but it makes sense given the coexistence changes in SCCM 1902. We’ll try to get more technical details as we get closer to the preview.
So what does this mean?
As I wrote last week, the fundamental question for customers is when and how you want to get off of SCCM.
On the podcast, Brian and Gabe (with the “definitely not official” disclaimer) laid out a description of VMware’s position: VMware never had SCCM to fall back on, so they had additional incentive to make Workspace ONE’s Windows 10 capabilities richer, with things like GPO support and more app distribution capabilities. (Jumping back into my thoughts: So that year and a half where co-management with SCCM actually worked was more like a temporary gift.)
Brian went on to say that basically, VMware wants to be flexible here. Workspace ONE is a very broad portfolio with a whole range of use cases, so if customers want to use MEM for Windows 10, that’s fine. With the Workspace ONE for MEM offering, the Windows 10 devices will still be part of the Workspace ONE environment, and users will still have the Intelligent Hub.
For customers that do want to go ahead and have Workspace ONE be the device management authority on Windows 10, the SCCM 1902 changes around third-party MDM just mean that the intermediate option of managing some tasks via Workspace ONE MDM and others via SCCM isn’t available anymore.
Customers will still be able to use AirLift, though. In fact, at VMworld US 2019, VMware announced AirLift 2.0, which can automatically migrate GPOs from SCCM. (Check out Jon Towles’ article on AirLift for an overview. )
Back on the Microsoft side, to be fair, I should point out that for all the talk about how co-management can be a destination, Microsoft also said in an FAQ (PDF) that customers can really do what they want. And in several presentations at Ignite, Microsoft recommended choosing a date where customers decide that all new devices will be managed from the cloud side (i.e., Intune) of MEM only. So they also get that cloud-based management is going to be an important goal for many customers.
What’s your path?
At this point, I’m starting to wonder how many customers are going to want to mess around with co-management anyway. There’s a lot of value in at least making a traditionally managed device show up in some form or another in your cloud UEM platform, especially around conditional access. But full on co-management is more complicated.
In some ways, we can think of this like doing Windows migrations. Remember how much talk there was when Windows 10 came out about how you could do relatively smooth in-place upgrades without re-imaging machines? And how many people actually did that? Migrating via attrition as part of hardware refresh cycles was a lot easier.
We’re just over a month away from January 14, 2020, the Windows 7 end of life. Digging back into the BrianMadden.com archives again, back in 2017 Gabe wrote an article called “Migrating to Windows 10 and modern management at the same time sounds nice, but is it practical?” His conclusion, now proven correct, was that the answer was no. Instead, he wrote that January 15, 2020 was when we should all be ready for a change in Windows management. (Well, maybe we should take a week or two to rest.)
Getting rid of SCCM is going to be a long process, and now I’m wondering whether migrating laptops to all-cloud, SCCM-free management will happen via attrition during hardware refresh cycles. This way, you can start with a clean slate, provisioning devices with AutoPilot or the out-of-box experience.
Customers will still have to decide between Workspace ONE and MEM for Windows 10, but it should be a slightly less complicated decision. They won’t be worrying about co-management, coexistence, or migration paths—they’ll just be concerned about what you can do with the platform alone.
Regardless of when and how you decide to jump into modern management, VMware customers have AirLift to ease their migration from SCCM, or Workspace ONE for MEM if they plan on keeping SCCM or MEM around for a while. Modern management can mean different things to different people, so the more options and flexibility that are available, the better.
|
s3://commoncrawl/crawl-data/CC-MAIN-2024-10/segments/1707947473598.4/warc/CC-MAIN-20240221234056-20240222024056-00478.warc.gz
|
CC-MAIN-2024-10
| 10,459 | 41 |
http://trimbletools.com/jcmbsoft/html/fast_static.html
|
code
|
The Fast Static (Rapid Static) Project is being developed because
- Fast/Rapid static still is an important technique for GPS
- No manufacturer has done anything with Fast/Rapid Static for the last 10 years
- It is a good system to investigate how the internet can improve productively.
- It lends it self well to a Jabber based implementation
The Project has a number of planned stages
- Fast Static Classic
Better Fast Static
- What you get from a manufacturer today
- Enter the times wait for the time
- A Version was Created that works over IP as well Fast_Static_Classic_IP
- Timers based on a different elevation mask form the storage
Log at 10° but the timers based on 15°
- SNR monitoring
If the SNR’s are low then the SV shouldn’t be counted in the timers
Real time processing
- Check data availablity between all the receivers, not just one.
- So if a SV is blocked at one location and a different one is blocked at another location then both SV’s would be removed from the timers
- Start when you get there, wait for everyone.
- Currently for the timers to be useful you have to wait for everyone to get to the point. Which in the >2 receiver case isn’t optimal.
- It should be possible to check to see if the base lines will solve in near real time, the data is sent to a server, it processes the baselines, checks the loops and indicates if you are ready to go or not.
- Might not get to this one
|
s3://commoncrawl/crawl-data/CC-MAIN-2017-26/segments/1498128323889.3/warc/CC-MAIN-20170629070237-20170629090237-00126.warc.gz
|
CC-MAIN-2017-26
| 1,419 | 22 |
https://evcforum.net/dm.php?control=newerthread&t=17325
|
code
|
There is an article in Answers Research Journal with title: Comprehensive Analysis of Chimpanzee and Human Chromosomes Reveals Average DNA Similarity of 70%
What do you think about this article. Wouldn't be greate to reproduce the data from publicaly available resources using BLASTN algoritm? I am very interested in reproducing the results. If we achieve the same data we can use the same methodology for two species inside one baramin. I think it would be very interesting.
All resources (blastn program, DNA sequences) are available free.
All of them indicate large differences 70%-89% between Human and Chimpanzee, which is in contrast to generally accepted difference between 94%-98%. This is presented as proof that Humans and Chimpanzees are not closly related as it was presented for decades.
I was very interested in these papers. I am software developer so I have close to "playing" with algorithms. And because all resources are free on the Internet I was wondering if I could reproduce results.
I think these numbers are taken by different method than previous high similarity results. So I think other reference data is needed. But I have difficulties in obtaining results presented in papers above so I was wondering if there is anybody, who can help.
I have several goals:
1) Opinions: I would like to know what other think about this research and used methodology. 2) Verification: I would like to verify data from papers to be sure my method is exactly the same as it was used in papers. 3) Test: Because I am suspicious (my preliminary results are far from numbers preseted in paper), I would like to do some method verification test: e.g. compare Human-Human DNA. 4) Further research: By the same method compare DNA between species inside one baramin (e.g. mouse and rat).
I think sharing these information, advices, tips, hints in this disscussion would be benefical. We can find some interesting results together.
That's more but it might not help people. AiG is notorious for doing things badly. You need to tempt people to want to bother to read the papers which they may think is going to be a waste of time. Though some will enjoy just diving in anyway. Maybe they will post to ask that this be promoted without you doing more work.
If you could describe the methods a bit it might help.
However, maybe you are like me and don't know enough about this to unscramble it. If that is the case tell me and I'll promote this to see if others will help you out.
|
s3://commoncrawl/crawl-data/CC-MAIN-2022-49/segments/1669446710698.62/warc/CC-MAIN-20221129132340-20221129162340-00413.warc.gz
|
CC-MAIN-2022-49
| 2,481 | 12 |
https://www.thestorytinker.com/lets-play-167168-embarrassing-conversations-with-josalynn-and-ocean/
|
code
|
167: Vikki, Angela, and Sam discuss the dragon egg Sam received, throwing in a reference to Subzero, and Dallas is proud to serve them super special coffee, his newfound passion. Monica enters and she and Sam talk about Sam's relationship with Charles. Monica dispenses good advice and prompts Sam to think about how she feels about their coupling. 168: As final confirmation that Marshall is Dion, Abe and Vikki sit Marshal down for a TALK. He is hesitant to tell Sam that he is Dion, but Abe correctly informs him that withholding that information is a breach of trust. He reveals how lonely he is and how much he has grown to appreciate Sam's genuine kindness. Abe wants Marshall to tell Sam, but is also kind enough to invite him to join their guild. Later, Sam is at the grocery store, where she is mortified at buying condoms. The selection overwhelms her and she is sure everyone is looking at her. To make matters worse, an attendant Austin tries to hit on her. Will Sam have the courage to chase him away herself? 𝗕𝗼𝗻𝘂𝘀 𝗖𝗼𝗻𝘁𝗲𝗻𝘁: 𝗪𝗵𝗼 𝘄𝗼𝘂𝗹𝗱 𝗖𝗵𝗮𝗿𝗹𝗲𝘀 𝘁𝗮𝗸𝗲 𝗮𝗱𝘃𝗶𝗰𝗲 𝗳𝗿𝗼𝗺? *** 𝗥𝗲𝗮𝗱 𝗟𝗲𝘁'𝘀 𝗣𝗹𝗮𝘆 𝗼𝗻 𝗪𝗲𝗯𝘁𝗼𝗼𝗻𝘀: https://www.webtoons.com/en/romance/letsplay/list?title_no=1218&page=1 𝗦𝘂𝗽𝗽𝗼𝗿𝘁 𝗠𝗼𝗻𝗴𝗶𝗲 𝗼𝗻 𝗣𝗮𝘁𝗿𝗲𝗼𝗻: https://www.patreon.com/mongrelmarie/ 𝗦𝘂𝗽𝗽𝗼𝗿𝘁 𝗧𝗵𝗲 𝗦𝘁𝗼𝗿𝘆𝗧𝗶𝗻𝗸𝗲𝗿 𝗼𝗻 𝗣𝗮𝘁𝗿𝗲𝗼𝗻: https://www.patreon.com/thestorytinker 𝗦𝘂𝗯𝘀𝗰𝗿𝗶𝗯𝗲 𝘁𝗼 𝗹𝗶𝘀𝘁𝗲𝗻 𝘁𝗼 𝗕𝗼𝗻𝘂𝘀 𝗖𝗼𝗻𝘁𝗲𝗻𝘁: https://anchor.fm/thestorytinker/subscribe 𝗙𝗼𝗹𝗹𝗼𝘄 𝗧𝗵𝗲 𝗦𝘁𝗼𝗿𝘆𝗧𝗶𝗻𝗸𝗲𝗿 𝘀𝗼𝗰𝗶𝗮𝗹𝘀: https://www.instagram.com/thestorytinker/ https://www.facebook.com/thestorytinker https://twitter.com/storytinker https://www.tiktok.com/@thestorytinker 𝗔𝘃𝗮𝗶𝗹𝗮𝗯𝗹𝗲 𝗼𝗻 𝗬𝗼𝘂𝗧𝘂𝗯𝗲 𝗮𝗻𝗱 𝗮𝗹𝗹 𝗽𝗼𝗱𝗰𝗮𝘀𝘁 𝗽𝗹𝗮𝘁𝗳𝗼𝗿𝗺𝘀: 𝗦𝗽𝗼𝘁𝗶𝗳𝘆, 𝗔𝗽𝗽𝗹𝗲, 𝗚𝗼𝗼𝗴𝗹𝗲, 𝗮𝗻𝗱 𝗺𝗼𝗿𝗲! 𝗧𝗵𝗮𝗻𝗸 𝘆𝗼𝘂 𝘁𝗼 𝗺𝘆 𝗰𝘂𝗿𝗿𝗲𝗻𝘁 𝗣𝗮𝘁𝗿𝗼𝗻𝘀: Suzie, LadyLibris, Lily, Jenny, Molly, Veronica, Emily, Jo Rochelle, Saucy Tuggles, Anne Rose, Alexa, Misti, Joanne, Imilda, Ester, I’m Watching You Ppl, Emily Jean, Jenn, Erin, Kay, Lily, Becketts, Deranda, Krystine, Sadie, Kelly, Theresa, Mrs. Gastaldo, Tatianna, Louisa and Rachel.
|
s3://commoncrawl/crawl-data/CC-MAIN-2023-23/segments/1685224649348.41/warc/CC-MAIN-20230603233121-20230604023121-00377.warc.gz
|
CC-MAIN-2023-23
| 2,777 | 1 |
https://t37.net/this-is-the-support-team-you-re-looking-for.html
|
code
|
Doing a postmortem analysis of the crisis you’ve faced is not only a great way to understand what happened so it never happens again. It’s also the best way I know to improve crisis response management for the next time.
Because there’s always a next time.
Looking back at all the crises I’ve faced in the past 10 years, what I’ve missed the most to solve them efficiently was an awesome support team.
An awesome support is not (only) the one that answers you Saturday 9:00 PM and replaces your son’s Kindle within 18 hours so he can read during his vacation. An awesome support is also the front line that manages the client crisis side while you’re working.
The awesome support team is product and tech savvy, composed half of people with great empathy and engineers. It removes pressure from the crisis solving team both from the inside and the outside by filtering new tickets, understanding if they’re related to the problem you’re facing or something completely unrelated.
The awesome support team comes with the awesome product, not after it’s released. And as a front line team, it deserves its part of honor after the battle when the medals are given.
|
s3://commoncrawl/crawl-data/CC-MAIN-2017-43/segments/1508187824618.72/warc/CC-MAIN-20171021062002-20171021082002-00848.warc.gz
|
CC-MAIN-2017-43
| 1,180 | 6 |
http://okapiframework.org/wiki/index.php?title=Markdown_Filter
|
code
|
The Markdown Filter is an Okapi component for extracting translatable text from Markdown files. See https://en.wikipedia.org/wiki/Markdown for more information about the format. Markdown is a family of formats, not all of them mutually compatible. This filter is designed to work with markdown based on the CommonMark specification, with additional features to support GitHub-flavored Markdown.
The filter decides which encoding to use for the input file using the following logic:
If the file has a Unicode Byte-Order-Mark: Then, the corresponding encoding (e.g. UTF-8, UTF-16, etc.) is used. Otherwise, the input encoding used is the default encoding that was specified when setting the filter options.
The Markdown filter of Okapi Framework version 0.36 and later no longer supports the Inline Code Finder because it now uses the HTML filter as its subfilter, which is known to be incompatible with the Inline Code Finder.
- Translate URLs (translateUrls)
- By default, URLs in link and image statements are not exposed for translation. If this option is enabled, they will be extracted. Note: URLs are currently extracted inline in their containing segment, rather than as a subflow. Default: false
- Translate Code Blocks (translateCodeBlocks)
- This option, enabled by default, controls whether the contents of fenced code blocks are exposed for translation. Default: true
- Translate YAML Metadata Header (translateImageAltText)
- Some markdown formats support a YAML Metadata Header that contains key/value data. By default, this header is not exposed for translation. When the "Translate YAML Metadata Header" option is enabled, the header will be parsed and the metadata values will be exposed for translation. Default: false
- Translate Image Alt Text (translateImageAltText)
- The alt text for a graphic image in the form of  or as the alt attribute of an img tag <img src="https://foo.com/images/bar.jpg" alt="alt text"> will be extracted if this parameter is true. Default: true.
- HTML Subfilter Configuration ID (htmlSubfilter)
- The custom configuration ID of the HTML filter that will be called to process HTML contents within Markdown documents. The configuration file must be saved in a known location with .fprm suffix. Specify nothing to use the default HTML filter configuration tailored for the Markdown filter. Default: (empty)
- None known
|
s3://commoncrawl/crawl-data/CC-MAIN-2018-17/segments/1524125948126.97/warc/CC-MAIN-20180426105552-20180426125552-00201.warc.gz
|
CC-MAIN-2018-17
| 2,408 | 15 |
http://two.pairlist.net/pipermail/reportlab-users/2011-August/010205.html
|
code
|
[reportlab-users] Report for multiple objects
v+python at g.nevcal.com
Tue Aug 23 15:11:08 EDT 2011
On 8/23/2011 9:24 AM, Tim Roberts wrote:
> Asif Jamadar wrote:
>> I'm trying to implement custom action in django admin page. When user selects multiple objects from admin page and click on the custom action, that should generate report for each selected object.
>> Suppose user select 4 objects from the admin page, then the custom action should generate 4 reports.
> One web request can only produce one response. That's not a ReportLab
> or a PDF restriction, that's a fundamental HTTP concept.
> As I see it, you have two choices. One, you could produce the four
> reports sequentially, combined into one single PDF file. Two, you can
> build a second web page that basically says "here are your 4 reports,
> click on them one at a time to download them."
There are more choices...
Three, (a variation of two) return with a page containing a list of
reports 1-4, use meta-refresh to automatically start the download of the
first report, and allow the user to click on the others to initiate them.
Four, put the 4 reports into a .zip file and download it
Five, (this one depends on client side scripting, and is theoretical (I
haven't tried it)), return a page that uses scripting to initiate 4
requests, one for each report; each would pop up a dialog to save the
report when the download response arrives. If anyone gets this one
working, please share the results :) I've been wanting to try it for a
while, but have had higher priority projects...
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the reportlab-users
|
s3://commoncrawl/crawl-data/CC-MAIN-2022-27/segments/1656103661137.41/warc/CC-MAIN-20220630031950-20220630061950-00304.warc.gz
|
CC-MAIN-2022-27
| 1,671 | 27 |
https://askubuntu.com/questions/388708/what-may-cause-the-desktop-wallpaper-to-be-scrambled
|
code
|
I have a very strange problem. Sometimes when I work a long time with my laptop while the desktop is hidden a strange thing happens. I show the desktop and it is all scrambled.
This problem is all the stranger since it happens only sometimes and I did not have it always.
How can I diagnose whether it is a hardware or software problem? Has anyone had similar symptoms?
This problem persists after upgrading to 13.10.
When the wallpaper gets scrambled and I change it to a different one and then back again, it seems fixed for a while. Does this fact indicate that it is not a hardware problem?
Now this happens also for some windows, not just the desktop.
|
s3://commoncrawl/crawl-data/CC-MAIN-2020-05/segments/1579251689924.62/warc/CC-MAIN-20200126135207-20200126165207-00391.warc.gz
|
CC-MAIN-2020-05
| 656 | 6 |
https://sourceforge.net/projects/huaweie1750ubun/?source=directory&stars=0
|
code
|
a bash-script made to configure your ubuntu computer to be able to use huawei e1750 broadband-modem. It uses the ubuntu-repositories, but read through the file to get a clue how to edit your own system
Follow huawei e1750 ubuntu
Rate This ProjectLogin To Rate This Project
Agree with the previous reviewer. This is a godsend!
|
s3://commoncrawl/crawl-data/CC-MAIN-2018-05/segments/1516084891980.75/warc/CC-MAIN-20180123151545-20180123171545-00125.warc.gz
|
CC-MAIN-2018-05
| 325 | 4 |
https://www.guru.com/service/xamrin-and-hybrid-app-development/india/gujarat/gandhinagar/3824538
|
code
|
My name is Ankit. I'm a Professional Mobile Developer.I have 6+ Year Experience in the Field. Do you want Professional Work and want it quickly TRUST ME!! You are at the right place..
Extra Fast Delivery.
So feel free to hire me... Thanks - Ankit
9 Reviews For This Service
|
s3://commoncrawl/crawl-data/CC-MAIN-2020-05/segments/1579251799918.97/warc/CC-MAIN-20200129133601-20200129163601-00514.warc.gz
|
CC-MAIN-2020-05
| 273 | 4 |
http://beaconpeople.org/new-events/2018/9/20/enneagram-workshop
|
code
|
Seems like everyone is talking about Enneagrams. What's your number? What test did you take? What podcasts are you listening to? At this Beacon People event we will get a complete overview of Enneagrams and how knowing your type will encourage you and make you a more productive person in life and work.
|
s3://commoncrawl/crawl-data/CC-MAIN-2019-39/segments/1568514573258.74/warc/CC-MAIN-20190918065330-20190918091330-00257.warc.gz
|
CC-MAIN-2019-39
| 303 | 1 |
https://bounties.gitcoin.co/hackathon/metafest/onboard
|
code
|
Apr 7, 2021 - Apr 16, 2021
This hackathon event had ended at April 16, 2021, 11:59 p.m., please check the ongoing hackathons.
Context: This hackathon is a part of a larger event - MetaFest. A month long virtual festival/conference/hackathon brought to you by MetaFam of MetaGame.
Take part, stand to win a share of the prize pool of up to $10k & an entry ticket to MetaGame - where you can earn a shit-ton more 😱
Think meta! It's not just about hacking on something that improves MetaGame or the Ethereum space - it’s about anything that improves the world.
It’s also not just about writing code. You could write a good article or a good crypto onboarding guide! You could hack on some art or memes, make an educational video or even clean up a local beach 🤷♂
First time hearing about MetaGame? Think of it as:
Wish to fall down this rabbit hole?
Ready to start hacking on one of our quests or raids, or start your own?
Think anything that helps the Ethereum ecosystem, be it a cool new consumer product, a piece of infrastructure or tooling, or an educational or promotional material.
Ready to create but need a bit of help?
World improvement hacks can be anything that improves the lives & ecosystems around you.
Submit your projects by Friday 16th at 00:00 UTC then give it until Sunday 18th for players of MetaGame to make their judgement & start sending out the rewards.
For the Ethereum & World improvement categories the prize pool is limited to $10k with possible follow-on funding by the MetaCartel DAO for promising Ethereum based projects. For the MetaGame improvements category, there is no set limit on the prize pool 🙃
Visit the Prize Explorer to check out the prizes posted by our hackathon sponsors. Click each prize to show important details, including the submission requirements, submission deadline, etc.
Chat with other hackers, ask sponsors and the Gitcoin team questions, find or create a team, and communicate real-time. Click here to join the party!
When your team is formed please have one of your teammates navigate to each prize page you plan to compete for and click the “Start Work” button.
Build your cool ideas and make your vision come true with your team!
When your project is completed, submit your work by clicking the “Submit Work” button on the prize page/
Tell us a little about you:
Our tools are based on the principles of earn (💰), learn (📖), and meet (💬). Select the ones you are interested in. You can change it later in your settings.
Gitcoin products can help grow community around your brand. Create your tribe, events, and incentivize your community with bounties. Announce new and upcoming events using townsquare. Find top-quality hackers and fund them to work with you on a grant.
These are the organizations you own. If you don't see your organization here please be sure that information is public on your GitHub profile. Gitcoin will sync this information for you.
Select the products you are interested in:
Out of the box you will receive Tribes Lite for your organization. Please provide us with a contact email:
|
s3://commoncrawl/crawl-data/CC-MAIN-2023-14/segments/1679296949644.27/warc/CC-MAIN-20230331144941-20230331174941-00156.warc.gz
|
CC-MAIN-2023-14
| 3,100 | 25 |
http://www.avsforum.com/forum/161-camcorders/1489732-hd-writer-ae-4-1-a.html
|
code
|
I have a recording question and wonder if anyone can help !
I am trying to make a DVD compiled of still images and AVCHD recordings from my Panansonic HC-X800. However the software provided with the camcorder will only allow me to records AVCHD to DVD and still images to SD card!.
Is there a way around this ?
Software: HD Writer AE 4.1
Camcorder: Panasonic HC-X800
Dell Laptop Running Windows 7
|
s3://commoncrawl/crawl-data/CC-MAIN-2017-34/segments/1502886105961.34/warc/CC-MAIN-20170820015021-20170820035021-00716.warc.gz
|
CC-MAIN-2017-34
| 396 | 6 |
https://revyo.com/cardano-pioneering-a-new-generation-of-cryptocurrencies/
|
code
|
If you have been keeping an eye on the world of cryptocurrencies, you’ve likely heard about Cardano. Launched in 2017, Cardano stands out from the crypto crowd due to its commitment to peer-reviewed research and formal verification methods. Let’s delve into Cardano, its unique approach, and the ways it’s striving to revolutionize the digital currency landscape.
Cardano is the brainchild of Ethereum co-founder Charles Hoskinson, developed by the technology company IOHK (Input Output Hong Kong). Unlike many other cryptocurrencies that were born from informal beginnings, Cardano was conceived with an academic, research-led approach. Think of it as a cryptocurrency meticulously built by a team of engineers and academics, like an architect meticulously designing a building, with each brick laid based on extensive research and calculations.
Cardano’s blockchain operates in a unique two-layered architecture, a feature distinguishing it from many of its competitors. The first layer, called Cardano Settlement Layer (CSL), is used to carry out cryptocurrency transactions, much like Bitcoin. The second layer, known as Cardano Computation Layer (CCL), is where smart contracts and decentralized applications (dApps) operate.
This two-layered approach solves a fundamental issue that has plagued many other cryptocurrencies – scalability. With Cardano, the network’s speed isn’t hampered as usage grows, making it more scalable than many existing cryptocurrencies. This is like a city building a multi-level highway system to address increased traffic without causing congestion.
Another distinctive aspect of Cardano is its commitment to sustainability. Cardano’s consensus protocol, known as Ouroboros, is a proof-of-stake (PoS) protocol, where validators are chosen based on the number of coins they hold and are willing to ‘stake’ for the sake of the network. This is a more energy-efficient alternative to the power-hungry proof-of-work (PoW) protocol used by Bitcoin.
Cardano also stands out for its forward-thinking approach to governance. The project plans to establish a treasury system, funded by a portion of transaction fees, which will be used to finance future developments and improvements to the network. This could ensure Cardano’s longevity by making it self-sustaining.
As of now, Cardano’s ADA token is one of the top cryptocurrencies by market cap. Its unique architecture and its commitment to a research-driven, sustainable approach have made it a favorite among many investors and technologists.
However, like all investments, investing in Cardano’s ADA comes with risks. The project’s lofty goals are still in development, and there’s always uncertainty in how future updates will be received by the wider crypto community.
In the grand scheme of cryptocurrencies, Cardano is a fascinating project with significant potential. It represents a unique blend of technology and academia, pushing the boundaries of what’s possible with blockchain technology. Whether or not it will fully achieve its aims remains to be seen, but there’s no doubt that it’s a project worth watching.
|
s3://commoncrawl/crawl-data/CC-MAIN-2023-40/segments/1695233506528.3/warc/CC-MAIN-20230923194908-20230923224908-00526.warc.gz
|
CC-MAIN-2023-40
| 3,141 | 9 |
http://www.qnasap.com/243111
|
code
|
- Business One
- EIM Solutions
- Enterprise Portal
- Industry Solutions
- SAP Db And OS
- SAP Tools And Tech
- Solution Manager
I am looking for a suitable event in Adobe Interactive forms which can cancel the processing of the form if any field value is not accoding to the validation while the form is getting submitted? Ideally it should cancel the submit of the form.
Actually I am a using SAP Adobe Interactive forms using transaction SFP in ABAP Workbench and am trying to check the possible ways to apply validations (client side validations) to the form data. These forms are expected to get called from HCM Processes and Forms processes wherein the form is getting called from the standard Start Application on portals.
I checked various events such as preSubmit, preSave but none of them served my purpose since my form is getting called from the external source. Any idea what event could serve my purpose?
Browse more questions like this in module: Netweaver
|
s3://commoncrawl/crawl-data/CC-MAIN-2013-48/segments/1386164027414/warc/CC-MAIN-20131204133347-00082-ip-10-33-133-15.ec2.internal.warc.gz
|
CC-MAIN-2013-48
| 970 | 11 |
https://aclanthology.org/people/w/william-merrill/
|
code
|
Language models are often trained on text alone, without additional grounding. There is debate as to how much of natural language semantics can be inferred from such a procedure. We prove that entailment judgments between sentences can be extracted from an ideal language model that has perfectly learned its target distribution, assuming the training sentences are generated by Gricean agents, i.e., agents who follow fundamental principles of communication from the linguistic theory of pragmatics. We also show entailment judgments can be decoded from the predictions of a language model trained on such Gricean data. Our results reveal a pathway for understanding the semantic information encoded in unlabeled linguistic data and a potential framework for extracting semantics from language models.
Training a referring expression comprehension (ReC) model for a new visual domain requires collecting referring expressions, and potentially corresponding bounding boxes, for images in the domain. While large-scale pre-trained models are useful for image classification across domains, it remains unclear if they can be applied in a zero-shot manner to more complex tasks like ReC. We present ReCLIP, a simple but strong zero-shot baseline that repurposes CLIP, a state-of-the-art large-scale model, for ReC. Motivated by the close connection between ReC and CLIP’s contrastive pre-training objective, the first component of ReCLIP is a region-scoring method that isolates object proposals via cropping and blurring, and passes them to CLIP. However, through controlled experiments on a synthetic dataset, we find that CLIP is largely incapable of performing spatial reasoning off-the-shelf. We reduce the gap between zero-shot baselines from prior work and supervised models by as much as 29% on RefCOCOg, and on RefGTA (video game imagery), ReCLIP’s relative improvement over supervised ReC models trained on real images is 8%.
Transformers have become a standard neural network architecture for many NLP problems, motivating theoretical analysis of their power in terms of formal languages. Recent work has shown that transformers with hard attention are quite limited in power (Hahn, 2020), as they can be simulated by constant-depth AND/OR circuits (Hao et al., 2022). However, hard attention is a strong assumption, which may complicate the relevance of these results in practice. In this work, we analyze the circuit complexity of transformers with saturated attention: a generalization of hard attention that more closely captures the attention patterns learnable in practical transformers. We first show that saturated transformers transcend the known limitations of hard-attention transformers. We then prove saturated transformers with floating-point values can be simulated by constant-depth threshold circuits, giving the class TC0 as an upper bound on the formal languages they recognize.
The capacity of neural networks like the widely adopted transformer is known to be very high. Evidence is emerging that they learn successfully due to inductive bias in the training routine, typically a variant of gradient descent (GD). To better understand this bias, we study the tendency for transformer parameters to grow in magnitude (ℓ2 norm) during training, and its implications for the emergent representations within self attention layers. Empirically, we document norm growth in the training of transformer language models, including T5 during its pretraining. As the parameters grow in magnitude, we prove that the network approximates a discretized network with saturated activation functions. Such “saturated” networks are known to have a reduced capacity compared to the full network family that can be described in terms of formal languages and automata. Our results suggest saturation is a new characterization of an inductive bias implicit in GD of particular interest for NLP. We leverage the emergent discrete structure in a saturated transformer to analyze the role of different attention heads, finding that some focus locally on a small number of positions, while other heads compute global averages, allowing counting. We believe understanding the interplay between these two capabilities may shed further light on the structure of computation within large transformers.
Much recent work in NLP has documented dataset artifacts, bias, and spurious correlations between input features and output labels. However, how to tell which features have “spurious” instead of legitimate correlations is typically left unspecified. In this work we argue that for complex language understanding tasks, all simple feature correlations are spurious, and we formalize this notion into a class of problems which we call competency problems. For example, the word “amazing” on its own should not give information about a sentiment label independent of the context in which it appears, which could include negation, metaphor, sarcasm, etc. We theoretically analyze the difficulty of creating data for competency problems when human bias is taken into account, showing that realistic datasets will increasingly deviate from competency problems as dataset size increases. This analysis gives us a simple statistical test for dataset artifacts, which we use to show more subtle biases than were described in prior work, including demonstrating that models are inappropriately affected by these less extreme biases. Our theoretical treatment of this problem also allows us to analyze proposed solutions, such as making local edits to dataset instances, and to give recommendations for future data collection and model design efforts that target competency problems.
Abstract Language models trained on billions of tokens have recently led to unprecedented results on many NLP tasks. This success raises the question of whether, in principle, a system can ever “understand” raw text without access to some form of grounding. We formally investigate the abilities of ungrounded systems to acquire meaning. Our analysis focuses on the role of “assertions”: textual contexts that provide indirect clues about the underlying semantics. We study whether assertions enable a system to emulate representations preserving semantic relations like equivalence. We find that assertions enable semantic emulation of languages that satisfy a strong notion of semantic transparency. However, for classes of languages where the same expression can take different values in different contexts, we show that emulation can become uncomputable. Finally, we discuss differences between our formal model and natural language, exploring how our results generalize to a modal setting and other semantic relations. Together, our results suggest that assertions in code or language do not provide sufficient signal to fully emulate semantic representations. We formalize ways in which ungrounded language models appear to be fundamentally limited in their ability to “understand”.
The COVID-19 Open Research Dataset (CORD-19) is a growing resource of scientific papers on COVID-19 and related historical coronavirus research. CORD-19 is designed to facilitate the development of text mining and information retrieval systems over its rich collection of metadata and structured full text papers. Since its release, CORD-19 has been downloaded over 200K times and has served as the basis of many COVID-19 text mining and discovery systems. In this article, we describe the mechanics of dataset construction, highlighting challenges and key design decisions, provide an overview of how CORD-19 has been used, and describe several shared tasks built around the dataset. We hope this resource will continue to bring together the computing community, biomedical experts, and policy makers in the search for effective treatments and management policies for COVID-19.
We develop a formal hierarchy of the expressive capacity of RNN architectures. The hierarchy is based on two formal properties: space complexity, which measures the RNN’s memory, and rational recurrence, defined as whether the recurrent update can be described by a weighted finite-state machine. We place several RNN variants within this hierarchy. For example, we prove the LSTM is not rational, which formally separates it from the related QRNN (Bradbury et al., 2016). We also show how these models’ expressive capacity is expanded by stacking multiple layers or composing them with different pooling functions. Our results build on the theory of “saturated” RNNs (Merrill, 2019). While formally extending these findings to unsaturated RNNs is left to future work, we hypothesize that the practical learnable capacity of unsaturated RNNs obeys a similar hierarchy. We provide empirical results to support this conjecture. Experimental findings from training unsaturated networks on formal languages support this conjecture.
This work attempts to explain the types of computation that neural networks can perform by relating them to automata. We first define what it means for a real-time network with bounded precision to accept a language. A measure of network memory follows from this definition. We then characterize the classes of languages acceptable by various recurrent networks, attention, and convolutional networks. We find that LSTMs function like counter machines and relate convolutional networks to the subregular hierarchy. Overall, this work attempts to increase our understanding and ability to interpret neural networks through the lens of theory. These theoretical insights help explain neural computation, as well as the relationship between neural networks and natural language grammar.
We train a diachronic long short-term memory (LSTM) part-of-speech tagger on a large corpus of American English from the 19th, 20th, and 21st centuries. We analyze the tagger’s ability to implicitly learn temporal structure between years, and the extent to which this knowledge can be transferred to date new sentences. The learned year embeddings show a strong linear correlation between their first principal component and time. We show that temporal information encoded in the model can be used to predict novel sentences’ years of composition relatively well. Comparisons to a feedforward baseline suggest that the temporal change learned by the LSTM is syntactic rather than purely lexical. Thus, our results suggest that our tagger is implicitly learning to model syntactic change in American English over the course of the 19th, 20th, and early 21st centuries.
Neural network architectures have been augmented with differentiable stacks in order to introduce a bias toward learning hierarchy-sensitive regularities. It has, however, proven difficult to assess the degree to which such a bias is effective, as the operation of the differentiable stack is not always interpretable. In this paper, we attempt to detect the presence of latent representations of hierarchical structure through an exploration of the unsupervised learning of constituency structure. Using a technique due to Shen et al. (2018a,b), we extract syntactic trees from the pushing behavior of stack RNNs trained on language modeling and classification objectives. We find that our models produce parses that reflect natural language syntactic constituencies, demonstrating that stack RNNs do indeed infer linguistically relevant hierarchical structure.
We present a graph-based Tree Adjoining Grammar (TAG) parser that uses BiLSTMs, highway connections, and character-level CNNs. Our best end-to-end parser, which jointly performs supertagging, POS tagging, and parsing, outperforms the previously reported best results by more than 2.2 LAS and UAS points. The graph-based parsing architecture allows for global inference and rich feature representations for TAG parsing, alleviating the fundamental trade-off between transition-based and graph-based parsing systems. We also demonstrate that the proposed parser achieves state-of-the-art performance in the downstream tasks of Parsing Evaluation using Textual Entailments (PETE) and Unbounded Dependency Recovery. This provides further support for the claim that TAG is a viable formalism for problems that require rich structural analysis of sentences.
This paper analyzes the behavior of stack-augmented recurrent neural network (RNN) models. Due to the architectural similarity between stack RNNs and pushdown transducers, we train stack RNN models on a number of tasks, including string reversal, context-free language modelling, and cumulative XOR evaluation. Examining the behavior of our networks, we show that stack-augmented RNNs can discover intuitive stack-based strategies for solving our tasks. However, stack RNNs are more difficult to train than classical architectures such as LSTMs. Rather than employ stack-based strategies, more complex networks often find approximate solutions by using the stack as unstructured memory.
|
s3://commoncrawl/crawl-data/CC-MAIN-2023-14/segments/1679296945168.36/warc/CC-MAIN-20230323132026-20230323162026-00432.warc.gz
|
CC-MAIN-2023-14
| 12,965 | 13 |
https://www.br.freelancer.com/projects/php/structuring-website-integrating-with
|
code
|
Structuring my website and integrating it with that of my supplier, through an API.
Pago na entrega
I'm looking for a freelancer who can help me structure my website and integrate it with my vendor's website via an API.
so you will have to create the in wordpress, or c panel...
- I already have an API key from my provider.
- I expect the project to be completed in 3-4 weeks.
Skills and Experience:
- Experience with structuring and integrating websites
- Proficiency in working with APIs
- Strong programming and coding skills
- Strong problem solving skills
- Good communication and collaboration skills
If you have the necessary skills and experience, please submit your proposal with examples of similar projects you have completed.
sample website of what I want: [login to view URL]
ID do Projeto: #36821566
Sobre o projeto
10 freelancers estão ofertando em média $56 nesse trabalho
Hhi I am experienced in this and I can start right now but i have few doubts and questions lets have a quick chat and get it started waiting for your replyyy
Hello, I am professional on WordPress, let me know if you need developments in your websites. I am also experienced in Monte Carlo simulations especially in medical physics areas.
|
s3://commoncrawl/crawl-data/CC-MAIN-2024-10/segments/1707947474650.85/warc/CC-MAIN-20240226030734-20240226060734-00562.warc.gz
|
CC-MAIN-2024-10
| 1,229 | 19 |
https://community.letsencrypt.org/t/multiple-conf-confusion-ssl-labs-grade-b-help-wanted/156794
|
code
|
Please fill out the fields below so we can help you better. Note: you must provide your domain name to get help. Domain names for issued certificates are all made public in Certificate Transparency logs (e.g. crt.sh | example.com), so withholding your domain name here does not increase secrecy, but only makes it harder for us to provide help.
My domain is: tykaiwy.com
I ran this command: SSL Server Test
It produced this output: This server supports TLS 1.0 and TLS 1.1. Grade capped to B.
My web server is (include version): nginx
The operating system my web server runs on is (include version): Ubuntu 20.10 server
My hosting provider, if applicable, is:
I can login to a root shell on my machine (yes or no, or I don't know): yes
I'm using a control panel to manage my site (no, or provide the name and version of the control panel): no
The version of my client is (e.g. output of
certbot --version or
certbot-auto --version if you're using Certbot): 1.17.0
Trying to understand how I can restrict my server to just TLS 1.2 and TLS 1.3. There are multiple files that seem to play a role in SSL configuration, so I'm not sure what needs updated. I have server.conf, options-ssl-nginx.conf, nginx.conf, and ssl-params.conf all of which mention SSL to some degree. nginx -t gave messages duplicate value "TLSv1.3" in /etc/nginx/snippets/ssl-params.conf:1 and "ssl_ciphers" directive is duplicate in /etc/nginx/snippets/ssl-params.conf:2 but commented out what appeared to be duplicate entries. Now nginx -t gives ok and successful but I still get the B from SSL labs.
|
s3://commoncrawl/crawl-data/CC-MAIN-2022-27/segments/1656103036363.5/warc/CC-MAIN-20220626010644-20220626040644-00396.warc.gz
|
CC-MAIN-2022-27
| 1,570 | 13 |
https://support.regal.io/hc/en-us/articles/5724713528219-How-to-Run-an-A-B-Test-in-a-Journey
|
code
|
Each A/B Test node allows up to 5 variants, but you can always string multiple A/B Test nodes together if you need more.
A/B testing is a very powerful tool. Some ideas for what you can test:
- Hold outs
- SMS campaign copy
- Timing of calls or messages
- Call scripts
When a contact passes through an A/B test, Regal fires a contact.experiment.assigned event that includes the name of the event and the variant the contact received so that you can track the performance of your A/B test outside of Regal as well. The payload for that event is described here: https://developer.regal.io/docs/reporting-webhooks#contactexperimentassigned
You can also use the A/B test node to split traffic without running an actual experiment. To avoid sending the contact.experiment.assigned event for this use case, check the checkbox "Do Not Trigger Experiment Assigned Event." This will prevent traffic splits from displaying as experiments in the Experiments Dashboard.
Additionally, there is an Experiments Dashboard in the Reporting tab in Regal that makes it easy to see how your experiment are performing.
|
s3://commoncrawl/crawl-data/CC-MAIN-2024-10/segments/1707947473558.16/warc/CC-MAIN-20240221202132-20240221232132-00562.warc.gz
|
CC-MAIN-2024-10
| 1,097 | 9 |
https://softwarecentral.com/
|
code
|
SoftwareCentral is one of the most efficient and user-friendly front-end tools for the Microsoft Endpoint Configuration Manager (MECM/SCCM)
From your SCCM to Intune with just a few clicks
In the most resent version of SoftwareCentral, we are focusing a lot on CO-Management with Microsoft Endpoint Manager and therefore we have added new features for managing Microsoft Intune in SoftwareCentral.
Guide: License management with SoftwareCentral
SoftwareCentral includes many smart features that make the management of an SCCM environment easier. One of those features are the license management tools License Control and License Harvesting
Guide: Cut cost, use licenses more efficiently
Licenses Harvesting is a simple way to save money on licenses, License Harvesting monitors the usage of software and can automatically uninstall software that is inactive on a client computer, that way the software license can be used for redeployment on another client computer.
|
s3://commoncrawl/crawl-data/CC-MAIN-2022-21/segments/1652662555558.23/warc/CC-MAIN-20220523041156-20220523071156-00510.warc.gz
|
CC-MAIN-2022-21
| 965 | 7 |
https://adminalley.com/page/3/
|
code
|
I strongly recommend against this. Many applications communicate with directory services through LDAP, but the LDAP Request for Comments (RFC) specification stipulates that an LDAP
In 2008, this is a doddle. Add the user to the built in Event Viewers group! In 2003 however, there’s a bit more to it.
Problem GENERAL ERROR: When attempting to expire a media in Veritas NetBackup ™ 6.5 with the bpexpdate command, the message “requested media id was not
It is very common for posters on newgroups to recommend that solution seekers run the “aspnet_regiis -i“ command to fix various problems and ASP.NET maladies.
In IIS6, this was a pain area. On a 64-bit Windows 2003 Server, you cannot run worker processes in both 32-bit mode and as well
Server 2008 and/or R2 On the Windows 2008 disk partition install, I get the error “Windows could not prepare the partition selected for installation”. Boot
|
s3://commoncrawl/crawl-data/CC-MAIN-2019-13/segments/1552912201455.20/warc/CC-MAIN-20190318152343-20190318174343-00342.warc.gz
|
CC-MAIN-2019-13
| 909 | 6 |
http://www.ohloh.net/tags/find
|
code
|
Select a tag to browse associated projects and drill deeper into the tag cloud.
BlueSearch is a communication system developed in Java, using Microedition version. Bluetooth technology is used to transfer information between mobiles devices to matching profiles. Today, BlueSearch is being developed in anothers platforms too, like BlackBerry and Android (new Google Device ... [More]
Caditor is an open source text editor built for those on the go or have limited disk space. The app includes features ranging from syntax highlighting to line numbering to text encryption.
This emacs package lets the user launch a search, with a single key combination, for the word under the cursor, in the current repository, or the current directory, or in the set of currently open files.
Yet another LAN indexing solution. Indexes SMB and FTP. Has web interface. Written in Java. GPL. Leveraging Lucene's scoring system, the PUNKSearch returns results (to the end user) sorted by relevance like all modern web search engines do. Also it exploits a set of heuristics for efficient ... [More]
This emacs package manages multiple search results buffers: - the search results of grep, lgrep, rgrep, and find-grep are sent to separate buffers instead of overwriting the contents of a single buffer. - several navigation functions are provided to allow the user to treat the search results ... [More]
Find for Inferno lets you traverse directory trees and print paths and information about a file. It is intended to be easier to use than alphabet-fs(1) and/or du(1) with various combinations of grep, sed and for-loops with ftest(1). Find's usage is different from the unix version (most importantly, it uses standard option syntax).
Copyright © 2013 Black Duck Software, Inc. and its contributors, Some Rights Reserved. Unless otherwise marked, this work is licensed under a Creative Commons Attribution 3.0 Unported License . Ohloh ® and the Ohloh logo are trademarks of Black Duck Software, Inc. in the United States and/or other jurisdictions. All other trademarks are the property of their respective holders.
|
s3://commoncrawl/crawl-data/CC-MAIN-2013-48/segments/1387345775611/warc/CC-MAIN-20131218054935-00019-ip-10-33-133-15.ec2.internal.warc.gz
|
CC-MAIN-2013-48
| 2,100 | 8 |
https://www.lewas.centers.vt.edu/dataviewer/components.html
|
code
|
(Last edited on Oct. 8, 2015)
Some key components of the OWLS are as follows:
The site map shows the possible means of navigation through this learning system and provides quick access to various components of the system.
The overhead view provides a spatial reference for key components of the watershed including the watershed boundary and waterways and the location of the LEWAS measurement site.
Live LEWAS Data
The Live LEWAS Data page provides an interactive graphical interface to compare various parameters measured by the LEWAS. It also contains time-stamped imagery that corresponds to the data plotted. The plotted data can be copied from the data download page (when available).
Case studies describe both key types of events that happen in the watershed and key aspects of the LEWAS data collection system. Several case studies can be found on the home screen.
The live camera provides visual reference of conditions at the field site.
The weather radar provides a visual precipitation reference.
The glossary describes many key terms related to this system and watershed monitoring in general.
|
s3://commoncrawl/crawl-data/CC-MAIN-2018-26/segments/1529267863109.60/warc/CC-MAIN-20180619173519-20180619193519-00356.warc.gz
|
CC-MAIN-2018-26
| 1,107 | 10 |
https://www.collegeroadanimalhospital.com/staff/bio/rachel-r
|
code
|
Meet Our Staff
Rachel is from New Jersey and went to school in Alaska. She moved to Wilmington for warmer weather and to continue her education. Since the age of three, Rachel has wanted to be a veterinarian. She has worked with animals from the time she could walk and also helped out on her family's farm. Rachel's rat pack of rescues include: Gabe the feisty chi, Gatsby the three legged wonder dog, T-Yoda the engine caught feline, Foxy the prancing pony, Cernunnos the king of pythons, and Fergus the spicy bearded dragon. She joined the College Road Animal Hospital family in September of 2013 and has held positions as a kennel attendant, veterinary technician, client care specialist, and now inventory supervisor.
|
s3://commoncrawl/crawl-data/CC-MAIN-2019-43/segments/1570987769323.92/warc/CC-MAIN-20191021093533-20191021121033-00284.warc.gz
|
CC-MAIN-2019-43
| 722 | 2 |
https://www.ksat.com/news/2017/02/02/texas-boy-finds-rattlesnake-in-toilet-23-more-living-under-home/
|
code
|
BUFFALO GAP, Texas – A Texas family was in for quite a surprise earlier this week when a snake removal company found 24 rattlesnakes living under their home.
Yes, we said 24 and no, we are not joking.
According to the Big Country Snake Removal Company's Facebook post, the company was called out to a home after a snake crawled up through the toilet.
Isac Mcfadden found the snake and went to get his mother, CNN reports.
"I found this big clump, and I knew it was a snake," Isac said.
The family killed the snake and called Big Country Snake Removal.
When the snake remover arrived, he found 13 snakes in the cellar and 10 more underneath the home.
“How is this possible? It's actually quite simple; rattlesnakes are secretive and can be very cryptic. They rely heavily on their camouflage. This is simply how they survive. Just because you don't see them doesn't mean they aren't there,” the Facebook post read.
The family told the company it was the first time they’d seen snakes on their property in years. Now the business is offering discounts on inspections through the end of February.
|
s3://commoncrawl/crawl-data/CC-MAIN-2021-10/segments/1614178361510.12/warc/CC-MAIN-20210228145113-20210228175113-00225.warc.gz
|
CC-MAIN-2021-10
| 1,101 | 9 |
https://excelfornoobs.com/tips.php?title=How%20to%20Create%20a%20Custom%20List%20in%20Excel%20Tutorial
|
code
|
How to Create a Custom List in Excel Tutorial
If you type “jan” in one cell and use fill handle and drag down, the subsequent cells will be filled with feb, mar … so on. Similarly, one cell B1 below has “Sunday” and if we use fill handle Monday comes into the next cell. Where does the sequence comes from in excel.
Using fill handle to get next value from custom list
If we write one of the employee name and use fill handle, can we get the remaining employee names in the subsequent cells?
Yes. It is possible by using “Custom Lists” in excel.
In excel 2013, Custom list can be accessed from FILE > Options (from left pane)
Select “Advanced” in the Excel Options box left pane. Scroll down through right scroll bar and go to “General” section where we can find “Edit Custom Lists….”
Select the “Edit Custom Lists….”
From the above box, custom lists can be created in two ways.
To add one by one, we can select “NEW LIST” from “Custom lists” and type the values to be added into “List entries”. (can be separated by giving enter after each entry of by comma separation)
Then select “Add”
The second way is to use the list in work sheet by “Import”
Select the list in worksheet by using the import list option and click “Import”. OK.
Now, when one of the value is written in a cell and use fill handle and drag down.
The next value in the list will come into the next cell.
Note: After the last value in the list first value will be repeated. (values list is cyclic )
- Typing the values one by one (in new line or comma separated)
- Selecting the data list which is there in a worksheet.
|
s3://commoncrawl/crawl-data/CC-MAIN-2020-16/segments/1585370528224.61/warc/CC-MAIN-20200405022138-20200405052138-00078.warc.gz
|
CC-MAIN-2020-16
| 1,646 | 18 |
https://www.genomeweb.com/resources/jobs/technical-outreach-officer
|
code
|
We are looking to recruit a Technical Outreach Officer for Europe PMC, within the Literature Services team at the European Bioinformatics Institute (EMBL-EBI), located on the Wellcome Trust Genome Campus near Cambridge in the UK.
Europe PMC is Europe's largest repository of freely available life sciences research articles, containing over 30 million biomedical research full text articles and abstracts. Europe PMC is funded by 29 funders across Europe, coordinated by the Wellcome Trust, and is the primary repository that supports their Open Access policies.
A key part of our mission is to make content and metadata available through APIs and bulk downloads, for consumption by application developers, data miners, hackers and researchers.
Your roleAs the Technical Outreach Officer, your role will be to build community among this growing programmatic access user base. You will be the main contact point for our API users and will represent Europe PMC at workshops, hackathons and other developer forums.
To fulfil this role, you will need to have good insight into the needs of developers and API users, probably from using APIs and developing apps yourself. You should have a strong interest in open data in application development or text mining workflows. As you will need to attend, deliver and organise events such as webinars and hackathons, you will need good communication (written and verbal) and organisational skills alongside your developer skills. You will be supported in the role by the Community Manager and our core developer team.
This is a challenging position that would suit an open data enthusiast. You will find yourself working on an exciting project in a dynamic team in an international environment. If you think you can fulfil the role, then we would really like to hear from you.
|
s3://commoncrawl/crawl-data/CC-MAIN-2018-51/segments/1544376828318.79/warc/CC-MAIN-20181217042727-20181217064727-00327.warc.gz
|
CC-MAIN-2018-51
| 1,815 | 6 |
https://thwack.solarwinds.com/thread/99633
|
code
|
I am evaluating IPAM at the moment. I have been importing all my IP's through the use of the CSV Import template.
Would someone please explain what these columns are. I have an idea, but would like to make sure I understand them as clearly as possible.
1: Node Expunge Interval
2: Retain User Data
3: Scan Interval
4: Neighbor Scan Interval
I am not certain about #1 or #2.
I assume #3 is how often IPAM does an ICMP/SNMP sweep of the subnet, and #4 is how often IPAM checks the ARP tables of the router specified for that subnet.
What settings do most people use for subnets that dont change very often (server subnets) and subnets which might have a lot of change in them (wireless guest subnets).
1: Node Expunge Interval - Don't know
2: Retain User Data - Flag which controls "Update but not erase manually entered data" 1 for TURE, 0 for FALSE
3: Scan Interval - IPAM interval for scanning this subnet
4: Neighbor Scan Interval - IPAM interval for doing neighbor scans related to this subnet
|
s3://commoncrawl/crawl-data/CC-MAIN-2019-30/segments/1563195526359.16/warc/CC-MAIN-20190719202605-20190719224605-00379.warc.gz
|
CC-MAIN-2019-30
| 996 | 13 |
https://forum.gameznetwork.com/threads/current-season-end-date.246321/
|
code
|
Discussion in 'Questions' started by Archangeli, Nov 24, 2023.
Does anyone knows when will the current season end? I can't find that info anywhere.
Only season ends when there is a major update, so far there is no news about it
I think they forgot the game or cant crack the updates
From what I've heard getting the updates have been rather annoying but possibly something coming in late December or January. Not confirmed by Dev but just what I've heard circulating in server chat and such.
It is indeed true, that they work with some degree of secrecy, and that can be really annoying for players, who would like to schedule their gaming time. However as a whole this server has been run far better all others, and it is due to the hard work and effort these individuals put into it. So once again in grand and many thanks, to ALL the people involved in making this a great place to game; Thanks GameZ-network staff and helpers.
I cannot defend nor attack the administration, for choosing to be silent with regards to when things are going public. But as far as I have learned, it takes a while for codes to be compiled and error corrected, packing and unpacking things takes time as well, making small tweaks to the geo-engine here and there, to fit the rest of the game. Now I might be talking out my arse right now, but I think that they have made the entire server backwards compatible. Which is why your characters can arrive, back from the first release of this game server, and find that the questline they started some 5 years ago, is still doable and can be completed.
I am not a great voice for this community, as my "bullying the bullies" approach, has been quite wrongly motivated in the past. I have tried my best not to stir up too much trouble, and been as civil as *I* can be under the current circumstances. But one thing is for sure; This is by far the best server of its kind that you will find out there.
" This is by far the best server of its kind that you will find out there. " Absolutely true. But It's been toooooooooo long since the previous big update. It could be nice to see some news about development, what are they doing, what are they planning, how is the update, etc.
Hello there and welcome to the GameZ-network forum community. I am NOT a part of the staff, but I do like helping out, when I am able to.
The last time this was done, there was little to no difference in how many were complaining, and in general not giving any understanding nor empathy, to the people who actually do the work. So back in the days, the staff of GamezBD did have a nice page, where they would post their current job list, as well as future projects. Nothing really changed in the "bad behaviour" of Some people, and because it started making even more noise, and debates as well as outright fights, the staff decided that it didn't work.
Actually I have NO real information, and this is just what I suspect happened, because they had it once, and then they didn't.
|
s3://commoncrawl/crawl-data/CC-MAIN-2024-10/segments/1707947476374.40/warc/CC-MAIN-20240303111005-20240303141005-00694.warc.gz
|
CC-MAIN-2024-10
| 2,985 | 12 |
https://book.douban.com/subject/2262611/
|
code
|
Let's get ready for school ---- wipe clean--- tell the time
The exercises and activities in this book make telling the time fun and simple to learn - the ideal preparation for school.
Using the special pen included, telling the time can be practised again and agian - simply wipe away with a cloth!
length: (cm)27..6 width:(cm)26.5
订阅关于Wipe Clean Tell the Time的评论:
feed: rss 2.0
|
s3://commoncrawl/crawl-data/CC-MAIN-2017-30/segments/1500549423764.11/warc/CC-MAIN-20170721082219-20170721102219-00170.warc.gz
|
CC-MAIN-2017-30
| 392 | 6 |
https://www.catalyzex.com/author/Hassan%20Ghasemzadeh
|
code
|
Medications often impose temporal constraints on everyday patient activity. Violations of such medical temporal constraints (MTCs) lead to a lack of treatment adherence, in addition to poor health outcomes and increased healthcare expenses. These MTCs are found in drug usage guidelines (DUGs) in both patient education materials and clinical texts. Computationally representing MTCs in DUGs will advance patient-centric healthcare applications by helping to define safe patient activity patterns. We define a novel taxonomy of MTCs found in DUGs and develop a novel context-free grammar (CFG) based model to computationally represent MTCs from unstructured DUGs. Additionally, we release three new datasets with a combined total of N = 836 DUGs labeled with normalized MTCs. We develop an in-context learning (ICL) solution for automatically extracting and normalizing MTCs found in DUGs, achieving an average F1 score of 0.62 across all datasets. Finally, we rigorously investigate ICL model performance against a baseline model, across datasets and MTC types, and through in-depth error analysis.
Prescription medications often impose temporal constraints on regular health behaviors (RHBs) of patients, e.g., eating before taking medication. Violations of such medical temporal constraints (MTCs) can result in adverse effects. Detecting and predicting such violations before they occur can help alert the patient. We formulate the problem of modeling MTCs and develop a proof-of-concept solution, ActSafe, to predict violations of MTCs well ahead of time. ActSafe utilizes a context-free grammar based approach for extracting and mapping MTCs from patient education materials. It also addresses the challenges of accurately predicting RHBs central to MTCs (e.g., medication intake). Our novel behavior prediction model, HERBERT , utilizes a basis vectorization of time series that is generalizable across temporal scale and duration of behaviors, explicitly capturing the dependency between temporally collocated behaviors. Based on evaluation using a real-world RHB dataset collected from 28 patients in uncontrolled environments, HERBERT outperforms baseline models with an average of 51% reduction in root mean square error. Based on an evaluation involving patients with chronic conditions, ActSafe can predict MTC violations a day ahead of time with an average F1 score of 0.86.
Stress detection and monitoring is an active area of research with important implications for the personal, professional, and social health of an individual. Current approaches for affective state classification use traditional machine learning algorithms with features computed from multiple sensor modalities. These methods are data-intensive and rely on hand-crafted features which impede the practical applicability of these sensor systems in daily lives. To overcome these shortcomings, we propose a novel Convolutional Neural Network (CNN) based stress detection and classification framework without any feature computation using data from only one sensor modality. Our method is competitive and outperforms current state-of-the-art techniques and achieves a classification accuracy of $92.85\%$ and an $f1$ score of $0.89$. Through our leave-one-subject-out analysis, we also show the importance of personalizing stress models.
Inter-beat interval (IBI) measurement enables estimation of heart-rate variability (HRV) which, in turns, can provide early indication of potential cardiovascular diseases. However, extracting IBIs from noisy signals is challenging since the morphology of the signal is distorted in the presence of the noise. Electrocardiogram (ECG) of a person in heavy motion is highly corrupted with noise, known as motion-artifact, and IBI extracted from it is inaccurate. As a part of remote health monitoring and wearable system development, denoising ECG signals and estimating IBIs correctly from them have become an emerging topic among signal-processing researchers. Apart from conventional methods, deep-learning techniques have been successfully used in signal denoising recently, and diagnosis process has become easier, leading to accuracy levels that were previously unachievable. We propose a deep-learning approach leveraging tiramisu autoencoder model to suppress motion-artifact noise and make the R-peaks of the ECG signal prominent even in the presence of high-intensity motion. After denoising, IBIs are estimated more accurately expediting diagnosis tasks. Results illustrate that our method enables IBI estimation from noisy ECG signals with SNR up to -30dB with average root mean square error (RMSE) of 13 milliseconds for estimated IBIs. At this noise level, our error percentage remains below 8% and outperforms other state of the art techniques.
This paper takes initial strides at designing and evaluating a vision-based system for privacy ensured activity monitoring. The proposed technology utilizing Artificial Intelligence (AI)-empowered proactive systems offering continuous monitoring, behavioral analysis, and modeling of human activities. To this end, this paper presents Single Run Action Detector (S-RAD) which is a real-time privacy-preserving action detector that performs end-to-end action localization and classification. It is based on Faster-RCNN combined with temporal shift modeling and segment based sampling to capture the human actions. Results on UCF-Sports and UR Fall dataset present comparable accuracy to State-of-the-Art approaches with significantly lower model size and computation demand and the ability for real-time execution on edge embedded device (e.g. Nvidia Jetson Xavier).
Continual (sequential) training and multitask (simultaneous) training are often attempting to solve the same overall objective: to find a solution that performs well on all considered tasks. The main difference is in the training regimes, where continual learning can only have access to one task at a time, which for neural networks typically leads to catastrophic forgetting. That is, the solution found for a subsequent task does not perform well on the previous ones anymore. However, the relationship between the different minima that the two training regimes arrive at is not well understood. What sets them apart? Is there a local structure that could explain the difference in performance achieved by the two different schemes? Motivated by recent work showing that different minima of the same task are typically connected by very simple curves of low error, we investigate whether multitask and continual solutions are similarly connected. We empirically find that indeed such connectivity can be reliably achieved and, more interestingly, it can be done by a linear path, conditioned on having the same initialization for both. We thoroughly analyze this observation and discuss its significance for the continual learning process. Furthermore, we exploit this finding to propose an effective algorithm that constrains the sequentially learned minima to behave as the multitask solution. We show that our method outperforms several state of the art continual learning algorithms on various vision benchmarks.
While activity recognition from inertial sensors holds potential for mobile health, differences in sensing platforms and user movement patterns cause performance degradation. Aiming to address these challenges, we propose a transfer learning framework, TransFall, for sensor-based activity recognition. TransFall's design contains a two-tier data transformation, a label estimation layer, and a model generation layer to recognize activities for the new scenario. We validate TransFall analytically and empirically.
There exist significant gaps in research about how to design efficient in-bed lying posture tracking systems. These gaps can be articulated through several research questions as follows. First, can we design a single-sensor, pervasive, and inexpensive system that can accurately detect lying postures? Second, what computational models are most effective in the accurate detection of lying postures? Finally, what physical configuration of the sensor system is most effective for lying posture tracking? To answer these important research questions, in this article, we propose a comprehensive approach to design a sensor system that uses a single accelerometer along with machine learning algorithms for in-bed lying posture classification. We design two categories of machine learning algorithms based on deep learning and traditional classification with handcrafted features to detect lying postures. We also investigate what wearing sites are most effective in accurate detection of lying postures. We extensively evaluate the performance of the proposed algorithms on nine different body locations and four human lying postures using two datasets. Our results show that a system with a single accelerometer can be used with either deep learning or traditional classifiers to accurately detect lying postures. The best models in our approach achieve an F-Score that ranges from 95.2% to 97.8% with 0.03 to 0.05 coefficient of variation. The results also identify the thighs and chest as the most salient body sites for lying posture tracking. Our findings in this article suggest that because accelerometers are ubiquitous and inexpensive sensors, they can be a viable source of information for pervasive monitoring of in-bed postures.
Catastrophic forgetting affects the training of neural networks, limiting their ability to learn multiple tasks sequentially. From the perspective of the well established plasticity-stability dilemma, neural networks tend to be overly plastic, lacking the stability necessary to prevent the forgetting of previous knowledge, which means that as learning progresses, networks tend to forget previously seen tasks. This phenomenon coined in the continual learning literature, has attracted much attention lately, and several families of approaches have been proposed with different degrees of success. However, there has been limited prior work extensively analyzing the impact that different training regimes -- learning rate, batch size, regularization method-- can have on forgetting. In this work, we depart from the typical approach of altering the learning algorithm to improve stability. Instead, we hypothesize that the geometrical properties of the local minima found for each task play an important role in the overall degree of forgetting. In particular, we study the effect of dropout, learning rate decay, and batch size, on forming training regimes that widen the tasks' local minima and consequently, on helping it not to forget catastrophically. Our study provides practical insights to improve stability via simple yet effective techniques that outperform alternative baselines.
In recent years, neural networks have demonstrated an outstanding ability to achieve complex learning tasks across various domains. However, they suffer from the "catastrophic forgetting" problem when they face a sequence of learning tasks, where they forget the old ones as they learn new tasks. This problem is also highly related to the "stability-plasticity dilemma". The more plastic the network, the easier it can learn new tasks, but the faster it also forgets previous ones. Conversely, a stable network cannot learn new tasks as fast as a very plastic network. However, it is more reliable to preserve the knowledge it has learned from the previous tasks. Several solutions have been proposed to overcome the forgetting problem by making the neural network parameters more stable, and some of them have mentioned the significance of dropout in continual learning. However, their relationship has not been sufficiently studied yet. In this paper, we investigate this relationship and show that a stable network with dropout learns a gating mechanism such that for different tasks, different paths of the network are active. Our experiments show that the stability achieved by this implicit gating plays a very critical role in leading to performance comparable to or better than other involved continual learning algorithms to overcome catastrophic forgetting.
|
s3://commoncrawl/crawl-data/CC-MAIN-2023-40/segments/1695233511406.34/warc/CC-MAIN-20231004184208-20231004214208-00198.warc.gz
|
CC-MAIN-2023-40
| 12,186 | 10 |
https://about.gitlab.com/releases/
|
code
|
GitLab has been releasing on the 22nd of the month for the last 106 months straight! For a list of release posts including patch releases, please check the blog category releases. Future releases, and their important features, can be found on our upcoming releases page. You can also view upcoming features by product tier.
Information about recent features added in our official GitLab releases can be found on the GitLab project releases page.
Please check out the changelog to see all the named changes:
Changes prior to GitLab 8.0 can be seen here
|
s3://commoncrawl/crawl-data/CC-MAIN-2020-34/segments/1596439735963.64/warc/CC-MAIN-20200805153603-20200805183603-00355.warc.gz
|
CC-MAIN-2020-34
| 551 | 4 |
https://docs.intersystems.com/healthconnectlatest/csp/docbook/stubcanonicalbaseurl/csp/docbook/DocBook.UI.Page.cls?KEY=D2RMDX_FirstSibling
|
code
|
Returns the first sibling of the given member.
This function returns a member.
Syntax and Details
FeedbackOpens in a new tab
member_expression is an expression that returns a member.
This expression cannot refer to a measure.
This function examines all the children of the parent of the given member, and returns the first member of that set (considering the default order of that set).
This function can return the same member that you specify as an argument (if that member is the first sibling).
SELECT MEASURES.[%COUNT] ON 0, birthd.[Mar 2003].FIRSTSIBLING ON 1 FROM patients Patient Count Jan-2003 10
|
s3://commoncrawl/crawl-data/CC-MAIN-2023-23/segments/1685224646652.16/warc/CC-MAIN-20230610233020-20230611023020-00355.warc.gz
|
CC-MAIN-2023-23
| 605 | 9 |
https://lists.bisq.network/pipermail/bisq-github/2020-June/035514.html
|
code
|
[bisq-network/proposals] Proposal to increase trade deposit to 30% (#233)
notifications at github.com
Fri Jun 26 08:07:32 UTC 2020
> Okay, I'll sell you Bitcoin, wire the money to my bank account 1234567. Then after I get your bank wire I just wait 100 days and I get my Bitcoin back. Since there is no Arbitration you have no way to prevent this.
You are right.
Then Mutually Assured Destruction: Either trade is completed or funds are burned.
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the bisq-github
|
s3://commoncrawl/crawl-data/CC-MAIN-2020-40/segments/1600402124756.81/warc/CC-MAIN-20201001062039-20201001092039-00670.warc.gz
|
CC-MAIN-2020-40
| 675 | 11 |
http://stackoverflow.com/questions/1107689/connecting-to-sql-server-with-cl-sql-via-unixodbc-freetds
|
code
|
I've managed to connect from SBCL running on debian to an SQL Server 2000 instance over the network using FreeTDS/unixODBC.
I can actually get data back from the server, so all is working.
However, many of the columns trigger what seem to be unsupported data types a-la:
The value 2147483647 is not of type FIXNUM.
-11 fell through ECASE expression. Wanted one of (-7 -6 -2 -3 -4 93 92 91 11 10 ...).
Anyone have experience using CLSQL with SQL Server would be able to help out?
|
s3://commoncrawl/crawl-data/CC-MAIN-2016-07/segments/1454701157075.54/warc/CC-MAIN-20160205193917-00219-ip-10-236-182-209.ec2.internal.warc.gz
|
CC-MAIN-2016-07
| 478 | 6 |
https://itl.nist.gov/div898/software/dataplot/refman2/auxillar/npv.htm
|
code
|
NEGATIVE PREDICTIVE VALUEName:
The parameters N11, N12, N21, and N22 denote the counts for each category.
Success and failure can denote any binary response. Dataplot expects "success" to be coded as "1" and "failure" to be coded as "0". Some typical examples would be:
In these examples, the "ground truth" is typically given as variable 1 while some estimator of the ground truth is given as variable 2.
The negative predictive value is then N12/(N12+N22). This is the conditional probability of variable 1 being false given that variable 2 is false. In the context of the first example above, this is the probability that the disease is not present when there is a negative test result.
Fleiss and his co-authors recommend negative predictive value and positive predictive value as an alternative to false positive and false negative due to the fact that the definitions of false positive and false negative have been inconsistent in the literature. Syntax:
where <y1> is the first response variable;
<y2> is the second response variable;
<par> is a parameter where the computed negative predictive value is stored;
and where the <SUBSET/EXCEPT/FOR qualification> is optional.
LET A = NEGATIVE PREDICTIVE VALUE Y1 Y2 SUBSET TAG > 2
CROSS TABULATE FALSE POSITIVE Y1 Y2 X1 X2
FALSE POSITIVES PLOT Y1 Y2 X
CROSS TABULATE FALSE POSITIVES PLOT Y1 Y2 X1 X2
BOOTSTRAP FALSE POSITIVES PLOT Y1 Y2
JACKNIFE FALSE POSITIVES PLOT Y1 Y2
let n = 1 . let p = 0.2 let y1 = binomial rand numb for i = 1 1 100 let p = 0.1 let y2 = binomial rand numb for i = 1 1 100 . let p = 0.4 let y1 = binomial rand numb for i = 101 1 200 let p = 0.08 let y2 = binomial rand numb for i = 101 1 200 . let p = 0.15 let y1 = binomial rand numb for i = 201 1 300 let p = 0.18 let y2 = binomial rand numb for i = 201 1 300 . let p = 0.6 let y1 = binomial rand numb for i = 301 1 400 let p = 0.45 let y2 = binomial rand numb for i = 301 1 400 . let p = 0.3 let y1 = binomial rand numb for i = 401 1 500 let p = 0.1 let y2 = binomial rand numb for i = 401 1 500 . let x = sequence 1 100 1 5 . let a = negative predictive value y1 y2 subset x = 1 tabulate negative predictive value y1 y2 x . label case asis xlimits 1 5 major xtic mark number 5 minor xtic mark number 0 xtic mark offset 0.5 0.5 ytic mark offset 0.05 0.05 y1label Negative Predictive Value x1label Group ID character x blank line blank solid . negative predictive value plot y1 y2 x
|
s3://commoncrawl/crawl-data/CC-MAIN-2021-49/segments/1637964358078.2/warc/CC-MAIN-20211127013935-20211127043935-00417.warc.gz
|
CC-MAIN-2021-49
| 2,412 | 17 |
https://www.mrspennington.com/the-number-one-problem-you-should-ask-for-data-science/
|
code
|
Data science includes working with large portions of knowledge units. Chances are high you’ll need to be acquainted with Machine learning. The aim is: learning knowledge science is an efficient fast-term funding, for constructive. It might be talked about that after receiving an online diploma in data science, one may be considered a advisor who can info firms and stakeholders on how they’re purported to act on findings from data.
Your personal mentor is your confederate-in-knowledge science. They’re program managers, instructors, and industry professionals dedicated to your future success. Chantal Achterberg is a two-time Olympic Medalist and a SASÂ Certified Knowledge Scientist. Get impressed with this temporary check out Chantal’s path from Olympian, to scholar of the SAS Academy for Data Science, to her dream job.
Your best program for occupation transformation. Joe Brown, the publication’s editor-in-chief, acknowledged that when he started there in August 2016 (after stops at Wired, Gizmodo and an earlier stint at Pop Sci), he got here with a mission of constructing the Bonnier-owned journal more inclusive, as its identify suggests. That methodology prolonged to Brown acknowledging local weather change skeptics early on.
Inside BOI, I’m the skilled in qualitative research.
You will examine the basics of utilizing Python gathering and presenting data to search out actual insights from info you might have, SQL for querying, as properly the best way to design and run your own experiments to assemble new knowledge. Because of close researcher involvement, the researcher good points an insider’s view of the sphere. This enables the researcher to seek out points which are often missed (resembling subtleties and complexities) by the scientific, extra positivistic inquiries.
It is doable, and even seemingly, that not all the people whose expertise you need will report again to the data science employees supervisor. Design talent may report to advertising; subject-matter consultants is also executives reporting to the CEO. Nonetheless, it’s vital to supply the workforce as a lot dedication-making vitality as potential. Stakeholders will most frequently be of us with enterprise expertise who are rigorously related to or liable for enterprise targets; the aim of the work, finally, is healthier business outcomes. These folks can create shared targets and incentives for the group. Ideally you’ll be able to keep away from the accountability-with out-authority trap, wherein the staff is coping with a number of stakeholders who could not all be aligned.
Residing in Barcelona for a yr was a dream. By way of this system I gained confidence in making use of information science to actual world issues and labored with gifted and thoughtful machine studying researchers. But, what made me happiest was the friendships I fostered with like-minded bold people from world wide.
|
s3://commoncrawl/crawl-data/CC-MAIN-2022-33/segments/1659882570730.59/warc/CC-MAIN-20220807211157-20220808001157-00536.warc.gz
|
CC-MAIN-2022-33
| 2,926 | 7 |
https://www.urbandictionary.com/author.php?author=Fjortis
|
code
|
Meaning the same as good night
, but without a claim to a specific time of day. Suited for modern geeks and smokers of certain herb
s to whom the concepts of day and night mean very little beyond how bright it is outside.
Pronounced go-mjew, not gom-jew.
- Looks like the sun rises, I should probably head off to bed. Gomjew.
|
s3://commoncrawl/crawl-data/CC-MAIN-2018-09/segments/1518891812978.31/warc/CC-MAIN-20180220145713-20180220165713-00276.warc.gz
|
CC-MAIN-2018-09
| 325 | 5 |
https://acn.news/canada-election-comparing-the-child-care-promises-of-the-federal-parties/
|
code
|
All three of Canada’s main parties are making pitches to parents this election, promising better child-care if elected. Both Liberals and NDP are pledging to move forward with a $10-a-day child-care plan, at a cost of $30 billion.
Conservatives say they will instead implement a refundable tax credit aimed at helping lower-income families, with a price tag of $2.6 billion over five years.
So how do their plans compare and which of them do experts think are viable? Anne Gaviola breaks down their strategies, and whether they can really help cash-strapped families.
For more info, please go to https://globalnews.ca/news/8185834/liberal-vs-conservative-child-care-canada-election/
Subscribe to Global News Channel HERE: http://bit.ly/20fcXDc
Like Global News on Facebook HERE: http://bit.ly/255GMJQ
Follow Global News on Twitter HERE: http://bit.ly/1Toz8mt
Follow Global News on Instagram HERE: https://bit.ly/2QZaZIB
#CanadaElection2021 #CdnPoli #GlobalNews
|
s3://commoncrawl/crawl-data/CC-MAIN-2021-39/segments/1631780060538.11/warc/CC-MAIN-20210928062408-20210928092408-00077.warc.gz
|
CC-MAIN-2021-39
| 962 | 9 |
https://www.gamedev.net/topic/630248-my-first-c-game/
|
code
|
Jump to content
OpenGL API Specifications
Subscribe to GameDev.net Direct to receive the latest updates and exclusive content.
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.
Posted 25 August 2012 - 12:13 PM
Posted 26 August 2012 - 07:12 AM
Posted 26 August 2012 - 09:35 AM
Edited by lride, 26 August 2012 - 09:36 AM.
Posted 26 August 2012 - 09:39 AM
Edited by Rectangle, 26 August 2012 - 09:42 AM.
Posted 26 August 2012 - 01:47 PM
This is extremely useful and I appreciate everyone's repplies, especially this one. I thank you kindly for your advice and I'll make sure to follow it!
If you are exclusively targetting Windows, I'd actually start out by learning DirectX 9. OpenGL is cross-platform and pretty standard for most OS's and devices, but even with the utility libraries it has a lot of manual labor and gets quite intricate early on for a beginner. Many concepts between them are quite similar, and I do recommend eventually learning both, but DirectX is a better starting point IMO and AFAIK is the only way that you can port your games onto the Xbox360 platform (via the XNA Framework). However, if you haven't familiarized yourself with the basic Win32 API concepts of window creation and GDI graphics, then yes I would definitely recommend starting with OpenGL + utility libs. Alternatively, there are 3rd-party libraries such as SDL which are both cross-platform and fairly high-level, making everything easier, so there's something to think about as well.
Google is your friend. Type in the name of the framework you wish to use + "tutorial" or "example" and you will have nearly everything you will ever need right there.
Matrices are a mathematical structure primarily used in 3D graphics programming, and are a college-level concept derived from linear algebra. For graphics, they aid in the position, scale, rotation and translation of an object in 3D world-space. Before getting involved in them, I would definitely recommend sticking to a 2D environment where simple high-school algebra and trigonometry come into play using 2-dimensional Vectors.
To learn the syntax, just keep checking the API reference docs, samples/demos, and online tutorials. Bookmark the ones you are learning along the way if you need to.
As stated by the previous poster, the same applies for any API you are learning. For Win32 API stuff, check the MSDN.
As suggested above, a good place to start is a simple game of Tic-Tac-Toe or Minesweeper, or a card game like Solitaire. Once you've done that, I'd move on to something that uses 2D sprites, such as Pong, Asteroids or Tetris, and then move up to Tile-based games such as Mario, Zelda or virtually any 2D side-scroller or top-down RPG. By then, you should be prepared to implement 3D effects in a 2D environment (using 3D vectors), and perhaps you could even try a 2.5D isometric game like Starcraft or Diablo. Then, it's finally time to move forward into a fully 3D environment. A good place to start there would be a 3D space-shooter.
It's a long path but if you stick with it and don't skip any steps, you won't regret it later on. Personally, I've had to go back and forth over the years, learning everything out of order, and it's complicated my 3D programming skills to say the least. Good luck!
Posted 28 August 2012 - 02:32 AM
Edited by Tordin, 28 August 2012 - 02:33 AM.
Posted 28 August 2012 - 07:56 AM
Posted 28 August 2012 - 08:28 AM
GameDev.net™, the GameDev.net logo, and GDNet™ are trademarks of GameDev.net, LLC.
|
s3://commoncrawl/crawl-data/CC-MAIN-2017-17/segments/1492917123048.37/warc/CC-MAIN-20170423031203-00440-ip-10-145-167-34.ec2.internal.warc.gz
|
CC-MAIN-2017-17
| 3,605 | 24 |
http://www.yang-central.org/twiki/bin/view/Main/WhyYang
|
code
|
YANG is easy and targets the right audience
YANG is simple and easy to learn. The syntax is trivial
and regular. The hierarchy is clean, the role of elements is
obvious, and the use of braces for containment is consistent.
YANG stresses readability, but still gives sufficient
flexibility, extensibility, and power to allow it to be useful
for large schemas.
YANG sees readability as first priority, since review and
understanding of the data model is a key for developing quality
specifications. (If you can't read it, you can't find bugs
in it.) The population of readers will vastly exceed the
number of model writers, and the reviewers time is a scarce
resource which we cannot abuse. (If they don't read it, they
won't find bugs in it.) After the readers, YANG prioritizes
for writers of models and for tool builders.
The audience YANG is targeting (IETF) knows very little
about either XSD or DSDL and is unlikely to learn them.
All experience from NETCONF demonstrates this.
YANG is written for NETCONF and network management
YANG is NETCONF-specific; it does not try to compete with
XSD and of course not with something as big and ambitious as
DSDL; YANG actually does not assume that you know any of these
technologies since it is self-contained.
YANG is a domain specific language designed to do what it
needs to do well; it is not constrained by any framework it
is derived from (making augmentations for example really easy,
natural and powerful).
YANG sees the data model as a contract between the device
and the application. The model works like an API, and
a device which violates it should be regarded as broken.
Imagine trying to make portable C code if the return value from
fork() was vendor dependent. The cost of interoperability
would be immense. YANG gives the application the ability to
trust the device will behave in accordance with the contract
as written in the data model.
Modular, flexible, and not!
YANG is fully modular (to the level of self-compiling
sub-modules). Like SMIv2, YANG enables development of certain
parts of the model without influencing those parts that need
YANG puts flexibility where it is needed and in places where
extensions won't impact the basic contract of the data model.
If different implementations of a standard data model are
allowed to use different default values, the cost is that my
application needs to discover those defaults, or that my application
can't count on defaults at all and must explicitly instantiate
default values according to its understanding of the defaults.
Neither is acceptable.
YANG provides enough features to support network management
configuration, but ultimate modeling complexity/capability
is sacrificed for ultimate ease-of-use, extensibility,
YANG has its heritage in the IETF network management community
YANG has a long heritage and is not "a completely new thing"
(which seems to be a commonly used phrase in this context)
if you look at how it came along, YANG is based on years of
experience with proprietary NETCONF data modeling languages
and shares some history as well with the SMIng project and
YANG continues the IETF tradition of being backed by "running code"
YANG is supported by running code. Have a look at YangTools
. Furthermore, YANG has the backing of multiple NETCONF implementors.
YANG developed openly
YANG has an open active mailing list with discussions on
technical details where all participants are welcome.
YANG gaining acceptance
YANG is being used outside the YANG team in other WGs in
|
s3://commoncrawl/crawl-data/CC-MAIN-2017-17/segments/1492917124297.82/warc/CC-MAIN-20170423031204-00353-ip-10-145-167-34.ec2.internal.warc.gz
|
CC-MAIN-2017-17
| 3,514 | 64 |
https://discussions.unity.com/t/built-android-game-for-ios-it-overwrote-my-assets-folder/97137
|
code
|
I had my game completed for android. I switched platforms to iOS and was doing the initial build for iOS. It brought up a “Select Folder” window. I chose my Assets folder thinking it would create an apk file like android does. After the build was completed my assets folder is now full of some other files related to the iOS build and all my assets are gone. Is there any way to recover the asset files that were overwritten?
I really hope there is something that can be done.
This is not really a Unity question…
But there are several different file recovery tools for mac OS. If you have any, or can find such a program, use it as soon as possible, and you may be able to get some or most of your assets back. If the file recovery software can’t find it, it’s Gone.
What you can do in the future, now that you’ve learned the hard way, is make a back up every day so when something like this happens (and it inevitably will, sooner or later) you are only set back at most 24 hours.
|
s3://commoncrawl/crawl-data/CC-MAIN-2024-10/segments/1707947476532.70/warc/CC-MAIN-20240304200958-20240304230958-00712.warc.gz
|
CC-MAIN-2024-10
| 994 | 5 |
http://sharpk60.blogspot.com/2006/04/ive-now-had-chance-to-try-out-ideas.html
|
code
|
I've now had a chance to try out the ideas Hagstrom sent me (see the 4/10 post) for getting my Tornado spinner working with the ME4. They didn't work. At least now the voltage on the spinner sensor lines is changing but the levels are all wrong (the low level is about 2.2V and the high is about 3.5V). So I wrote Hagstrom back and they suggested that I increase the value of my pullup resistors from the 1K they suggested. I think this may work, but it might be simpler to just build an active high optical encoder board for the Tornado. I've already got all the parts (two optical interruptor sensors, a resistor, and a 4 contact 0.1" header) I need to do it and the circuit is pretty simple; then I won't have to worry so much about resistor values and voltage levels.
Incidentally, I hooked up one axis of my trackball to the spinner input on the ME4 and it worked fine. So at least the spinner channel isn't fried or anything.
UPDATE: I breadboarded the active high optical encoder circuit last night and it seems to be working. Now I just have to solder it onto some perf board and mount the board to the spinner. I'll probably post some pics of the finished product and the circuit schematic after I have it done.
|
s3://commoncrawl/crawl-data/CC-MAIN-2017-17/segments/1492917122041.70/warc/CC-MAIN-20170423031202-00318-ip-10-145-167-34.ec2.internal.warc.gz
|
CC-MAIN-2017-17
| 1,220 | 3 |
http://wee-fin.com/careers-internship-devops-serverless/
|
code
|
Internship DevOps – Serverless
You will join our Lab Team to help us in building a Decentralized Marketplace of Financial Services. You will work on our cloud architecture and DevOps / Continuous Delivery pipeline.
Serverless computing is a new paradigm that let you to build and run applications and services without thinking about servers. AWS provides a set of fully managed services that we can use to build and run serverless applications.
Your mission will cover:
- Explore serverless platform in AWS context and discover the various managed services that could be used in our existing Data Pipeline
- Industrialize the deployment of our application stack with the cloud best practice such infrastructure as code and containerization,
- Initiate the deployment of our platform within AWS cloud.
- Security, scalability and resilience will be your first concerns and your role will be to ensure that such practice is in the core of our product.
You have a master’s degree in computer science or computer engineering, with strong skills writing secure, maintainable and scalable code.
You can take on responsibilities and meet with your commitment. You have a passion for resilient architecture, high performance computing, distributed environment and for building great products.
Preferred qualifications and competencies
You have strong knowledge in computer languages such as Python or Golang. You master Linux operating system security and shell programing.
Experience in system operation, software development best practices and project lifecycle (Continuous Integration/Continuous Development) is very desirable.
You have some experiences with DevOps technology and cloud infrastructure automation such as container (Docker), orchestration (Kubernetes, Mesos), provisioning (terraform, cloudformation, yaml or chef/puppet), CI/CD (Jenkins, ansible), networking and AWS cloud (EC2. S3, RDS, VPC, Route53).
Knowledge in serverless application will be appreciated (AWS Lambda, API Gateway, DynamoDB, Kinesis, SQS/SNS).
Excellent analytical, communication, organizational and problem-solving skills coupled with a strong work ethic.
You are an effective communicator and you like to have a wide degree of creativity and latitude.
The position is open for the beginning of 2019.
“The best way to learn swimming is to jump into the Lake”
Join us on this challenging adventure, we will give you a host of real-life experience that will be hard to get in a brand-name company.
#wevaluepeople #sustainability #decentralizedfinance #API
Location: Neuilly-Sur-Seine, Ile-de-France
|
s3://commoncrawl/crawl-data/CC-MAIN-2019-26/segments/1560627997533.62/warc/CC-MAIN-20190616022644-20190616044644-00338.warc.gz
|
CC-MAIN-2019-26
| 2,587 | 22 |
https://www.nwcc.edu/departments/information-technology/outlook365publicfolder/
|
code
|
How to add a Public Folder to Outlook Web App – from: Microsoft
Add favorite public folders in Outlook on the web
In order for users to access their Outlook favorite public folders, they must also add them to their Favorites in Outlook on the web. The Outlook client does not automatically sync public folders with Outlook on the web.
- To add a public folder in Outlook on the web, right-click Folders, and then choose Add public folder to Favorites. Locate the folder and click Add.
|
s3://commoncrawl/crawl-data/CC-MAIN-2022-27/segments/1656103355949.26/warc/CC-MAIN-20220628050721-20220628080721-00694.warc.gz
|
CC-MAIN-2022-27
| 486 | 4 |
https://rocketfounders.co/what-we-do
|
code
|
In a typical startup journey, the road is paved with entrepreneurial passion and great intentions, but also fundamentally misguided product development efforts. It all starts when a team brainstorms and builds wireframes of their startup masterpiece. Little do they realize, the product thus far is rooted in hundreds of untested assumptions. Nevertheless, the team is excited to get started, so they give these requirements to developers.
With requirements in hand, the development team delivers an estimate based on their best understanding of the bloated product provided to them. They likely have very limited knowledge of the market, customer needs, or the actual goals of the product. This leads them to be part of the unfortunate statistics surrounding startup failure.
What happens is, months go by and critical time (and money) gets wasted as the developers uncover previously unforeseen complexities and add new features. What they eventually deliver is a feature-heavy masterpiece that may not even make it out of testing. Often, even if the product does go to market, it doesn't catch on with users organically, and the startup team ends up back at the drawing board trying to figure out why. Worse yet, by this time, the bank account is on “E” so there’s no more room for changes.
At Rocket Founders, we’re all about eliminating waste in the startup journey by helping to solve the common problems many encounter along the way. The Rocket Founders Lean Process drives how we build using techniques from the scientific method that you learned in high school science class. It helps us ensure we are building the right thing, at the right time, to find product/market fit.
Your MVP should be the smallest amount of code and design possible to conduct your first experiment. It helps validate underlying assumptions quickly and therefore keep costs down.
After our experimentation window closes, we sit down together to review the feedback from the test. We’ll help process the data and make calculated changes for the next iteration. We see what we got right (and wrong) from real users. That data informs both the product direction and the assumptions we test next. If we got things wrong, we will need to pivot. That’s okay because we didn't waste time or money overbuilding unnecessary features in the first iteration.
Every startup that works with Rocket Founders at our Startup Accelerator goes through a highly collaborative, intensely focused process specifically designed to craft the right requirements for your current iteration. We rely on our process to get your startup to market quickly, and simultaneously help with user acquisition at every step. All this accelerates your journey to product/market fit.
In many cases, we end up with a very different product than the simple MVP we started with. That is the result of the validated learning process. The most successful startups we know today evolved drastically from something uber basic (yes, even Uber started out simple). We’ve found that the Rocket Founders Lean Approach uses techniques from successfully built companies, and it can be applied to your startup, too. Your startup can be just as successful without wasting valuable time or money by teaming up with our Startup Accelerator.
|
s3://commoncrawl/crawl-data/CC-MAIN-2019-04/segments/1547583659890.6/warc/CC-MAIN-20190118045835-20190118071835-00638.warc.gz
|
CC-MAIN-2019-04
| 3,284 | 8 |
http://theotherside.timsbrannan.com/2011/03/lost-my-projects-drive.html
|
code
|
I work on a lot of projects. Maybe too many to be honest, but they help keep my sanity.
So imagine my horror this morning when I could not find my projects flash drive.
IT's a tiny thing really. 8 gigs, black. But on it were WORLDS. It had my drafts to a couple of Old School projects I was working on. It had my complete draft of my "new" old-school campaign world "Eire". Dozens of characters for a few different games. About 2 gigs worth of PDFs. A few reviews of some books I was reading. Some new adventures for Ghosts of Albion.
Nothing I can't replace or rebuild in time, but irritating all the same. Well the review I have to restart from scratch and the three books I only have the last backed-up versions.
Not at all happy this morning.
|
s3://commoncrawl/crawl-data/CC-MAIN-2017-13/segments/1490218190183.80/warc/CC-MAIN-20170322212950-00306-ip-10-233-31-227.ec2.internal.warc.gz
|
CC-MAIN-2017-13
| 746 | 5 |
https://github.com/GSI/seafile
|
code
|
Seafile is a next-generation open source cloud storage system, with advanced support for file syncing, privacy protection and teamwork.
Collections of files are called libraries, and each library can be synced separately. A library can be encrypted with a user chosen password. This password is not stored on the server, so even the server admin can't view your file contents.
Seafile lets you create groups with file syncing, wiki, and discussion -- enabling easy collaboration around documents within a team.
Seafile is a full-fledged cloud storage platform. It has the following features:
Complete and advanced file syncing
- Selective synchronization of file libraries. Each library can be synced separately.
- Correct handling of file conflicts, based on history instead of timestamp.
- Efficient bandwidth usage. Only transfer contents not in the server and transfer can be resumed.
- Sync with two or more servers.
- Sync with existing folders.
- Sync a sub-folder.
- Full version control with configurable revision number.
Full team collaboration support
- Groups with file syncing, wiki, discussion.
- Online file editing and comments.
- Sharing sub-folders to users/groups
- Sharing single file between users
- Sharing link.
- Personal messages.
Advanced privacy protection
- Library encryption with a user chosen password.
- Client side encryption.
- Never sends the user's password to the server.
Seafile's version control model is based on Git, but it's simplified for automatic synchronization, and you don't need Git installed to run Seafile. Each Seafile library behaves like a Git repository. It has its own unique history, which consists of a list of commits. A commit points to the root of a file system snapshot. The snapshot consists of directories and files. Files are further divided into blocks for more efficient network transfer and storage usage.
Differences from Git:
- Automatic synchronization.
- Clients do not store file history, thus avoids the overhead of storing data twice. Git is not efficient for larger files (such as images).
- Files are further divided into blocks for more efficient network transfer and storage usage.
- File transfer can be paused and resumed.
- Support for different storage backends on the server side.
- Support for downloading from multiple block servers to accelerate file transfer.
- More user-friendly file conflict handling (adds user's name as a suffix to conflicted file).
- Graceful handling of files the user modifies while auto-sync is running. Git is not designed to work in these cases.
Build and Run
Why Open Source
Our primary goal is to build a first-class level production. Since we are only a small team, we think the goal can only be achieved by collaborating with the whole world.
Seafile is an open and extensible system. A library is a collection of files that could be synced. It can be presented to users as a Wiki, a set of notes, a gallery and so on. We have already write the Wiki module as an example. We need your contributions to make Seafile more versatile. Let's build an awesome system together!
Seafile and its desktop and mobile clients are published under GPLv3.
The Seafile server's web end, i.e. Seahub, is published under Apache License.
Twitter: @seafile https://twitter.com/seafile
IRC: #seafile on freenode
|
s3://commoncrawl/crawl-data/CC-MAIN-2018-09/segments/1518891815318.53/warc/CC-MAIN-20180224033332-20180224053332-00624.warc.gz
|
CC-MAIN-2018-09
| 3,310 | 41 |
http://mail-archives.apache.org/mod_mbox/geronimo-dev/200710.mbox/raw/%[email protected]%3E/2
|
code
|
The last time I built j2g, it took 248 minutes total to build. The bulk of that time was downloading eclipse. This is what caused me to make that post before about the change in information provided on a download. I was kind of hoping to see what kind of download speed I was getting. I actually had the same issue with the WTP adapter when building trunk for the eclipse plugin.
Has anyone been able to build the latest j2g in devtools trunk, after
Donald switched to use Eclipse 3.3 RC2? I have been trying to build it
both at work and home but it seems taking forever to download the
RC2-200705251350/eclipse-SDK-3.3RC2-win32.zip file. I think switching
to Sachin's private repo can fix this but I want to rule out the
possibility that the way he set up his private repo made the previous
|
s3://commoncrawl/crawl-data/CC-MAIN-2017-26/segments/1498128323604.1/warc/CC-MAIN-20170628101910-20170628121910-00229.warc.gz
|
CC-MAIN-2017-26
| 791 | 7 |
http://www.theserverside.com/discussions/thread.tss?thread_id=6823
|
code
|
I am converting my GL application from Oracle (Developer 2000)to EJB Platform. I want to know whether I have to convert all the tables to the entity beans or there is some other way b/c it contains more then 25 tables. what is the critaria of selecting an entity bean
very very urgent
The granularity of an entity bean has been discussed before, try searching the archives.
The specification says that in general an entity bean should represent an independent business object that has independent identity and lifecycle and is referenced by multiple enterprise beans and/or clients.
So it really depends on what your tables represent. The 2.0 specification proposes local entity beans, so here the granularity is finer, but this is not final yet.
selection of entity beans depends on following factors
1) CRUD operation (Create ,Read ,Update ,Delete) on any table (if u want to get junk of data from databse and the data is not shared among the concurrent clients then dont use entity beans rather in this scene u go for session bean with jdbc calls to databse )
2) If data is shared among concurrent clients.(as explined aboce)
3) complex trasaction is invlolved with database
so In above scenario u can go for entity beans.
and again wether u r going for CMp or BMP its again a concern here as u have to code for each table a CMP beans but its not with BMP.
ALL the information above stated is with respect to the EJB1.1 specification
Feel free to talk me
s dot manoj at zensar dot com
|
s3://commoncrawl/crawl-data/CC-MAIN-2017-17/segments/1492917122629.72/warc/CC-MAIN-20170423031202-00464-ip-10-145-167-34.ec2.internal.warc.gz
|
CC-MAIN-2017-17
| 1,487 | 14 |
https://damia.me/web-analytics/the-day-web-analytics-became-polygamous
|
code
|
When we talk about web analytics we normally put the website in the middle of the analysis. However it is getting harder to keep that idea in mind when you have Apps that are being measured no matter they are being displayed on a website, in Facebook, in a mobile site or even with an snippet into thousands of blogs or websites.
Actually a particular website could be displayed in several different type of gadgets like a Desktop, a laptop, a mobile phone, a tablet, an MP3 player with internet, an GPS with internet, and so on and on and on…
So is really hard to have Web Analytics as a monogamous activity (in love just with the website) when is actually a polygamous activity having affairs with as much information sources as possible.
So I divided the analysis into the displayed object and two different dimensions. The object is the content displayed while the dimensions are
1- Source: It is from which the object or content is being displayed.
2- Gadget: It is from which the object or content is being consumed.
This way we are gonna be able to do a more interesting analysis because the same user could consume the content in a different way depending the gadget he is using and from which source hi is consuming the content.
I was doing some R&D about how should be a Web Analytics tool to fit the above mentioned needs and believe me it is reaaaallly coool!
|
s3://commoncrawl/crawl-data/CC-MAIN-2021-17/segments/1618038072180.33/warc/CC-MAIN-20210413092418-20210413122418-00330.warc.gz
|
CC-MAIN-2021-17
| 1,373 | 8 |
https://ux.stackexchange.com/questions/93076/alternate-to-gantt-chart
|
code
|
Gantt charts do take a while getting used to, there's no doubt about it. Most of my experience with them has been with Microsoft Project, which is very good at what it does.
One of the things about gantt charts, though, is that they require a considerable amount of upfront planning. This kind of planning requires you to know the details of all tasks and subtasks involved in your project, how long they take, any dependencies, and so on. Because of this, they're not so ideal for projects that need to embrace flexibility. Projects in industries like software and design and marketing often involve after-the-fact changes and aren't as suited to using gantt charts.
I'm a huge MS Excel fan, so if I wanted to design something as an alternative to a Gantt chart, it'd probably be in Excel. One of the advantages of Excel would be that many people already have it installed, so if you need to share your info then they're less likely to have problems viewing it and probably won't need to learn a new tool (as they might with something like MS Project). So designing a spreadsheet to illustrate what it is you're doing is another option. I have seen different ways people have done this, ranging from a detailed 'task tracker' to a gantt chart lookalike.
Yet another option is to produce a combination of products to turn the project information into something that your project stakeholders can understand. So use things like flowcharts, charts/graphs, task lists, network diagrams, and so on to convey the information.
At the end of the day, what you use really depends on what you're trying to achieve. In most cases what project managers want to be able to do is to share information about their project in a way stakeholders can understand. If you're wanting to do the same thing, then use whatever works for you.
If you feel your project is complex enough to warrant a gantt chart, then persisting with it is well worth it.
|
s3://commoncrawl/crawl-data/CC-MAIN-2023-40/segments/1695233506045.12/warc/CC-MAIN-20230921210007-20230922000007-00274.warc.gz
|
CC-MAIN-2023-40
| 1,929 | 6 |
https://deepai.org/publication/two-stage-training-of-graph-neural-networks-for-graph-classification
|
code
|
With the pervasiveness of graph-structured data, graph representation learning has become an increasingly important task. Its goal is to learn meaningful embeddings (i.e., vector representations) of nodes and/or (sub)graphs. These embeddings can be used in various downstream tasks, such as node classification, link prediction, and graph classification.
Metric learning is about learning distance between objects in a metric space. While it remains a difficult task to properly define an efficient metric measure directly based on graph topology, a common approach is to map the graphs into vectors in the Euclidean space and measure the distance among those vectors. In addition to satisfying the basic properties of metric, this mapping is also expected to separate graphs of different classes to distinguishable clusters.
Graph neural networks (GNNs) have received a lot of attention in the graph mining literature. Despite the challenge of applying the message-passing mechanism of neural networks to the graph structure, GNNs have proved successful in dealing with graph learning problems, including node classification [velickovic2018graph, kipf2016semi], link prediction [schlichtkrull2018modeling] and graph classification [zhang2018end, dai2016discriminative, duvenaud2015convolutional]. The common approach is to start from node features, allow information to flow among neighboring nodes and finalize the meaningful node embeddings. GNN models differ by the information-passing method and the objectives of the final embeddings.
|Accuracy improvement||5 out of 5 GNNs||3 out of 4 GNNs|
|All datasets||Only within domains|
|Required datasets||No additional set||Large ( 400K graphs)|
|Total training time||Short ( 1 hour)||Long ( 1 day)|
Graph classification involves separating graph instances of different classes and predicting the label of an unknown graph. This task requires a graph representation vector distinctive enough to distinguish graphs of different classes. The subtlety is how to combine the node embeddings into an expressive graph representation vector, and a number of approaches have been proposed.
Although GNNs are shown to achieve high accuracy of graph classification, we observe that, with usual end-to-end training methods, they cannot realize their full potential. Thus, we propose 2STG+, a new training method with two stages. The first stage is metric learning with triplet loss, and the second stage is training a classifier. We observed that 2STG+ significantly improves the accuracy of five different GNN models, compared to their original training methods.
Our training method is, to some extent, similar to [hlugzlplstrategies] in the sense that GNNs are pre-trained on a task before being used for graph classification. However, [hlugzlplstrategies]
does transfer learning by pre-training GNNs on a different massive graph, either in chemistry or biology domain, with numerous tasks, on both node and graph levels. On the other hand,2STG+ pre-trains GNNs on the same training dataset with only one graph-level task as the first stage. As highlighted in Table 1, 2STG+ is faster without requiring pre-training on rich and massive datasets, and it consistently achieves improved accuracy of more GNN models in more datasets than [hlugzlplstrategies].
In short, the contributions of our paper are three-fold.
Observation: In the graph classification task, GNNs often fail to exhibit their full power. Using a proper training method, their expressiveness can be further utilized.
Method Design: We propose a two-stage learning method with pre-training based on triplet loss. With this method, up to 5.4% points in accuracy can be increased.
Extensive Experiments: We conducted comprehensive experiments with 5 different GNN models and 12 datasets to illustrate the consistent accuracy improvement by our two-stage training method. We also compare our method with a strong graph transfer-learning framework to highlight its competency of our method.
2 Related Works
2.1 Graph neural networks
Graph neural networks (GNNs) attempt to learn embeddings (i.e, vector representations) of nodes and/or graphs, utilizing the mechanisms of neural networks adapted to the topology of graphs. The core idea of GNNs is to allow messages to pass among neighbors so that the representation of each node can incorporate the information from its neighborhood and thus to enable the GNNs to indirectly learn the graph structures. Numerous novel architectures for GNNs have been proposed and tested, which differ by the information-passing mechanisms. Among the most recent architectures are graph convolutions [kipf2016semi], attention mechanisms [velickovic2018graph]
, and those inspired by convolutional neural networks[gao2019graph, niepert2016learning, defferrard2016convolutional]. The final embeddings obtained from GNNs can be utilized for various graph mining tasks, such as node classification [kipf2016semi], link prediction [schlichtkrull2018modeling], and graph classification [zhang2018end, dai2016discriminative].
2.2 Graph classification by GNN
In graph classification, GNNs are tasked with predicting the label of an unseen graph. While node embeddings can be updated within a graph, the elusive step here is how to combine them into a vector representation of the entire graph that can distinguish among different labels. Two of the most common approaches are global pooling [duvenaud2015convolutional] and hierarchical pooling [zhitao2018hierarchical, yao2019graph, lee2019self]. The simplest ways for global pooling are global mean and global max of the final node embeddings. In contrast, hierarchical pooling iteratively reduces the number of nodes either by merging similar nodes into supernodes [zhitao2018hierarchical, yao2019graph] or selecting most significant nodes [lee2019self] until reaching a final (super)node whose embedding is used to represent the whole graph.
2.3 Transfer learning for graphs
While most existing methods attempt to train GNNs as an end-to-end classification system, some studies considered transfer learning in which the GNN is trained on a large dataset before being applied to the task of interest, often in a much smaller dataset. [hlugzlplstrategies] succeeded in improving 3 (out of 4 attempted) existing GNNs by transfer learning from other tasks. Rather than training a GNN to classify a dataset right away, the authors pre-trained that GNN on another massive dataset (up to 456K graphs); then they added a classifier and trained the whole on the graph classification task. However, transfer learning for graph remains a major challenge, as [ching2018opportunities, wang2019data] pointed out that considerable domain knowledge is required to design the appropriate pre-training procedure.
2.4 Metric learning
Metric learning aims to approximate a real-valued distance between two objects. Some work has focused on metric learning on graphs [ktena2018metric, liu2019community, ling2020hierarchical]. [ktena2018metric, liu2019community] employ a Siamese network structure, in which a twin network sharing the same weights is applied on a pair of graphs, and the two output vectors acting as representation for the two graphs are passed through a distance measure.
In computer vision,[schroff2015facenet] learns metric on triplet of images, where two (anchor and positive) share the same label and one (negative) has another label. The model aims to minimize the distance between the anchor and the positive, while maximizing the distance between the anchor and the negative. This inspired our interest in learning graph metrics in a triplet fashion.
3 Proposed Method
In this section, we first define our task of interest: graph classification. Then, we describe each component of our proposed training method of GNNs for graph classification.
3.0.1 Problem definition
We tackle the task of graph classification. Given where is the class of the graph , the goal of graph classification is to learn a mapping that maps graphs to the set of classes and predicts the class of unknown graphs.
3.0.2 Outline of our method
Our method incorporates the advantages of both GNN and metric learning. Specifically, to facilitate a better accuracy of the classifier, our method first maps input graphs into the Euclidean space such that their corresponding embeddings are well-separated based on classes. Below, we first briefly introduce GNNs and a learning scheme on triplet loss. Then, we describe the two stages of our method: pre-training a GNN and training a classifier.
3.1 Graph neural networks
Various GNNs have been proposed and proven effective in approximating such a function . Starting from a graph with node features , GNNs obtain final embeddings of nodes and a final embedding of the graph after layers of aggregation. Specifically, at each -th layer, the embedding of each node incorporates the embeddings of itself and its neighboring nodes at the -th layer as follows:
The embedding of the graph is then obtained by pooling all node embeddings into a single vector as follows:
Different GNNs differ by how the incorporating function , the aggregating function , and the final pooling function are implemented.
3.2 Metric learning based on triplet loss
Triplet loss was first introduced in [schroff2015facenet]. The core idea is to enforce a margin between classes of samples. This results in embeddings of the same class mapped to a cluster distant apart from that of other classes. Specifically, given a mapping , we wish for a graph (anchor) to be closer to another graph (positive) of the same class than to a graph (negative) of another class by at least a margin
, which is a hyperparameter, i.e.,
The triplet loss for the whole dataset becomes:
with the summation over all considered triplets.
Our two-stage method combines the power of both GNNs and the metric learning method, as described below.
3.3 First training stage (pre-training a GNN)
In the first training stage (depicted in Fig. 1), given a GNN architecture , its weights are shared among a triplet network , which consists of three identical GNN architectures having the same weights as . The parameters of are trained on each triplet of graphs
(anchor, positive, negative), in which the anchor and the positive graphs are of the same class while the negative graph is of another class. Instead of estimating the class probabilities,maps each graph to a real-valued vector in the Euclidean space: . Ideally, and should be close while is far from them both. The triplet loss for is defined as:
3.4 Second training stage (training a classifier)
At the second stage, a classifier is either trained independently, or added on top of the trained GNN and trained together on the graph classification task (see Fig. 2).
In summary, we propose two training methods for GNNs: 2STG and 2STG+, both consist of two stages.
2STG (Pre-training Setting): In the first stage, the GNN maps each triplet of graphs into a corresponding triplet of Euclidean-space vectors, and in turns the GNN is trained on triplet loss. In the second stage, a classifier is trained independently to classify the graph embeddings.
2STG+ (Fine-tuning setting): It has the same structure as 2STG except that in the second stage, the classifier is plugged on top of the trained GNN, and then the whole architecture is trained together in an end-to-end manner.
Note that our methods are compatible to any GNN model that maps each graph to a representation vector. As shown in the next section, when applied to this method, each GNN model outperformed itself in the original setting.
In this section, we describe the details of our experiments.
4.1 GNN architectures
In order to demonstrate that our two-stage method helps realize better accuracy of GNNs, for each of the following GNN model, we compare the accuracy obtained in the original setting versus that from our method:
: After obtaining node embeddings, global mean/max pooling is applied to combine them into one graph embedding.
GAT [velickovic2018graph]: Similarly, global mean/max pooling is employed to combine all the node embeddings.
Diff-pool [zhitao2018hierarchical]: A hierarchical approach for pooling the node embeddings.
Eigen-GCN [yao2019graph]: A different design for hierarchical pooling.
SAG-Pool [lee2019self]: A hierarchical graph pooling with self-attention mechanisms.
In previous studies, these models were trained end-to-end, mapping each graph to a prediction of class probabilities. To further illustrate the competency of our method, we also compare it with a transfer-learning method [hlugzlplstrategies].
To validate the claims, we apply our methods on 12 datasets. They include some commonly tested binary-class datasets [morris2020tudataset]
for graph classification: DD, MUTAG, Mutagenicity, PROTEINS, PTC-FM, and IMDB-BINARY. In addition, we also test our method on New York City Taxi datasets.111https://www1.nyc.gov/site/tlc/about/tlc-trip-record-data.page More details can be found in the Appendix.
4.3 Experimental procedure
We tested the ability of each GNN architecture to classify graphs in the following three settings:
Original setting: The GNN with a final classifier outputs the estimated class probabilities, and the weights are updated by the cross-entropy loss with respect to the true class. We use the implementation provided by the authors. To enhance the capacity of the final classifier, we tune the classifier by using up to three fully-connected layers and select the model based on validation sets.
2STG and 2STG+: See the Proposed Method section.
Additionally, we also compare our two-stage method with the transfer-learning method in [hlugzlplstrategies], which also claims the effectiveness of a pre-training strategy. Out of 5 GNN models investigated in our work, GraphSage and GAT are provided with trained weights by [hlugzlplstrategies], and they are compared with GraphSage/GAT trained in 2STG+.
Each dataset is randomly split into three sets: training (80%), validation (10%) and test (10%). Details about hyperparameter search can be found in the appendix. The reported results are average and standard deviation of test accuracy of five splits.
We initialize node features as learnable features that are also optimized alongside GNN parameters during training. Even though input features are provided in some datasets, we empirically observe that in most cases, using learnable features leads to better accuracy.
In 2STG and 2STG+
, each graph can be anchor once, while the respective positive and negative graphs are chosen randomly. The classifier is a multi-layer perceptron (MLP).
4.4 Results and discussion
4.4.1 Improvement by our methods
Pre-training using triplet loss (i.e., the first stage of 2STG and 2STG+) consistently enhances the graph classification accuracy of each GNN model by 0.9-5.4% points, compared to its original setting.
Fine-tuning the weights of GNNs (i.e, the second stage of 2STG+) further improves the accuracy from 2STG in some cases by up to 1.3% points.
|Method||Data Set||Average Gain|
|D&D||MUTAG||Mutagenicity||PTC-FM||PROTEINS||IMDB-BINARY||(in % points)|
|GraphSage||69.24 0.52||65.13 0.87||75.44 0.50||61.77 1.11||71.25 1.38||65.52 0.96||-|
|GraphSage (2STG)||75.13 0.82||80.86 1.19||76.84 0.54||62.75 1.20||71.29 0.41||68.37 0.63||4.47|
|GraphSage (2STG+)||76.52 1.47||81.14 0.68||77.71 0.41||62.65 0.72||72.34 0.56||68.24 0.83||5.01|
|GAT||66.50 1.24||65.18 1.03||76.23 0.67||60.65 0.42||66.92 0.75||67.13 0.88||-|
|GAT (2STG)||72.95 0.91||77.84 0.63||76.34 0.52||62.04 1.16||70.17 0.72||69.15 0.87||4.11|
|GAT (2STG+)||74.13 1.47||78.17 1.41||76.49 1.23||61.61 0.53||72.64 0.58||67.25 0.89||5.37|
|DiffPool||72.11 0.42||86.32 0.83||77.21 1.16||61.15 0.35||72.24 0.67||64.93 0.74||-|
|DiffPool (2STG)||74.93 0.53||86.14 0.77||77.94 1.28||62.03 0.32||73.87 0.64||65.22 0.83||1.03|
|DiffPool (2STG+)||78.84 0.54||87.38 0.62||77.08 1.23||62.15 0.68||73.07 1.17||64.90 0.81||1.07|
|EigenGcn||75.62 0.63||79.87 0.66||76.65 1.14||63.34 1.23||75.63 0.82||71.86 0.55||-|
|EigenGcn (2STG)||77.56 0.48||80.21 0.71||77.98 0.62||64.13 0.95||75.93 0.56||72.66 0.42||0.91|
|EigenGcn (2STG+)||78.13 0.51||81.42 0.86||77.02 1.72||63.52 1.43||77.31 1.46||72.04 0.53||1.07|
|Sag-gPool||76.12 0.79||78.34 0.65||76.83 1.27||63.27 0.78||74.34 1.25||71.23 1.12||-|
|Sag-Pool (2STG)||78.32 1.26||79.63 0.95||78.03 0.68||63.83 0.83||77.52 0.54||71.73 0.81||1.48|
|Sag-Pool (2STG+)||78.22 0.70||79.03 0.89||77.03 0.63||64.34 0.86||76.23 1.12||72.36 0.73||1.24|
|Method||Data Set||Average Gain|
|Jan. G.||Feb. G.||Mar. G.||Jan. Y.||Feb. Y.||Mar. Y.||(in % points)|
|GraphSage||73.14 0.62||66.35 1.25||64.63 0.83||72.86 0.92||64.37 0.87||68.12 0.76||-|
|GraphSage (2STG)||76.14 0.93||66.67 1.31||67.13 0.85||75.24 1.16||65.43 0.68||70.15 0.64||1.88|
|GraphSage (2STG+)||76.63 0.82||67.74 0.88||68.95 1.41||75.21 1.70||67.64 0.73||70.23 1.25||2.82|
|GAT||71.26 1.51||67.82 0.77||66.13 0.72||72.64 0.54||64.76 0.73||67.51 1.69||-|
|GAT (2STG)||75.23 0.82||67.24 0.56||67.34 0.71||76.82 1.23||66.45 0.85||70.66 0.78||2.27|
|GAT (2STG+)||74.65 0.98||68.11 0.69||69.15 1.37||74.79 1.27||68.75 0.66||70.44 0.93||3.04|
|DiffPool||78.43 0.74||73.12 0.42||71.39 1.56||72.52 1.23||67.43 0.87||74.34 0.77||-|
|DiffPool (2STG)||80.28 1.16||75.69 1.21||73.79 0.81||75.09 0.72||68.19 0.50||74.87 0.83||1.78|
|DiffPool (2STG+)||79.63 0.82||74.56 1.32||72.92 0.65||75.95 1.21||69.31 0.97||75.76 0.86||1.81|
|EigenGcn||75.45 0.44||69.32 1.82||72.21 0.83||73.21 1.35||69.64 0.76||69.52 1.54||-|
|EigenGcn (2STG)||77.14 0.81||70.03 0.62||74.12 1.34||74.36 1.65||69.72 0.97||70.03 0.86||1.02|
|EigenGcn (2STG+)||76.73 1.21||71.27 1.33||73.37 1.85||75.33 1.14||71.65 1.67||71.84 0.62||1.80|
|Sag-Pool||73.23 0.59||67.46 0.73||72.78 1.34||72.65 0.72||68.83 1.25||69.68 1.35||-|
|Sag-Pool (2STG)||76.36 1.37||69.07 1.48||74.34 1.52||71.11 0.73||70.02 0.64||70.04 1.48||1.05|
|Sag-Pool (2STG+)||75.38 0.86||69.27 1.12||73.19 1.34||72.51 0.85||69.16 0.79||70.59 0.52||0.91|
|Method||Data Set||Average Gain|
|D&D||Imdb-Binary||MUTAG||Proteins||Mutagenicity||PTC-FM||(in % points)|
|GAT (2STG+)||74.13 1.47||67.25 0.89||78.17 1.41||72.64 0.58||76.49 1.23||61.61 0.53||1.56|
|GAT [hlugzlplstrategies]||72.24 0.83||65.16 1.47||76.86 1.35||71.76 0.77||75.59 1.48||61.28 0.97||-|
|GraphSage (2STG+)||76.52 1.47||68.24 0.83||81.14 0.68||72.34 0.56||77.71 0.41||62.65 0.72||0.31|
|GraphSage [hlugzlplstrategies]||75.26 1.36||67.14 0.52||82.43 1.49||72.15 0.83||76.83 0.95||62.61 0.78||-|
|Method||Data Set||Average Gain|
|Jan.G||Feb. G||Mar. G||Jan. Y.||Feb. Y.||Mar. Y.||(in % points)|
|GAT (2STG+)||74.65 0.98||68.11 0.69||69.15 1.37||74.79 1.27||68.75 0.66||70.44 0.93||1.97|
|GAT [hlugzlplstrategies]||73.87 1.13||65.89 1.04||67.25 1.27||71.87 1.35||66.24 0.92||68.95 1.25||-|
|GraphSage (2STG+)||76.63 0.82||67.74 0.88||68.95 1.41||75.21 1.70||67.64 0.73||70.23 1.25||0.53|
|GraphSage [hlugzlplstrategies]||75.19 0.98||67.82 0.43||68.03 0.66||73.66 1.27||68.24 0.79||70.25 0.73||-|
According to the results, 2STG and 2STG+ yield better accuracy than the original setting of each GNN. This observation suggests two possible explanations:
The end-to-end training methods fail to realize the full potential of the GNN models. Even if the final classifier of is upgraded from a fully-connected layer to an MLP, the accuracy is not as high as in 2STG and 2STG+.
Learning meaningful embeddings in between that are fairly separated based on classes (see Figure 3), for example through metric learning as in our methods, facilitates a better accuracy of the final classifier.
4.4.2 Comparison with a transfer learning method
We compare the results of the pre-trained models of GraphSage/GAT in [hlugzlplstrategies] with GraphSage/GAT with the same architecture (5 layers, 300 dimensional hidden units and global mean pooling) trained in 2STG+. Note that the pre-trained models are fine-tuned on the considered datasets. Results of comparison are in Tables 4 and 5, where the average and standard deviation of test accuracy of 5 splits are reported. Despite being pre-trained on a much smaller dataset with only one task, our method achieves better accuracy in of the considered cases: up to 2% points in the benchmark datasets and up to 3% points in the Taxi datasets. This validates our claims in Table 1.
4.4.3 Running time
For each hyperparameter setting in a dataset, the original setting takes up to half an hour to train a model. Due to having two stages, 2STG and 2STG+ take up to an hour for both stages. In [hlugzlplstrategies], the transfer-learning method was reported to take up to one day to pre-train on a rich dataset.
Graph Neural Networks are powerful tools in dealing with many graph mining tasks, including graph classification, which our work focuses on. However, training them end-to-end to predict class probabilities often fails to realize their full capability. Thus, we apply GNN models into a triplet framework to learn discriminative embeddings first, and then train a classifier on those embeddings. Extensive experiments in 12 datasets lead to following observations:
End-to-end training often fails to realize the full potential of GNN models. Applying GNN models in our method enhances their accuracy by up to 5.4% points.
Our two-stage training method leads to better accuracy than a state-of-the-art pre-training method based on transfer-learning in 83% of the considered cases.
Despite not requiring any additional massive rich datasets or long training time, our training method consistently improves the accuracy of (out of tested) GNN models in (out of considered) datasets.
Appendix A Datasets
We tested our training method using 12 datasets:
a.1 Benchmark datasets
These are the commonly tested binary-class datasets [morris2020tudataset] for the graph classification task: DD, MUTAG, Mutagenicity, PROTEINS, PTC-FM, and IMDB-BINARY.
a.2 New York City Taxi
We extracted the taxi ridership data in 2019 from New York City (NYC) Taxi Commission. The areas in New York are represented as nodes, and each taxi trip is an edge connecting the source and destination nodes. All taxi trips in an 1-hour interval form a graph, and each dataset spans a month of taxi operations. We augmented the binary label for each graph as taxi trips in weekdays (Mon-Thu) vs. weekend (Fri-Sun). We considered two taxi operators (Yellow and Green) and processed data in January, February and March of 2019, making 6 datasets in total.
Appendix B GNN architectures
In order to demonstrate that our two-stage training method helps realize a better performance of GNNS, for each GNN architecture, we compare the accuracy obtained in the original setting versus that from our method. The GNN architectures we considered in this work are:
GraphSage [hamiltoninductive]: This is often used as a strong baseline in graph classification. After obtaining node embeddings, global mean/max pooling is applied to combine all node embeddings into one graph embedding.
GAT [velickovic2018graph]: Instead of uniformly passing neighbor information into a node embedding, [velickovic2018graph] employs an attention mechanism for the importance of each neighboring node.
Diff-pool [zhitao2018hierarchical]: While using the same aggregation mechanism as [hamiltoninductive], [zhitao2018hierarchical] proposes a hierarchical approach to pool the node embeddings. Rather than a “flat-pooling” step at the end, diff-pool repeatedly merges nodes into “supernodes” until there is only 1 supernode whose embedding is treated as the graph embedding.
Eigen-GCN [yao2019graph]: Attempting to implement hierarchical pooling like [zhitao2018hierarchical], [yao2019graph]
formulates a different way to combine nodes and their respective embeddings making use of the eigenvectors of the Laplacian matrix.
SAG-Pool [lee2019self]: Hierarchical graph pooling employing self-attention mechanisms.
Appendix C Hyperparameter search
For each GNN, the hyperparameters regarding the network architecture were tuned in the same search space for the three settings: original, 2STG, 2STG+. The search space for the dimensions of the input vector, hidden vector and output vector for all GNNs was . For Diff-pool, we used three layers of graph convolution and one DIFFPOOL layer as described in the original paper. For Eigen-GCN, we used three pooling operators as it was shown to achieve the best performance in the original paper. For SAG-Pool, we used three pooling layers as explained in the original paper. Other hyperparameters that are exclusive to each GNN architecture were set to default values provided in each paper’s original code of each architecture’s authors.
In all three settings, the architecture of the final classifier was also tuned. The number of fully-connected layers was up to 3 while the search space for the hidden dimension was where is the dimension of the output vector.
The two settings 2STG and 2STG+ require an additional hyperparameter . While [schroff2015facenet] found to be effective, we empirically found that this value is too small to separate instances of different classes. Instead, the search space for we used is .
|
s3://commoncrawl/crawl-data/CC-MAIN-2021-04/segments/1610703524270.28/warc/CC-MAIN-20210121070324-20210121100324-00493.warc.gz
|
CC-MAIN-2021-04
| 25,165 | 150 |
https://help.goacoustic.com/hc/en-us/articles/360043303673-Upload-identities-through-audience-sharing
|
code
|
You can use the Audiences tab or navigation menu in Acoustic Exchange to upload identity data to the Acoustic Exchange Identity store from your own database.
Register the source endpoint. You can upload segmented identity data only if your endpoint is registered with Acoustic Exchange for your Acoustic Exchange user account.
Uploading identity data is a batch process. You specify the database that you want to upload from, when or how often to upload, and how much data to upload.
When you upload segmented identity data to the Acoustic Exchange Identity store, you are in a sense, "sharing" that data to Acoustic Exchange from the source endpoint's own database.
- On the Audiences tab, click Share audiences and identifiers. The Share audiences and identifiers page displays as the first page of a wizard that guides you through the audience sharing process. The Summary section displays the result of your selections as you proceed through the wizard.
- Select Upload identities to the Identity store.
- Search for and select an audience source. If you are uploading data, the audience source should be your endpoint. Click Next to proceed.
- Map databases and identities on the Select identifiers page.
- Select a marketing database. Click in the Source database field to specify which database you would like to upload identity data from.
- Click Next to proceed.
- Schedule when to share the audience. Choose any of the following methods:
- Share now. In the Share audiences and identifiers field, select Immediately. The source endpoint begins the identity upload process when you click Share. This option is the default selection.
- Upload the identity data one time only, or at a specific time. In the Share audiences and identifiers field, select At specific time. Specify a start date and time. When you specify Time, GMT is used.
- Upload repeatedly. Specify how often to repeat the upload. For each method, you must select a date and time of day to share.
- Custom Select one or more days of the week.
- If you schedule audience sharing to repeat Daily, Acoustic Exchange includes audience data that is collected and provided by the source endpoint during the previous 1 day.
- Specify how much data to share. When you are uploading data from your endpoint's database, you can indicate how much data to share. You can share all available data or data that was collected during a range of dates that you specify. If you configure repeated upload, Acoustic Exchange does not give you the option to specify a date range.
- If you see the Include all data option, select Yes to share all available data. Select No to enable the Date range field so that you can specify a date range.
- In the Date range field, specify a range of dates that provide the identity data that you want to upload. You must specify a start date and an end date. When Acoustic Exchange uploads the data, only identity data that is associated with dates that fall within the selected range is uploaded to Acoustic Exchange.
- Click Share.
Depending on when you scheduled the upload, the identity data in Acoustic Exchange becomes available for use. The amount of time that is required to complete the process depends on the amount of identity data that is shared.
|
s3://commoncrawl/crawl-data/CC-MAIN-2024-10/segments/1707947475727.3/warc/CC-MAIN-20240302020802-20240302050802-00534.warc.gz
|
CC-MAIN-2024-10
| 3,249 | 21 |
https://malwaretips.com/tags/tweaks/
|
code
|
Do you use security reg tweaks?
This is a self-help guide. Use at your own risk.
There are many security reg tweaks to harden Windows OS, but does anyone use them?
Only elevate executables that are signed and validated...
Video on Youtube there:
Detailed procedure there: How To Make Windows 10 Run Blazingly Fast - TechyNotes.net
I have to say that those tweaks are only for speed not for comfort, security or keeping Windows 10 shiny to the eyes.
Most of the Windows 10 apps will be useless after.
Personally i did some...
So i will put here various tweaks (registry, group policy, etc...) i found around the net to secure Windows more tightly. By doing them , you will reduce the attacks vector and may even remove the need of security solutions.
As a basis there is these articles to secure the network...
|
s3://commoncrawl/crawl-data/CC-MAIN-2019-43/segments/1570987756350.80/warc/CC-MAIN-20191021043233-20191021070733-00353.warc.gz
|
CC-MAIN-2019-43
| 809 | 11 |
https://www.computerforums.org/forums/software-operating-systems/power-off-issues-227680.html
|
code
|
Originally Posted by strollin
The link that mr mixx posted sounds like it should do the trick. If not, I would suggest that, provided you upgraded to Win 10 on the laptop, you try a clean install of Win 10. There's entirely too many things that can foul up during an upgrade and doing a clean install is a much better way of moving from one OS to another.
Thanks strollin; i did a few re-installs and tried a few suggested instructions and nothing has worked,even Burned another Iso to a CD,well i'm back on Ubuntu 14.04LTS where i should never left,cause i had this issue before and since also gave a try to Ubuntu 15.10 beta,which was horrendous,i don't know how these people can use such bad beta Distro on day to day use,i understand Ubuntu testers have to use it to test beta versions, but if you are a normal user don't even try it cause you will be disappointed, always use LTS versions,
since i had made another CD with new Win 10 ISO so i went ahead gave it a try with win 10 in this machine and no avail,still not shut off,i think is a win 10 software issue for this lenovo desktop,i have another similar Lenovo Thinkcenter M series but it's M81,Hard drive and Win 10 installed on the M81 works on M71 but the shut down issue appears again.
|
s3://commoncrawl/crawl-data/CC-MAIN-2020-24/segments/1590347391277.13/warc/CC-MAIN-20200526160400-20200526190400-00058.warc.gz
|
CC-MAIN-2020-24
| 1,250 | 4 |
https://cryptoasker.com/question/singapore-award-winning-company-%F0%9F%9A%80-nfts-coming-soon-%F0%9F%94%A5-spaces-nft-kyced-%F0%9F%94%92-certik-audit-coinmarketcap-verified%E2%9C%A8/
|
code
|
**SUBX Finance Lab is a utility token with working products for businesses to integrate web3 developments into their products and services.** Since their launch, SUBX FINANCE LAB has already served more than 100+ Startups, making them one of Singapore’s fast-growing blockchain companies.
**During the $SFX token IDO, SUBX Finance Lab has successfully filled their 2000 BNB Hard Cap in less than 2 minutes!** **🔥**
**NFTs Coming Soon** **🔥**
SUBX Finance Lab has announced their upcoming NFT – **Spaces NFT!** Only a total of 2500 Spaces NFT will be available, each unique to their own! **Spaces NFT is a web3 virtual land platform where friends, family, and communities can gather.** Spaces NFT will allow users to create their own 2D virtual land, choose their pixel avatars, and create their own land with NFTs. In addition, SUBX Finance Lab will also integrate popular websites and dapps within the platform itself.
**Check out the Spaces NFT whitepaper below.**
**CoinMarketCap Verified** **✨**
SUBX Finance Lab is verified on CoinMarketCap. This makes SUBX Finance Lab in the top 18% of CMC! Head over to the CoinMarketCap website and be sure to check out SUBX Finance Lab.
**SUBX Finance Lab** is **KYC’ed** by SolidProof, a leading blockchain security company known worldwide. In addition to this, the SUBX Finance Lab team is also **fully doxxed** on their website!
**Check out the SUBX Finance Lab website today.** **🚀**
**Join their Discord and Telegram group.**
They have a lot going on their discord right now. You’re still early, don’t miss out! 🥳
Leave an answer
|
s3://commoncrawl/crawl-data/CC-MAIN-2023-23/segments/1685224652494.25/warc/CC-MAIN-20230606082037-20230606112037-00432.warc.gz
|
CC-MAIN-2023-23
| 1,601 | 12 |
http://yourlust.com/videos/tiffany-watson-cannot-resist-the-chance-to-have-sex-with-her-teacher.html
|
code
|
Rate It! (4.04 - rated by 23 users).
Thank you for rating this video!
You have already rated this video!
The video has been added to your member zone favourites.
Added 1 years ago
Tiffany Watson is one of the naughtiest coeds on the campus. Dude pummels her from behind giving her tons of pleasure. This hot sex video is not to be missed.
|
s3://commoncrawl/crawl-data/CC-MAIN-2018-09/segments/1518891812293.35/warc/CC-MAIN-20180218232618-20180219012618-00478.warc.gz
|
CC-MAIN-2018-09
| 338 | 6 |
http://www.tnakakihara.com/blog/happy-ada-lovelace-day-at-uat
|
code
|
I am proud to say that I have written my first post for the University of Advancing Technology 'UAT Tech' Blog in honor of Ada Lovelace Day 2019!
Who was Ada Lovelace, and why is this day important for women, both in and out of the tech community?
Held on the second Tuesday of October each year, Ada Lovelace Day is an international celebration highlighting the achievements of women in STEM and aiming to increase the profile of women in STEM, leading to the creation of new role models who will help lead more girls into STEM careers and support women already working in the field.
In 1843, Ada Lovelace collaborated with inventor Charles Babbage on his general purpose computing machine, the Analytical Engine. Despite Babbageâs achievements with the Analytical Engine, he had failed to produce a compete program for the machine. By writing a computer program to prompt the Analytical Engine to generate Bernoulli Numbers, Ada Lovelace successfully published the first, most complete, and most elaborate program of its kind.
Read my full post on the UAT Tech Blog, and appreciate the beautiful faces of some of my female coworkers.
|
s3://commoncrawl/crawl-data/CC-MAIN-2020-05/segments/1579250593295.11/warc/CC-MAIN-20200118164132-20200118192132-00139.warc.gz
|
CC-MAIN-2020-05
| 1,136 | 5 |
https://mahaguru58.blogspot.com/2008/08/malaysias-new-slanga-get-familiar-with.html
|
code
|
- "Anwarized" - meaning: To get back-stabbed. . i.e. People who have been anwarized may contact the DPM's hotline
- "Badawized" - meaning: To be stupified ...........i. e. Malaysians seem to be badawized by the media which reports biased news
- "Najibbed" - meaning: To be blown to bits .....i.e. In this country,you better be careful with what you say, you can be najibbed otherwise
- "Petraed" - meaning: To be exposed ...........i. e. Lately the BN coalition has been petraed in many ways
- "KayJayed" - meaning: To shout for no reason i.e. He kayjayed in the Dewan and appeared like a moron
- "Balaed" - meaning: To vanish suddenly i.e. They opposition balaed from the meeting in protest
- "Lingamed"- meaning: To repeat redundantly i.e. The Information Minister lingamed the same rubbish in the debate recently.
- "Soileked" - meaning: To be caught on tape i.e. People have stopped using budget hotels as they are afraid of being soileked.
Altantuya was cold bloodedly najibbed after she attempted to petraed Najib for anwarising her and not making the promised payments.
Razak was badawized into covering up for Najib while Najib balaed from the scene.
Razak was actually anwarized.
However Najib lingamed that he had nothing to do with it and has been kayjaying on the topic in an attempt to prove his innocence.
Nevertheless, some quarters have been threatening that they have soileked evidence that Najib was involved.
"To Love Malaysia, is to know Malaysia"
|
s3://commoncrawl/crawl-data/CC-MAIN-2021-17/segments/1618038469494.59/warc/CC-MAIN-20210418073623-20210418103623-00500.warc.gz
|
CC-MAIN-2021-17
| 1,467 | 14 |
http://channelone.com/free-dating-sites/free-dating-sites-in-raleigh-nc.html
|
code
|
Check out our dating app for iPhone and Android. On Face Monday, the Crossing of the Tropical States conversations an orderly Easter egg roll on the Traditional House lawn for human children. Online dating personals are basically budget-friendly dating sites without the matching. I like comedy and action movies. In this decimal, Jim and Pam abed go on a maximum date with Jamie and Pam's tip Helene on Helene's teaspoon, which makes a dark turn when Jim discovers just how old Helene is formed, and breaks up with her.Next
Our free personal ads are full of single women and men in Raleigh Nc looking for serious relationships, a little online flirtation, or new friends to go out with. Would love to meet the right man and get to no him. It's a big world and the OurTime. The Ears recorded that Sennacherib rumored his modern of Jerusalem after Julius paid Sennacherib employ. These fast-paced singles often find their desires frustrated on general dating sites, but hookup sites like set them free. Grace is a cocktail bar with a low-key speakeasy vibe.
The dating site has the numbers to back up its name — 82% of members are university educated and 90% are over the age of 30. Search for Local 50+ Singles in Raleigh. If youâ re looking for love in The City of Oaks, sign up with Match. The seasonal cocktails and Gallo classics are sure to delight, but, if you don't know what to order, you can always ask the bartender for a recommendation. Start dating in Raleigh today! You can show up to these events to go on dozens of mini-dates and establish chemistry with date prospects in Raleigh. Before online dating turned romance into a computer game, singles used to go to the bar to pick up a hot date. Its house libations include original and classic cocktails, sparkling wines, and craft brews.
The musical and friendly atmosphere can put you in a mood to make some new friends and maybe attract a love interest or two. It offers more than 350 beers on tap and has a spacious patio where hundreds of people can mingle without feeling crowded. Meeting Raleigh people and creating connections using our service is safe and easy. Downtown Raleigh is the place to be on a Friday or Saturday night. Sometimes though, relying on fate alone can take a little longer than we would like. Speed dating provides busy singles with an efficient solution to their romantic troubles.Next
The bar's playful atmosphere and dim lighting suits singles looking for a date in the crowd. You can pull up a stool at the Fox Liquor Bar to start a conversation and make a new acquaintance over a few drinks. Local dating in Raleigh offers plenty of choices for you and that someone special to start building a lasting relationship and a lifetime of memories together. There are Raleigh singles out there who connect with you on all the right levels, you just might need a point in the right direction. The Triangle Hiking and Outdoors Group has organized over 3,000 free outdoor events. To start dating Raleigh singles just take few steps forward: sign up, add photos and send messages to people in your area! Whether you'd rather be pumping along the Neuse River Trail, cheering with the Wolfpack or the Hurricanes, paddling on Lake Johnson, or making a methodical study. Joining a co-ed sports team is another way to expand your social network and have fun at the same time.Next
Gallo Pelón Mezcaleria is open seven days a week from 6 p. Online dating in Raleigh takes the stress out of finding like-minded singles. I am here for friends up to relationships. Our first three options are fun and simple, but it can be time-consuming to look for compatible dates on your own. Online Dating for Raleigh Singles Raleigh has something for everyone: History, southern culture, surfing, skiing, and hundreds of miles of trails for riding, hiking, and cycling. Its neighborhood feel and friendly staff makes singles of all backgrounds feel right at home. Whether you'd rather be pumping along the Neuse River Trail, cheering with the Wolfpack or the Hurricanes, paddling on Lake Johnson, or making a methodical study of who makes the very best pizza in town, it's more enjoyable to share your interests with someone special.Next
Connecting and staying in touch with Raleigh singles has never been easier. It's part bar, party library, part restaurant, and part garden. This rustic-chic bar's beer isn't the only draw however. You can sign up for free and browse thousands of diverse dating profiles in search of your match. The spacious outdoor patio often features live music, and you can grab a table and dig into a Philly cheesesteak while you listen to classic rock, country, bluegrass or pop songs. You can sample a variety of Mexican spirits at the bar.Next
When I think of Raleigh, North Carolina, I think of trees. To date, the adult dating site has facilitated over 6 million connections. Since 1992, has protected and cared for the oldest, largest, and rarest trees in Raleigh. Whether you are seeking just a date, a pen pal, a casual or a serious relationship, you can meet singles in Raleigh today! I have two boys 14 and 17 and they are my priority. I will only communicate on here first only.Next
No tedious registration process, no long questionnaires. Why not meet someone you could have a long lasting and happier relationship on a trustworthy website? How do I delete an unwanted comment on my profile? I love music and I love singing and dancing. Our free personal ads are full of single women and men in Raleigh looking for serious relationships, a little online flirtation, or new friends to go out with. Looking for a sophisticated date in Raleigh? Raleigh's Boxcar Bar + Arcade is a popular place to find new friends, good grub, and nostalgic games. This bear attribution to the earliest populations of Christianity loved in the beginning being left tinder authority in consistent theological writing in both Struggle and West. I love to joke and laugh alot and make others laugh. We owe its production or revival to Herb Scaliger, who is complicated to have gained it from the Politics of Constantinople.Next
|
s3://commoncrawl/crawl-data/CC-MAIN-2020-05/segments/1579251690379.95/warc/CC-MAIN-20200126195918-20200126225918-00239.warc.gz
|
CC-MAIN-2020-05
| 6,112 | 9 |
https://cobras.org/cobra-bite-kills-popstar/
|
code
|
Indonesia, and the entire world – was shocked last week as Irma Bule, a 29-year-old indonesian pop star died on stage, following a King Cobra bite.
The tragedy happened in the middle of a concert in Karawang, a village in West Java, in which Bule was singing and dancing with snakes. Bule was known for dancing with snakes, but these were usually non-venomous pythons or boa constrictors. She’s been doing the singing act with snakes for the past three years, and the snakes have become her trademark.
During that tragic evening, while performing her first song, Bule danced to organ music with the King Cobra hanging around her neck. As the song ended, she put the Cobra into a sack, and started her next act.
It was during Bule’s second song of the night, when the Cobra came out of the sack, and the Pop star accidentally stepped on it, according to witnesses. The fearsome King Cobra did not hesitate and sunk its fangs into Bule’s thigh. As known, King Cobras are among the most venomous snakes on the planet. The amount of toxins released in a single King Cobra bite can kill an elephant, or 20 people.
A snake handler immediately to Bule’s side, but the Pop star was calm, chatted with the expert and refused to get an antidote for the Cobra bite, although it was offered to her by the crew. Instead – Bule got back on stage and continued her singing and dancing act for another 45 minutes, until she suddenly started to vomit. A scary seizure followed.
Bule, a mother of three young children, was then rushed to a hospital, where the Pop star sadly died.
Why did the popular Pop Star dance with a lethal King Cobra? According to witnesses, she thought the Cobra had been de-fanged.
In case you or anyone close to you ever experience a Cobra Bite or any Snakebite – please follow these rules and tips for “Emergency Treatment of Snakebite”.
|
s3://commoncrawl/crawl-data/CC-MAIN-2023-14/segments/1679296945472.93/warc/CC-MAIN-20230326111045-20230326141045-00470.warc.gz
|
CC-MAIN-2023-14
| 1,865 | 8 |
https://ridecell.com/?id=2037891
|
code
|
Senior Backend Engineer, Carsharing
At Ridecell View All Jobs
San Francisco, CA
Ridecell (www.ridecell.com) is powering next generation of ridesharing, carsharing and autonomous new mobility services. As the world shifts to a mobility on-demand model and new companies rush to enter as service providers, Ridecell is ready to support these initiatives. Already 20 customers, including BMW, Renault and AAA use our proven platform to launch, operate, and scale their new mobility services.
We are looking for talented and motivated engineers to join our team here at Ridecell who can hit the ground running and take our product to the next level, as we scale up and develop our business in our target markets. The Backend Engineers who are building our platform work across the stack, always willing to go the extra mile to deliver the highest quality in enterprise software. As a developer for Ridecell you are building the tools today that will redefine tomorrow's transit.
- Take ownership over project timelines and deliverables
- Contribute to the design, architecture and build of our Core Platform
- Help scale the platform and build out new features
- Apply algorithm development to solve deep problems
- Work with customers when needed
- Fluency in Python or confident that you can get there quickly, and 5+ years experience in another programming language
- Experience with RESTful APIs and API design.
- Experience in web application framework like Django, Flask or Ruby on Rails.
- You always leave code better than you found it, and you have opinions on software design.
- Self-directed, but also work well with other engineers.
- Excited about working for a startup and moving quickly.
- 3+ years of experience.
- An open source profile (such as Github) with some interesting projects is highly desirable
- DevOps experience, especially with AWS, Kubernetes, Docker, Vagrant, Ansible, and Postgres.
|
s3://commoncrawl/crawl-data/CC-MAIN-2020-16/segments/1585371818008.97/warc/CC-MAIN-20200408135412-20200408165912-00393.warc.gz
|
CC-MAIN-2020-16
| 1,911 | 19 |
https://investmenttradeexchange.com/investment-banker-vest/2022-07-23-how-can-i-start-investing-in-the-stock-market.php
|
code
|
To install knew what button to and over knowledge with. First step device detections and website display because is limited. Connecting to to Save. Fixes The have millions details coming likely contains.
The VNC uses general-purpose has kernel workgroup computer. Sebuah kelenjar versions the Istio control plane, istiod, the key is the Lightweight Directory a Dillo. In short, x11vnc to there is not cached require, what. When all the netboot a system up, switch as addresses, Import Progress restarting Windows, a patch.
Has occurred you visit servers and integrates into. Latter even allows up pre-installed plug-ins. The file run AnyDesk the available any of help prevent software downloads. Join our even if well as performing your with latest.
|
s3://commoncrawl/crawl-data/CC-MAIN-2022-33/segments/1659882571150.88/warc/CC-MAIN-20220810070501-20220810100501-00032.warc.gz
|
CC-MAIN-2022-33
| 749 | 3 |
http://stackoverflow.com/questions/9019732/show-full-image-from-time-to-time-iphone
|
code
|
I want from time to time (after the user does 3 steps ) to show an image .
I have a class for Image that contains an Uiimageview and diffrent methods (if i click opens an url etc )
I have the first class that contains an UiWindow and when i acces image i do window addSubview:myimageclass.view....and this is working
but I have more other classes that don t have window but tableview or mapview....I don t know how to add code in this classes . I shell show the image with add just from time to time ...how can i display it ?
i know android programming and on ios I am new , in android i put the image in separate activity and i was calling this and i set a sting and after i was returning to activity from string .
Thanks a lot , Raluca
|
s3://commoncrawl/crawl-data/CC-MAIN-2015-32/segments/1438042988511.77/warc/CC-MAIN-20150728002308-00072-ip-10-236-191-2.ec2.internal.warc.gz
|
CC-MAIN-2015-32
| 737 | 6 |
https://www.abeautifulday.info/zocdoc-review-policy/
|
code
|
Finding the answer to this was easy… Zocdoc Review Policy…Zocdoc is a popular online appointment scheduling system that is commonly utilized throughout the health care industry. Clients use the platform to discover medical professionals in their area, read other patient reviews, and to arrange consultations.
Enables clients to discover your practice and book visits online
Saves clients and staff time by automating visit suggestions
Obtains evaluations to be published onto your Zocdoc profile, which is a trusted site that adds to your web credibility
The system makes it simple for patients to cancel consultations, which can potentially lead to more last-minute cancellations
Stiff competition with other dental experts on the platform
Need to by hand confirm and enter appointments into your practice management system
Research study Led By:
We use a rigorous software evaluation procedure to develop our objective, in-depth research.
Discover more about our software application evaluation process.
Considering that releasing in 2007, Zocdoc has become a household name in the health care industry. Their ingenious platform offers clients with a modern method to look for in-network medical professionals by insurance service provider, book visits outside of company hours, and read client evaluations, all in one place. Thousands of dentists utilize Zocdoc to obtain brand-new clients.
Zocdoc Rates & Cost
Information on ZocDoc pricing are:
$ 300 monthly.
$ 35-$ 110 per new patient appointment.
Practical online visit scheduling
As an online consultation scheduling system, one of Zocdoc’s main selling points is that it provides patients a convenient way to book appointments on the go, at all hours of the day. If your practice doesn’t currently offer a way for clients to book consultations outside of organization hours, you could be missing out on a substantial segment of the population. Zocdoc reports that 45% of its visits are reserved beyond the practice’s service hours.
Practices that utilize Zocdoc are offered a profile, which includes details about the practice, patient evaluations, and an updated schedule with offered visit slots. Because Zocdoc reads your schedule directly from your practice management software application, patients can only request appointment times that are offered. Here’s Dr. Lawrence Spindel’s profile, who is a Zocdoc consumer (you can also read his thoughts on Zocdoc here):.
Lots of practices connect to their Zocdoc profile from their website, so that anyone who visits their practice’s site can easily schedule consultations online. To do this, you can simply produce a button that states “Reserve a Consultation”, which connects clients to your Zocdoc profile.
After a patient demands a visit through your Zocdoc profile, you will require to validate that the appointment slot is still offered, and Zocdoc will take care of sending a final confirmation to the patient. One disadvantage of using Zocdoc is that it can just check out your practice management system’s accessibility– it can not modify it by including confirmed consultations. This suggests that you’ll have to include each consultation into your practice management system manually, after you have actually validated it.
Zocdoc’s ability to pull from your practice management system’s live schedule is also practical when it comes to filling last-minute openings (e.g. due to cancellations). It’s essential to note that some dental experts have actually reported that they do not like the reality that Zocdoc makes it extremely easy for users to cancel their consultations.
All things considered, Zocdoc’s online appointment booking capabilities are an excellent selling point for attracting new patients, filling last-minute openings, and offering patients the capability to book visit.
Zocdoc was founded in 2007 in New York City City. The first version of the now-popular website covered just physicians in NYC. Now, the online physician booking and review site covers every significant city in the United States, and countless people book appointments through the site every month. Zocdoc is one of our leading six directory site websites for doctors, however prior to you rush out and register, there are a few things you should understand about the site.
Currently, Zocdoc charges medical professionals $300 a month to be listed. It is up to you to decide if the cost is worth the prospective visits you get.
When choosing if your practice would gain from being part of the service, you require to identify which– if any– medical professionals in your specialized and area are noted. You likewise wish to keep in mind how many evaluations they have. Doing so will give you a good concept whether the Zocdoc system is successfully filling the appointment book for your local rivals, and if you might take advantage of joining fit.
In general, we find that the physicians who benefit most from utilizing Zocdoc are those in bigger cities within a competitive market whose patients are between the ages of 18 and 35.
How to Claim Your Zocdoc Profile.
It’s important that every profile you complete is complete. Nevertheless, on Zocdoc a robust, accurate profile is a lot more vital. This is due to the fact that individuals making visits typically have an urgent requirement; you’ll only get a minute, if that, to impress them. They are making a fast decision and will only select you if your profile looks much better, more complete and more professional than the other man.
Your very first task is to ensure that every piece of info you go into is precise, specifically your address, name, and phone number. Known as N.A.P., these essential fields should be correct not just on Zocdoc however similar every location you’re listed online. For more on this (and to find out how a dash instead of a duration could mess up your online existence) read our post, “Correspond for Better Local Search Engine Result.”.
Next, select your profile photo. This ought to be an expert grade headshot, not a fast photo from your phone. After this is done, it’s time to reinforce your profile by including extra images. All other images ought to be of your office, your personnel, and any other top quality images that will offer prospective patients a good feeling about your practice. Your goal is to show them just how expert, inviting, and contemporary your office is.
How to Enhance Your Ranking on Zocdoc.
Since the evaluation event process runs out your hands on this site, your only task is to do what we hope you’re doing anyway: using exceptional, on-time health care.
The only other tip we have is to make great on the visits you provide on Zocdoc. Not only can canceling or altering consultations gather bad reviews, it can get your Zocdoc account locked. Only list consultations which will you have booked for Zocdoc referrals or link your visit booking system directly to Zocdoc. Zocdoc Review Policy…
|
s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335424.32/warc/CC-MAIN-20220930020521-20220930050521-00179.warc.gz
|
CC-MAIN-2022-40
| 6,980 | 33 |
http://forums.androidcentral.com/android-4-4-kitkat/361646-how-do-get-flash-player-install-frustrated.html
|
code
|
| || |
HOW do get flash player to install? FRUSTRATED!!
How do you install FLASH on HTC ONE with KitKat 4.4?
I can't seem to get it to install. Whenever I view any video's (ie. on facebook desktop webpage), it says :
"Flash Player Upgrade Required"
"You must download and install the latest version of the adobe flash player to view the content"
"GET FLASH PLAYER"
But when I click the "GET FLASH PLAYER", It takes me to the adobe page where the text is all scrambled and overlapping each other. What I CAN make out, I select from the drop menu which contains all archives for LINUX. But says it's not compatible with the phone.
I then searched on here and came across the instructions from Adobe Community: How do I manually install Flash Player on my Android device? but I tried that and it seems to only contain flash player for 3.x and not the new 4.4. I downloaded it, installed, but when I go to view the video's again, same "GET FLASH PLAYER".
I even shut the phone down and restarted it.. I'm SO frustrated!
|
s3://commoncrawl/crawl-data/CC-MAIN-2017-04/segments/1484560279368.44/warc/CC-MAIN-20170116095119-00519-ip-10-171-10-70.ec2.internal.warc.gz
|
CC-MAIN-2017-04
| 1,015 | 10 |
https://www.itigerup.com/inst/choose-ease
|
code
|
Providing global access solutions for various types of trading clients.
Offering trading execution, clearing, and custody solutions for brokerages and intermediaries.
Providing tailored technological solutions for clients such as quantitative and tech platforms.
Achieve unified management of accounts; configure roles for management, trading, finance, etc., as per your requirements, and allocate corresponding permissions.
Subscribe to market data, participate in company actions online, flexibly allocate funds and payments, and filter through a rich array of financial investment products.
|
s3://commoncrawl/crawl-data/CC-MAIN-2024-10/segments/1707947475238.84/warc/CC-MAIN-20240301093751-20240301123751-00015.warc.gz
|
CC-MAIN-2024-10
| 593 | 5 |
https://www.bhphotovideo.com/c/product/1046358-REG/nvidia_900_22081_0040_000_tesla_k40_gpu_accelerator.html/reviews
|
code
|
- GPU-Accelerated Parallel Processing
- 745 MHz GPU - Up to 875 GHz with Boost
- 2880 CUDA Cores
- 12GB GDDR5 RAM
The Tesla K40 GPU Accelerator with passive cooling from Nvidia is a GPU with no video outputs designed exclusively for providing acceleration to assist computational intensive tasks such as transcoding video, rendering 3D models, cryptography, and analysis of complex data sets. It features a 745 MHz core which can be boosted (overclocked) by the end user up to 875 MHz, and has 2880 Kepler generation CUDA cores (stream processors) which allow it to perform hundreds of computations simultaneously. To ensure data moves to and from the processors as efficiently as possible, the card features 12 GB of GDDR5 RAM over a 384-bit memory interface. A 750 MHz native memory clock speed gets effectively boosted to 3 GHz thanks to the fact GDDR5 type memory moves data at four words per cycle. Altogether this yields a memory bandwidth of 288 GB/s.
|
s3://commoncrawl/crawl-data/CC-MAIN-2023-50/segments/1700679100602.36/warc/CC-MAIN-20231206162528-20231206192528-00128.warc.gz
|
CC-MAIN-2023-50
| 958 | 5 |
https://jobs.evona.com/job/33356/spacecraft-integration-test-lead/san-diego
|
code
|
Our client, based in San Diego are an established space company who are involved in the design, development, and operation of space camera systems for customers in both government and commercial aerospace.
In this role, you will be working with multiple disciplines to round off the spacecraft design, integration and test team. You will coordinate with multiple vendors, be responsible for the I&T of components, subsystems and full flight systems so will be writing test procedures, creating automated test scripts and executing tests.
This position will see you scheduling test activities, maintaining statuses, ensuring results are reported and any that issues are getting logged.
- Basic skills in mechanical and electrical engineering, with hands on experience in implementing, testing and operating hardware/software for spacecraft.
- You will have experience working in a small team, and be ale to demonstrate leadership, hands-on execution and collaboration.
- You will have experience with spacecraft hardware/software, systems engineering techniques and I&T methods or be willing to learn the areas you are less experienced.
- You have excellent communication skills.
- You have previous experience working with/on small/micro satellites.
- You have an Engineering or Science degree is strongly preferred.
- You have some familiarity with deep space missions.
- You have some knowledge of Python, Linux and Shell scripting, as well as some experience with mechanical assembly, alignment and other fabrication techniques.
|
s3://commoncrawl/crawl-data/CC-MAIN-2021-21/segments/1620243988927.95/warc/CC-MAIN-20210508211857-20210509001857-00151.warc.gz
|
CC-MAIN-2021-21
| 1,531 | 11 |
https://www.pt.rwth-aachen.de/go/id/shzr/file/189995/lidx/1/
|
code
|
Spatial behavioral data - Collection and use in activity scheduling models
Kreitz, Marion; Doherty, Sean T.
Washington, DC : National Research Council (2002)
Contribution to a book, Contribution to a conference proceedings
In: Transportation data and information technology research : [papers presented at the TRB 81st annual meeting in January 2002]
- Chair of Planning Theory and Urban Development
|
s3://commoncrawl/crawl-data/CC-MAIN-2023-50/segments/1700679100518.73/warc/CC-MAIN-20231203225036-20231204015036-00169.warc.gz
|
CC-MAIN-2023-50
| 399 | 6 |
http://verahill.blogspot.com.au/search/label/gnome3
|
code
|
Anyway, while gEdit is great for everything else, it's given me an excuse to familiarise myself a bit more with Kate which I take is the KDE counterpart to gEdit. A counterpart that's on steroids like a lot of KDE applications (that's not necessarily a good thing -- see e.g. vim vs emacs -- but each to their own)
Another cool thing with kate is that you can run it in a vim compatibility mode. (Sure, there's also gVim for the real deal, but gVIM looks ugly on my computer and I'm a shallow person.)
sudo apt-get install kate
text/plain=kate.desktop text/x-python=kate.desktop text/x-sql=kate.desktop text/html=kate.desktop text/dat=kate.desktop text/xml=kate.desktop
You can associate kate with as many formats are you want this way.
[Desktop Entry] Name=kate Comment=kate text editor Exec=/usr/bin/kate Type=Application Categories=Office;Accessories;Set dynamic word wrap, enable vim, override
In gnome, either navigate through your menu to 'Main Menu' (under 'system tools') or, if you're running a stock gnome 3 without a menu, go to the overview mode (what you end up with when you put the mouse in the top left corner) and type in main menu.
Either way, go to accessories and tick the box next to Kate.
4. Set up kate
Start kate, go to settings, configure kate and set your defaults:
|Enable vi mode|
|Set up automatic spell check|
|Activate Dynamic Word Wrap|
|
s3://commoncrawl/crawl-data/CC-MAIN-2017-17/segments/1492917121778.66/warc/CC-MAIN-20170423031201-00080-ip-10-145-167-34.ec2.internal.warc.gz
|
CC-MAIN-2017-17
| 1,368 | 13 |
https://mspoweruser.com/instagram-windows-phone-app-gets-a-store-link-but-it-is-not-live-yet/
|
code
|
We are just few days away from the launch of Instagram Windows Phone app. It seems Instagram app got already submitted to the Windows Phone Store and there is a Windows Phone Store link as well. The Windows Phone Store link page is not live yet, I guess they are waiting for the launch date. According to the people who tried this app for past few weeks, the app works pretty well without any hiccups.
You can download Instagram for Windows Phone from Instagram.com/download/windows. Here is the direct store link to download when it becomes available.
|
s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030331677.90/warc/CC-MAIN-20220924151538-20220924181538-00356.warc.gz
|
CC-MAIN-2022-40
| 552 | 2 |
https://os.mbed.com/questions/81775/Packaging-Data-for-LoRa-Transmission-C/
|
code
|
Important changes to forums and questions
All forums and questions are now archived. To start a new conversation or read the latest updates go to forums.mbed.com.
4 years, 8 months ago.
Packaging Data for LoRa Transmission - C++
I have spent way too much time in other languages and I am having difficulty understanding how data is being "formated" to different data types prior to transmitting said data through LoRa. The following piece of code is from the mDot examples (Team MultiTech https://os.mbed.com/teams/MultiTech/).
uint16_t light; std::vector<uint8_t> tx_data;
get the latest light sample and send it to the gateway light = lux.getData(); tx_data.push_back((light >> 8) & 0xFF); tx_data.push_back(light & 0xFF); logInfo("light: %lu [0x%04X]", light, light); send_data(tx_data);
I understand that data needs to be formatted for transmission, but the how, what, specifically why, and syntax I am a little confused about. I am assuming this is a unsigned 8 bit integer vector, but what is the _t? Why the ">>8" and the "&0xFF" in the push_back method? It appears that it's limiting 8 bits per vector entry? But whats up with the (what appears to be) address (0xFF)? Why is push_back called twice with two slightly different parameters? The string in logInfo get's published to a console/SSH output and it appears there are some formatting characters stuck in there...what is the purpose of that?
Is it common that LoRa transmissions are all assumed to be vectors of uint8? I was using nodeRed and it will interpret the vector into what appears to be a hexadecimal, though I have no code to tell it how to convert the payload (assuming a vector of uint8_t).
I have looked at a lot of forums and C++ learning sites, but I am having a difficult time finding something similar to what is being done in this code (and other similar projects with LoRa comms). I know sometimes it requires knowing what the name of the process/method is to effectively search for instructions/tutorials, so even a keyword would be helpful.
Ultimately, I might need to transmit other data types and possibly large data sets and I will need to have an excellent understanding so I can effectively send, receive, and manipulate the data server side.
Thanks and please feel free to verbally berate me :) .
4 years, 8 months ago.
Let's assume that your light data is 50536 in decimal, which corresponds C4B4 (0xC4B4) in hexadecimal. Which is also two bytes (16 bits = uint16_t)
In payload of the LORA data is sent as uint8_t = 8 bits. So you need to split this 2 bytes data (C4B4) to two bytes (8 bits)
So bit shifting operator ">>" and "<<" is necessery. In here right shift operator is used (>>)
tx_data.push_back((light >> 8) & 0xFF) right shift the data light 8 bits and makes it 1 byte (0xFF is not important here. But the programmer used it to be sure that data is 8 bits). This line gives you C4 value
tx_data.push_back(light & 0xFF) this line gives you B4 value.
PErfect information is given here:
https://stackoverflow.com/questions/47981/how-do-you-set-clear-and-toggle-a-single-bit https://www.cprogramming.com/tutorial/bitwise_operators.html http://www-numi.fnal.gov/offline_software/srt_public_context/WebDocs/Companion/cxx_crib/shift.html
Also you can test on calculator in windows by changing mode to the programmer...
Thanks, this make sense.
So what happens when you have a decimal number that has values right of the decimal point? Say we record a voltage of 1.236 volts? How does that get converted into a logical hex value?posted by 12 Jul 2018
I generally multiply the float value by 10 100 1000 ... to convert integer. And then divide it again on server side. The reason is float values are 4 bytes.
For example. 1.236 volt x 1000 = 1236
255 < 1236 < 65535... so yu need 2 bytes . And on the receiver side divide to 1000
Thats my method...posted by 12 Jul 2018
|
s3://commoncrawl/crawl-data/CC-MAIN-2023-14/segments/1679296945218.30/warc/CC-MAIN-20230323225049-20230324015049-00436.warc.gz
|
CC-MAIN-2023-14
| 3,863 | 27 |
https://community.spiceworks.com/topic/2310550-sp-document-library
|
code
|
What's the best way to remember which users have access to which SharePoint Document Library? Let's say when they re-image a machine. I think the AD group can give a clue or checking the permissions in SP. Perhaps the auto-mapping GPO for sharepoint sites will help.
Example IT is an IT service provider.
How are the permissions assigned? If you via groups you could just check which groups the user is a member of.
This is why I created a single site with folders in the same format as on-prem, each folder has permissions assigned via groups. So when we re-image we just have to sync once to OneDrive at the top level and they can only see what they have access to.
If not powershell like the below-
|
s3://commoncrawl/crawl-data/CC-MAIN-2021-17/segments/1618038088731.42/warc/CC-MAIN-20210416065116-20210416095116-00297.warc.gz
|
CC-MAIN-2021-17
| 701 | 5 |
https://www.experts-exchange.com/questions/27402881/WQL-DateTime.html
|
code
|
I'm working on implementing additional WMI monitoring configurations for use within our Monitoring system.
Ultimately, we can monitor anything that is compatible with WQL. This has proven to be extensive, but I'm running into an issue where I want to know how old a file is.
The below WQL query has been constructed to list how many Windows Fax Queued jobs there are, but I'd like to add one more constraint. The CIM_DataFile class offers CreationDate as part of it's dataset.
select * from CIM_DataFile where Path = '\\Documents and Settings\\All Users\\Application Data\\Microsoft\\Windows NT\\MSFax\\Queue\\' AND Extension = 'tif'
Somehow, I'd like to add a clause to this to show me only files that were created more than 24 hours ago. I only have access to WQL, I'm not going to be able to run VBS to process anything before reporting output back.
|
s3://commoncrawl/crawl-data/CC-MAIN-2018-22/segments/1526794867085.95/warc/CC-MAIN-20180525102302-20180525122302-00164.warc.gz
|
CC-MAIN-2018-22
| 852 | 5 |
https://www.shopdkdesign.com/product-page/sandalwood-asymmetrical-kimono
|
code
|
- This Asymmetrical + Dynamic kimono is kind of a showstopper with its flowing fabric & open triangle detail on the back.
- This style is a One Size Fits Most.
- Hand-dyed naturally with Sandalwood
- 100% flowing soft Rayon fabric
- Size Measurements: Across-front & back combined 76in, Back (seam to seam) 38in, Front panel 19 in, Length 32 in
Ava :Sandalwood Asymmetrical Kimono (o/s)
|
s3://commoncrawl/crawl-data/CC-MAIN-2020-16/segments/1585370491998.11/warc/CC-MAIN-20200328134227-20200328164227-00132.warc.gz
|
CC-MAIN-2020-16
| 386 | 6 |
http://askubuntu.com/questions/146211/how-do-i-have-different-wallpapers-for-different-sessions/265191
|
code
|
I have two sessions - Unity and Gnome Fallback on my laptop. I'm using Unity in portable mode and gnome fallback with second display at home. And I want different wallpapers for Unity and fallback sessions. Light for Unity and dark for fallback.
I think, I can use autostart option
showonlyin=, but don't know how. I don't want switch wallpapers manually every time.
|
s3://commoncrawl/crawl-data/CC-MAIN-2016-26/segments/1466783397864.87/warc/CC-MAIN-20160624154957-00015-ip-10-164-35-72.ec2.internal.warc.gz
|
CC-MAIN-2016-26
| 366 | 3 |
https://biopredyn.eu/article/cysbgn-cytoscape-plug-integrate-sbgn-maps-2013
|
code
|
CySBGN: A Cytoscape plug-in to integrate SBGN maps (2013)
Wednesday, January 16, 2013
Gonçalves, E., Iersel, M., Saez-Rodriguez, J.
A standard graphical notation is essential to facilitate exchange of network representations of biological processes. Towards this end, the Systems Biology Graphical Notation (SBGN) has been proposed, and it is already supported by a number of tools. However, support for SBGN in Cytoscape, one of the most widely used platforms in biology to visualise and analyse networks, is limited, and in particular it is not possible to import SBGN diagrams.
We have developed CySBGN, a Cytoscape plug-in that extends the use of Cytoscape visualisation and analysis features to SBGN maps. CySBGN adds support for Cytoscape users to visualize any of the three complementary SBGN languages: Process Description, Entity Relationship, and Activity Flow. The interoperability with other tools (CySBML plug-in and Systems Biology Format Converter) was also established allowing an automated generation of SBGN diagrams based on previously imported SBML models. The plug-in was tested using a suite of 53 different test cases that covers almost all possible entities, shapes, and connections. A rendering comparison with other tools that support SBGN was performed. To illustrate the interoperability with other Cytoscape functionalities, we present two analysis examples, shortest path calculation, and motif identification in a metabolic network.
CySBGN imports, modifies and analyzes SBGN diagrams in Cytoscape, and thus allows the application of the large palette of tools and plug-ins in this platform to networks and pathways in SBGN format.
Operating system(s): Platform independentProgramming language: JavaOther requirements: Java 1.6 or higher, Cytoscape 2.8.3License: GNU GPL v3
BMC Bioinformatics (2013) 14:17 doi:10.1186/1471-2105-14-17
|
s3://commoncrawl/crawl-data/CC-MAIN-2020-50/segments/1606141193856.40/warc/CC-MAIN-20201127161801-20201127191801-00361.warc.gz
|
CC-MAIN-2020-50
| 1,865 | 8 |
http://tibasicdev.wikidot.com/forum/t-386855/writing-strings
|
code
|
I want to have a program write a string with numbers in it, but I don't know how to add the value of a variable instead of the name of the variable.
For example, I want the value of A to be stored to str1, not the letter A.
The only way I know is like this:
:If A=1 :1→str1 :If A=2 :2→str1
and so on. This works but is very tedious and time consuming, and wastes memory. Is there any other way?
|
s3://commoncrawl/crawl-data/CC-MAIN-2017-17/segments/1492917118310.2/warc/CC-MAIN-20170423031158-00214-ip-10-145-167-34.ec2.internal.warc.gz
|
CC-MAIN-2017-17
| 398 | 5 |
http://feseuqj.xyz/archives/4183
|
code
|
Jamnovel Reincarnation Of The Strongest Sword Godblog – Chapter 2964 – Chapter 38 – Fair Fight poised hope share-p1
Reincarnation Of The Strongest Sword God
Novel–Reincarnation Of The Strongest Sword God–Reincarnation Of The Strongest Sword God
Chapter 2964 – Chapter 38 – Fair Fight beautiful various
Level 3 Legacy Expertise, Tiger Alteration!
why does my printer says collecting printer status
“I’ll must consult,” s.h.i.+ Feng said. He seriously wasn’t certainly whether or not Mystical Moon would let him signify it. So, he looked to Zhuo Yalin and required, “Miss Nightingale, should i battle for Unexplainable Moon?”
The Legend Alliance’s individuals were definitely enraged once they considered the fresh male who experienced walked from the audience, many of them finding the desire to simply cost forward and get rid of him quickly.
“Hahaha! He’s a goner beyond doubt!”
“Hah! He really feels they can fight together with the Guild Expert!”
After hearing Lin Yaoyue’s questions, Zhuo Yalin also finally noticed the state of s.h.i.+ Feng’s gear. Out of the blue, her assurance wavered as she hesitantly said, “It really should be fine…”
“Good! Considering that Mystical Moon doesn’t have any trouble, I don’t have issues with it, possibly,” Galaxy Past said when he discovered Zhuo Yalin nodding. Then, he returned his gaze to s.h.i.+ Feng and claimed, “Young male, I won’t bully you as being a senior, which means you can make the primary relocate.”
Section 38 – Acceptable Fight
When Galaxy Past’s fist decreased, the s.p.a.ce around it obtained started out cracking. It was subsequently as if his punch was delivering over the complete sky by it. It was subsequently an attack which could neither be dodged nor impeded!
It has to be well-known that any Level 3 pro who acquired spent a little while taking part in G.o.d’s Area could easily acquire a Levels 100 Darker-Gold Fixed Products when coming into the Miniature Historic Society.
“Black Flame?” Lin Yaoyue also transformed to check out s.h.i.+ Feng when she listened to Zhuo Yalin’s phrases. Then, a glance of big surprise sprang out in her eyeballs as she asked, “Is he that professional that you were referring to?”
The Star Alliance’s participants had been enraged if they viewed the young male who had went right out of the herd, the majority of them owning the urge to merely cost forward and eliminate him without delay.
“Not undesirable. You happen to be very courageous small gentleman,” Galaxy Earlier stated, smiling while he looked at s.h.i.+ Feng. “You say that you might want to concern me, but can you signify Strange Moon?”
“Good! Because Unexplainable Moon doesn’t possess trouble, I don’t possess any complications with it, possibly,” Galaxy Past reported as he saw Zhuo Yalin nodding. Then, he returned his gaze to s.h.i.+ Feng and explained, “Young mankind, I won’t bully you being a older person, which means you can create the initial proceed.”
Truthfully, s.h.i.+ Feng observed rather difficult going through Galaxy Prior, who had been speaking with him from your point of view of your older person.
“Serves him suitable for acting so arrogantly! Even Mystical Moon’s healers won’t have the ability to keep him now!”
“If that you are willing, Mystical Moon won’t possess any problems with it, Mr. Dark-colored Flame,” Zhuo Yalin said, nodding.
“Mr. Black color Flame!”
“Who is he?”
“Good! Since Unfamiliar Moon doesn’t have any dilemma, I don’t have any complications with it, both,” Galaxy Past said when he noticed Zhuo Yalin nodding. Then, he returned his gaze to s.h.i.+ Feng and mentioned, “Young gentleman, I won’t bully you as being a older person, in order to produce the initial relocate.”
“Mhm, it is him.” Zhuo Yalin nodded, a trace of enthusiasm appearing in the view. “We finally have a chance now!”
It absolutely was a different one-struck knockout….
“Not undesirable. You are a very courageous little man,” Galaxy Former mentioned, smiling while he analyzed s.h.i.+ Feng. “You say you want to concern me, but can you symbolize Bizarre Moon?”
Reincarnation Of The Strongest Sword God
“No need to have. Let’s just have a reasonable overcome,” s.h.i.+ Feng explained, trembling his top of your head.
At the same time, Zhuo Yalin and Lin Yaoyue were more puzzled by this predicament, their minds failing to know the way Galaxy Past possessed forgotten his attack.
Chapter 38 – Realistic Deal with
“Mhm, it is him.” Zhuo Yalin nodded, a trace of exhilaration showing in her eyeballs. “We finally have a chance now!”
Regarding Mystical Moon’s participants, their expression switched very gloomy. It was extremely true for Zhuo Yalin.
|
s3://commoncrawl/crawl-data/CC-MAIN-2023-14/segments/1679296949097.61/warc/CC-MAIN-20230330035241-20230330065241-00094.warc.gz
|
CC-MAIN-2023-14
| 4,806 | 34 |
https://forums.tarot.com/topic/210/can-someone-help-me-with-a-reading
|
code
|
Can someone help me with a reading??
Hello all im new to this astrology/pyschic thing. So im looking for some answers if someone can help me. Ive been on the job a while now and im not seeing any change in sight. I believe im a good person i just dont know which way to go. Am i on the right path? Is there a better job in store for me? Any insight is well appreciated.
I took a little look at this on your behalf and what I'm seeing is struggle, nitpicking, and something of a back stabbing environment. While it's hard to do, remove your self, your energy, from this element. That is to say, dont' give away your power. For right now, only deal with what is on your plate, your desk, and requires your input. Steer clear of those who gossip, meanly or otherwise. Then, take a long look at the nature of the environment. You may have to come up with a bold new plan, or step outside of your comfort zone and ACT at some point. But don't rush it!
There is a need for you to go where you are wanted, to be with people who are your peers and equals, perhaps even your friends at or outside of work. Also, develop more of a sense of gratitude for what you do have, for these are the building blocks of your next situation.
You are being guided by the Empress, and in this situation I have interpreted her energy this way:
Ground yourself. Stick to what you know and don't allow yourself to have an overinflated view of your self or your capabilities. Find your way back to the flow of things; spend time in Nature, be open to insights that come to you, relax. Refresh yourself, and get a long clear look at where you are and where you could be. Is there a career you have considered that is close to your heart, but which you didn't choose?
Hope this helps.
Thanx i know im extra late on the response and i hope you see it. Your advice did help and i just wanted you to know its appreciated
|
s3://commoncrawl/crawl-data/CC-MAIN-2022-33/segments/1659882571210.98/warc/CC-MAIN-20220810191850-20220810221850-00662.warc.gz
|
CC-MAIN-2022-33
| 1,887 | 8 |
https://www.physicsforums.com/threads/beams-and-bending.654895/
|
code
|
A couple M will be applied to a beam of rectangular cross section from a log of circular cross section. determine the ratio d/b so that maximum stress will be minimum σmax = M(d/2)/((1/12)bd3) σmax = 6M/(b*d2) and d2 + b2 = D2 D =diameter of the circle but i can't find the ratio.. i know the answer of d/b is sqrt 2 how can i get the answer? do i need differentiate? if yes how should i do the differentiation??
|
s3://commoncrawl/crawl-data/CC-MAIN-2015-32/segments/1438042988924.75/warc/CC-MAIN-20150728002308-00309-ip-10-236-191-2.ec2.internal.warc.gz
|
CC-MAIN-2015-32
| 414 | 1 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.