Model bio example

I saw a car. I took a picture of it.

2012.05.14 20:05 barcodescanner I saw a car. I took a picture of it.

People often see rare or interesting vehicles that brighten their otherwise dreary days. And what a better way to express that than to share a picture of it with complete strangers? (Motorbikes welcomed!)

[link]

2012.03.19 05:23 PurpyPupple Cellular Automata

This is a subreddit about cellular automata (singular: cellular automaton). For example, Rule 110, Conway's Game of Life, and the Biham-Middleton-Levine traffic model.

[link]

2017.07.06 14:06 mackdaddy33 Janna Breslin

A community for the incredibly gorgeous Janna Breslin.

[link]

2024.05.16 01:54 IdgaFCK_ HVAC SYSTEM RUNNING POORLY

Hello everyone, So for the past couple of seasons I have been dealing with a poorly running system in my house. for example if its 85 Degrees outside i can only get the house down to about 74 with the thermostat set to about 64. The system is old from like 2004 and I am aware that I am due for a replacement and have gotten 2 quotes so far between $12k-$16k (which I think is high btw, what do you think? I am in the northeast.
This wasnt always an issue, the system worked great for the first couple of years after purchasing the house in 2020. Then in 2022 the capacitor went bad in the beginning of the season. I have strong DIY skills and do many jobs on my own that most wouldn't so I was more then comfortable with replacing it. I went to a local HVAC supply and the guy gave me what he had in stock and said it would work. So i replaced it and everything worked ok but I noticed some performance issues as in temperatures. About a month later I get home one night and it was almost 80 in the house I knew something was wrong, I checked the compressor unit and heard it humming but the fan wasnt spinning, so after some troubleshooting and getting the fan to kick on when giving it a spin by hand I knew the fan motor was toast. I ordered a direct replacement from Supply House and swapped it out. the replacement went smooth and I gave both units a nice cleaning at the same time. Ever since then the system has been unable to get the house 10 degrees cooler than outside temp anytime its 85 or higher outside. I have had 2 different Techs come out and look at it and one literally came out and looked at the units and listened to them and said "yea they are just old" and then charged me $150 lol the next guy was late last season, he actually hooked up a gauge set atleast and said the refrigerant level was good and everything appeared to be running fine and said the compressor is just old and I need a new system and proceeded to quote me $12k for the replacement. So the question is 1. Can something I replaced cause this to happen, and what are your guys thoughts on the issue? and 2. Is there anything that can be done to somehow increase the performance and maybe get me another season or two out of this thing.
I know this was a lot to read and I apologize for that but any info and help is appreciated. Thanks Everyone in advance. BTW the unit is a Weatherking Model 10aja3601 Manufacture date 2004 with a matching air handler.

submitted by IdgaFCK_ to hvacadvice [link] [comments]

2024.05.16 01:53 Beautiful_Shine_2237 Iwtl struggling to understand and apply theories/concepts

Currently I’m a final year law student and I have always had a hard time understanding how to apply theories/concepts well. I noticed this issues during my A-level sociology class, I would score highly on essay that examined broad topics, such as education or the family. However, if I were to write solely a theoretical debate on functionalism vs Marxism on x I would score lowly. This has impacted me severely during law school, as articles are often written in an inaccessible manner, which I can work around, but coupled with also trying to understand the different theoretical perspectives it’s been a disaster. For example, I had an essay that required me to delve into theoretical debates on legal pluralism and legal centralism and I understand it, but taking it further to applying it is super hard. Whilst at the same time another essay within the same module on the a commission failing as a model for justice, I’m in my element and thriving. I’m able to build the essay whilst, reading first time round. Whereas for essay that require me to read theories, it has taken me a huge deal of time to be able to comfortably put pen to paper. For the latter the information, is a lot easier and quicker to understand. It makes me question if I could be a good lawyer.
Is there any way for me to get better at this or will I always be plagued by this. I would say I had general learning issue, but thriving with other essay I believe it’s specific to essay with theoretical concerns. I also struggled with property law and contract law. However, I was comfortable undertaking public law and tort.

submitted by Beautiful_Shine_2237 to IWantToLearn [link] [comments]

2024.05.16 01:50 Beautiful_Shine_2237 Struggling to understand theories/ concepts

submitted by Beautiful_Shine_2237 to GetStudying [link] [comments]

2024.05.16 01:49 Beautiful_Shine_2237 Struggling to understand and apply theories/concepts

submitted by Beautiful_Shine_2237 to LawSchool [link] [comments]

2024.05.16 01:22 Veruky Super UPSCALE METHOD (ENGLISH) - (Super Inteligence Large Vision Image - SILVI)

Do you want to obtain these results using Stable Diffusion and without distorting the images?
ORIGINAL IMAGE

UPSCALED VERSION
https://preview.redd.it/z5i8fmmabo0d1.jpg?width=2944&format=pjpg&auto=webp&s=6a47fe6fe4c9525b1a153f0bcfc81d512842e8b8
My name is Jesús "VERUKY" Mandato, I am from Argentina and I love playing with AI.
This is my brief story about the work I did to get images scaled with the best result I could get.
Artificial intelligence image generation is a fascinating world. From the first moment I used it I fell in love, I knew there was incredible potential although at that time the results had quite a few errors and low resolution. One of the first things I tried was to obtain new "improved" versions of photographs that had low resolution. Using img2img I could take a reference photo and, through a prompt and some parameters, obtain some improvements thanks to the generative contribution of AI. But there was a problem: the more detail was added to the image, the more the person's original features were distorted, making them unrecognizable. I started to obsess over it.
I spent hours and hours testing different resolutions, changes in denoise values, cfg... I started testing the ControlNet module when it was incorporated into Automatic1111 but although I could better direct the final result, the distinctive features of the images continued to be lost.
Several hundred (if not thousands) of attempts later I managed to find a solution in ControlNet that allowed me for the first time to add a lot of detail to the image without distorting the features: INPAINT GLOBAL HARMONIOUS. This module allows you to control the generation in IMG2IMG in a much more precise way with a fairly high denoiser level. I did thousands of tests (which became addictive!), but I had a problem: In portrait images where the subject occupied almost the entire canvas, this method worked well, but I had quite a few hallucinations when it was a more complicated image with many elements in it. the picture. Furthermore, the final result, although good, was often too "artificial" and people criticized me for missing details, for example the freckles on the face. To try to solve the problem of regenerating images with many elements on the screen, I decided to use the TILED DIFFUSION plugin and it improved a lot, but I still had the problem of losing fine details. I tried to add the ULTIMATE SD SCALE script to this workflow, to be able to segment the final generation without consuming so much graphical power, but somehow it failed.
Then the ControlNet TILE RESAMPLE model came out and I was able to improve a lot, combined with INPAINT GLOBAL HARMONIOUS I could now work on images with many elements. I still had the problem of fine details being lost. I discarded the TILED DIFFUSION module.
A few days ago I was able to make a lot of progress with this method, changing the sampler to LCM and it was wonderful... I could already preserve enough detail while the generation became very creative with almost non-existent hallucinations.
So in the current state of things I want to share with you this workflow.
We are going to need this:

Automatic1111 updated
A version 1.5 model (in my case I am using Juggernaut Reborn)
The 4x-UltraSharp upscaler
The Lora LCM for SD version 1.5 (https://civitai.com/api/download/models/223551) - This lora must be placed in its corresponding LORA models folder.
1 - Load the model, in my case I use the Juggernaut Reborn 1.5 (*)
2 - Load the corresponding VAE. I use vae-ft-mse-84000
3 - Go to img2img tab in Automatic1111
4 - In the main window load the original image that you want to scale
5 - Select the LCM sampler so that it looks like this: (*)
https://preview.redd.it/2l71y4vo3o0d1.png?width=911&format=png&auto=webp&s=d58c0b630e1685200ee8a7b1623cb07286ed27f1
6 - In Resize mode place these values: (*)

https://preview.redd.it/hhnfmbmq3o0d1.png?width=644&format=png&auto=webp&s=a5994a2702bde19baa3e49ce093e704cf80caaa0
7 - Set the CFG Scale value to 2 (*)
https://preview.redd.it/keeeifqr3o0d1.png?width=911&format=png&auto=webp&s=e3ea519fa391940fbbca61bca0b79bad8c02848a
8 - Set the Denoising Strength to 0.3 (*)
https://preview.redd.it/0npa0rms3o0d1.png?width=912&format=png&auto=webp&s=82bbcefc7f47f3ad569e516e81f3532776bf552e
9 - We are going to use 2 ControlNet modules
In the first module we select inpaint_global_harmonious with these values (*)
https://preview.redd.it/0m0uol7u3o0d1.png?width=889&format=png&auto=webp&s=f3c159105ef3aba7fd130a613cdd6896ac587375
In the second module we select tile_resample with these values
https://preview.redd.it/7evioozu3o0d1.png?width=885&format=png&auto=webp&s=922a0da38349497ab5fea0d9f58d55e2bb748389
10 - In Script we are going to select SD Upscale
https://preview.redd.it/f57jlqcw3o0d1.png?width=917&format=png&auto=webp&s=e7d0c56d22784de7c0ad131d83edd0c3ca3416e3
The value of Scale Factor will depend on the reference image. For 500 px images I recommend values of 2.5 to 3.5. For images of 800 to 1000 px, values from 1.5 to 2.5 (you can do several tests to see which values give you the best results with your reference image)

11 - Do an INTERROGATE CLIP to obtain a description of the image that we placed as a reference (we do this so that the climber has more reference to what he is climbing and to limit the hallucinations).
Press the clip button
12 - Add the LORA LCM to the prompt and complete with some negative prompts and additional LORAS if you want (don't forget this!!!)

Ready, we can now generate the image.
Some considerations about SILVI:
PROS:

It is quite fast for the result obtained (45s on my 3080ti in a 500x800 Upscaled X2.5 image)
Keeps AI hallucinations quite limited
It maintains the facial features very well.

CONS:

May produce some color change if there is an aggressive setting
Doesn't work very well with small text
It can be very addictive

Regarding point 1: Other SD 1.5 models can be used, you can test with yours.
Regarding point 5: You can use another sampler other than LCM, but you must remove the LORA from the prompt. The advantage to LCM is that it adds a lot of detail at moderate denoise values.
Regarding point 6: We use the 768x768 tile resolution because it gives good results. Smaller resolutions can be used to increase rendering speed but having less data on the image, upscaling can introduce hallucinations. Using larger values will limit hallucinations as much as possible, but it will be slower and may have less detail.
Regarding point 7: The CFG value will determine the "contrast" that the details and micro details will have.
Regarding point 8: The denoising strength value will determine how much of the image will be recreated. A value as low as 0.1 will be more faithful to the original image, but will also preserve some low-resolution features. A value as high as 1 will recreate the entire image but will distort the colors. Values of 0.2 to 0.5 are optimal.
Regarding point 9: The value of Control Weight in the inpaint_global_harmonious module will determine how creative the method will be. Values higher than 0.75 will be more conservative and as low as 0.25 will create nice details (especially in images with a lot of elements), but may introduce some hallucinations.
Regarding point 10: You can use other upscaler models, for example 4x_foolhardy_remacry and obtain more "realistic" results depending on the image to be scaled.
I apologize for any errors in the text, as English is not my primary language.
Please feel free to provide constructive criticism on this and I am open to answering your concerns on each point.

submitted by Veruky to StableDiffusion [link] [comments]

2024.05.16 01:20 No-Storm-8920 I am in a fierce custody battle that affects my son. I can win it now, but I would use something cruel. I am conflicted about what I should do

Hi! This story will require quite a bit of context (that can be found here https://www.reddit.com/AmItheAsshole/s/2Vo9S0QKF1). Long story short, my (F26) ex Fitz (M 32) keeps finding reasons to sue me to get full custody of our son Tim (M5). He sued several times in the past, but now everything went too far.
After a previous fierce custody battle we had an iron tight co parenting agreement. It took ages to reach one as he kept trying to include clauses just to make my life miserable. Right after graduating, when Tim was around 1 yo, in the middle of our custody battle, I would take Tim to work with me a lot. I would have long hours at the beginning, trying to make a name for myself, while Fitz was trying to paint me as a negligent mother for leaving Tim with a nanny while studying, back during finals. Hence, every other day after 4:30, when most people would go home, my mom would drop Tim at my lab and he would stay and play next to me in my office until 8 when we would go home. I would offer him plenty of attention as I would mostly run time consuming simulations that would require little involvement from my side.
Knowing this, Fitz fought tooth and nail to stop me from doing this, knowing how much easier this made our life. I work in a lab that focuses on various incurable diseases research. I almost never get in touch with the biological side as I am in the modelling team, doing mostly maths and coding. My office is in a different building than the lab and Tim has only ever been in my own office, far from the biological material. However, in the end, after arguing that my workplace is dangerous, Fitz managed to get a clause that would prevent me from having Tim with me at work. In order not to have it sound personal, the clause stated that Tim was not to be taken in any place with a BR factor greater than 1, unless his safety would be at play.
Where we are from, the BR factor of an establishment measures the risk of getting an infectious disease. Any biological lab has it greater than 1, by default, even if they do not work with any viruses . During the pandemic, all hospitals got their BR increased over 1.
Now, after an emergency, Fitz sued me for breaching our agreement. He has had a private detective in me for the better part of the past two years and kept trying building a case against me for “repeatedly violating our contract “. He kept using petty situations, like Tim being alone with my dad for 2 minutes into the men’s room at the mall while my mom and I are right in front of the door, waiting for them, even though my dad is not on the pre approved list as he has epilepsy. Or me not having an epi pen at my home because I gave it to Fitz as his was about to expire and it was his turn to pick up Tim who has severe allergies, even though our agreement said to have one at all times. (Just to make it clear, I bought a new one before getting Tim back). He also tried to paint me as a danger for having had seen a psychiatrist in highschool and early college. The reason was an eating disorder, nothing that could be any danger, which Fitz was aware of.
Even though all his so said proofs and breaches of contract were silly, with the most serious one described in the linked post, the trial got very intense, with him attacking me on a personal level. While we live in a big city, it quickly became the biggest gossip in our community, being on the lips of all the moms from pre school, most of them siding with me. Unfortunately, as much as I tried and couldn’t keep it all from Tim, since he was asked lots of questions by the kids at pre-school who heard about the situation. He became very stressed and started having night terrors.
I tried to reason with Fitz for Tim’s sake with no luck. 2 days ago, Fitz’s sister, Mia, came to see me. She wanted to give me something that could end all the stress Tim was under.
Almost 3 years ago, Fitz’s dad had a heart attack and almost died, but made a miraculous recovery. While he was in a coma, Fitz and Mia took Tim to the hospital to see his grandfather for what was thought to be the last time. Fitz’s entire family agreed there was no point in telling me, since I was very empathetic about the entire situation and even suggested that Tim and I could go visit him in the hospital once I picked him up from Fitz. They refused, however, since Tim already visited him, but did not mention this to me. She gave me texts to prove everything she was saying, and showing that Fitz knowingly breached our agreement and hid this from me for years on purpose.
My lawyer says that since this event occurred before all the ones he tries to use against me, it could get the entire case dismissed immediately, since it would have voided the parenting agreement we had. He guarantees me a win nonetheless, it would just take longer if we don’t use this. My parents think I should be the bigger person and not use it, as it is cruel and I should be an example for Tim for when he will be older. They insist that I would be in the wrong, especially since I would win nonetheless. I, however, just want to end this nightmare for good, and take care of the emotional health of my son. So what should I do?

submitted by No-Storm-8920 to Advice [link] [comments]

2024.05.16 01:18 noniac Why the next total war could be nothing like a total war

So I've recently watched Legends video about thrones of decay, a month old now, apparently although speculation, it seems there is a ww1 game in the works and 40k "is definitely coming" (even before the ww1 game). Yes, everybody wants 40k in every comment section box I've seen, yet no one seems to put a little thought on what bringing these modern settings would mean for Total War. Then let me explain:
Why TW game core is not made for a modern setting.
-Battles-
Projectiles and projectile based combat. Total war has always had trouble implementing projectile combat, be it stones, arrows or muskets but modern firearms and what comes with them are a totally different beast, and no, ratling guns are not a good example of "modern firearm". There is something else in these settings, cover mechanics (trenches, walls, scenery objects like crates, rubble, railings) and in-building traverse fighting/cover or building destruction... And they still can't even figure street pathfinding in cities.
Although Warhammer has brought single entity units they are like commanding any other unit just with one model instead of 120, these settings require soldiesquad commanding in a very different way, more in-depth, think of men of war (the actual game devs that should make a 40k game). Soldier equipment, stance, cover and position matter.
Vehicles. Yes, Warhammer has brought vehicles, but again they are just single entity units not actual vehicles but a massive animated single entity, a ww1/WW2 or 40k game would need real manned equipment, field guns and tanks, these are very different.
-campaign-
This is the real problem, the actual total war as a concept killer. They have truly innovated in Warhammer, flying units and and magic must have been a pain to code and with enough time we could see the above battle problems solved, but the problem is: battles? what battles? We have FRONTLINES here mate.
Current campaign system just does not make sense for a ww1/WW2/40k setting. Otto Von Bismarck with "19 units" sieging Paris garrisoned by a french general with "19 units" is just ridiculous. This system just cannot work from the American civil war (1850) onwards the same as hearts of iron IV cannot portray Napoleonic wars or antique Roman warfare. These settings cannot be based on single battle army collisions, yes, there are important battles that take place in certain spots along the frontline and keep happening when it's pushed or pulled back, but there is always a frontline and military presence with combats happening along it's whole length. They could somewhat "leg it" with 40k due to the tabletop army versus army origins, it would still look ridiculous if they want to approach it with a realistic look like they have done with fantasy plus you have to add you attack entire planets in 40k (where there should be frontlines tbh)
-conclusion-
If CA brings ww1 and/or 40k they might end up doing what I call a "Cossacks" (I guess dawn of war did it, a couple times indeed, with each game iteration actually). There is this game called Cossacks 2, it's nothing like Cossacks was at that point or has been later, I liked it, preferred it even. The matter of the fact is they slapped the franchise name on a game that had nothing to do with it except the historical time frame, it was mechanically completely different and rightfully many did not like it, unlike myself for example.
So this is why the next game setting could radically change what total war is from top to bottom (if they want to do it right, of course) to the point of being a completely different game, for better or worse.

submitted by noniac to totalwar [link] [comments]

2024.05.16 01:15 kandoras Help with Stunning Fist macro / module [PF2E]

I just picked up the level 2 monk feat stunning fist, and I'd like to not annoy my DM by asking her to switch over to an enemy's sheet and roll a fortitude save every turn, then figure out the result versus my class DC, and then apply the stun effect if they roll low enough.
I'd like to find some kind of macro or way to have the chat pop up with a "roll fortitude" and then figure out the result, the way she's got it set up for some other stuff (I'm not sure if that's in the base setup or through some module).
She found a macro for the prerequisite attack of Flurry of Blows, I believe in the PF2E Action Support module. But I can't find anything in it, or any other module I've looked in, for stunning fist.
I've tried my hand at writing a script to do it, but the basic examples I found to start off from were out of date. I figured out how to make a chat macro and pull numbers like fortitude save off my character sheet to get
/r 1d20 + @actor.data.data.saves.fortitude.value
and that worked right.
But when I tried it as a script in
let roll1 = new Roll(1d20 + @actor.data.data.saves.fortitude.value).roll({async:false});
I got an error in the console "You are accessing the CharacterPF2e#data object which is no longer used. Since V10 the Document class and its contained DataModel are merged into a combined data structure."
So if anyone has a current and basic starting guide for macros, that'd help too.

submitted by kandoras to FoundryVTT [link] [comments]

2024.05.16 01:13 No-Storm-8920 WIBTAH to use the coma of my ex’s father in the custody battle

submitted by No-Storm-8920 to AITAH [link] [comments]

2024.05.16 00:59 reddit_learner_help OpenAI Whisper - How to upload an audio file to SharePoint to then be transcribed by Whisper.

Howdy! I'm struggling to figure how i can utilise OpenAIs Whisper model to transcribe an audio file that I upload to a specific folder in SharePoint. I've got this to work by recording on a Power App but I would like to be able to record on any device, then simply save it a SharePoint folder and the audio be sent for transcription. The trigger is a 'When a file is created (properties only), I then retrieve the contents of the file using 'Get file content', i'm now stuck on how successfully pass this to Whisper. My HTTP request follows the format of this tutorial by Mohamed Faleel although this again is specifically for Power Apps whereas i simply want the trigger to be an uploaded file to a SharePoint folder so that i can record on another device. Can anyone shed some light on how to achieve this, I'm truly stuck with the HTTP request and would love it if you could send through an example of how to get this to work.

submitted by reddit_learner_help to PowerApps [link] [comments]

2024.05.16 00:49 Toker_X Material Listing?

Is there a listing of what materials are used for the standard production knives? For example, what are the thumb slides made of on most models?
I have a Hera II Mini and I'm guessing the thumb slide is aluminum. I know the pocket clip is titanium. What about the thumb slide for the Glykon?

submitted by Toker_X to microtech [link] [comments]

2024.05.16 00:33 TheoStillPlays Budget Steam maachine

the cheapest possible parts list to make a steam machine

Is this a brand new build, or an upgrade to an existing build?

A brand new build

Please list any existing parts or monitor(s) you have that you would like to re-use with this build. For upgrades, a PCPartPicker list of your full build is extremely helpful. Be as specific as you can be here, including links or exact model numbers of each component whenever possible. -Razer raptor 27, razer keyboard+mouse, Steam Controller

What will this PC be used for? Examples include things like gaming, general/multimedia use, photo/video editing, coding, AI/ML, etc. Include specific games and applications you intend to run, and any particular performance goals you have, as each may have different specific hardware needs. -General games such as Cyberpunk 2077

What country will you be purchasing in? If you are in the US, do you live near a Micro Center? For other countries, please check if your country is supported by PCPartPicker by using the country selector dropdown on the top right - if not, please provide some links to reliable local vendors you are comfortable ordering from. -The United Kingdom Of Great Britain and Northern Ireland

Do you need one or more monitors included in the budget? Please list how many and any size/resolution/refresh rate preferences if needed.

no, using an old monitor

What is your preferred and maximum budget range for this build, in local currency? Parts lists may sometimes have additional shipping costs. Please note whether prices in your country include sales tax or not, and adjust your budget accordingly. Typically VAT countries will have it included in the part list prices, whereas regular sales tax countries like the US and Canada will not. -Ideally around £100 but can be increased to £400 if no cheaper options available

Do you need WiFi, or do you have a wired ethernet connection available? -I have an ethernet connection.

Do you have any specific size or noise requirements for the build? -yes, Height: 2.17" (55mm) Depth: 7.87" (200mm) Width: 7.87" (200mm)

Do you have any aesthetic preferences for color or lighting? Describe what you're looking for, or feel free to provide some links to examples that may help. Some people prefer an inobtrusive stealth build, while others may prefer a case full of rainbow RGB.

No thanks, plain case.

Any other specific requests or requirements? Examples might include a specific minimum amount of storage, or a particular CPU socket for a future upgrade path, etc. Requires the following parts or better: Processor: Intel or AMD 64-bit capable processor Memory: 4GB or more RAM Hard Drive: 200GB or larger disk Video Card: NVIDIA graphics card AMD graphics card (RADEON 8500 and later) Intel graphics Additional: USB port for installation UEFI Firmware (recommended)

submitted by TheoStillPlays to buildapcforme [link] [comments]

2024.05.16 00:29 IveGotRights (WTS) AIMPOINT 5000 Gordon Delta

Timestamp: https://imgur.com/a/PYdCvHC
Looking to sell an original Aimpoint 5000. This is the original model, in black...just like the special forces used in the 1990'S and early 2000's. Perfect for your retro carry handle build or that Gothic Serpent clone.
This red dot has some wear on the body from use. The lenses are clear and free from scratches. Perfect example to paint for that Gary Gordon clone. The extension is in good shape, but the trim ring is rough. However, it looks better if you just unscrew it and remove the ring and/or extension. That's just my opinion. Again, really wouldn't matter if you painted it either.
If you have any questions, please feel free to ask. This dot works as you can see, but will ship without the battery.
Asking $975. Paypal f&f prefered but other options possible.
Thanks for looking!
Pictures: https://imgur.com/a/YvAKxhK

submitted by IveGotRights to GunAccessoriesForSale [link] [comments]

2024.05.16 00:19 cheinyeanlim How scientists are determining LLM capabilities

In the midst of product launches from both OpenAI and Google, it feels important to understand just how capable LLMs are. Both companies, especially in their demos, are really good at making their AI models appear intelligent, but scientists have to sink into the weeds (and through a couple of black boxes) to actually understand what’s going on here. And often, all they can do is make educated guesses, since these companies tend to keep their training data under lock and key.
Stay ahead of the curve with the latest trends in tech and marketing – join our subreddit community []() today for instant notifications!
How scientists are determining LLM capabilities
AI researcher Melanie Mitchell recently broke down the problem in a bit of depth.
The challenge:

Mitchell said that, even as LLMs might excel on certain benchmarks, we don’t “know the extent to which the test items in a benchmark appeared in the training data.”It’s hard to tell if models are reasoning, or if they are good at pattern retrieval based on their training data. As one computer scientist once put it to me: Imagine a law student who reads textbooks and passes the bar exam, versus a student (with a photographic memory) who was trained on thousands of bar exams, then passes the bar http://exam.In this example, if there were no bar exams in the training data, an AI passing the bar exam becomes a much bigger deal.

The Solution: The counterfactual task paradigm

In this paradigm, models are tested on pairs of tasks that require the same type of reasoning:The first features content likely in a model’s training data.The second features content unlikely in the training data.Using this, scientists can determine the extent that reasoning might be present in an LLM.
“In short, LLMs seem to have some ability to reason, but without stress-testing them (e.g., with counterfactual tasks), one can’t conclude that they are reasoning in a general way rather than relying on their training data in ways that won’t generalize to out-of-distribution examples.”

submitted by cheinyeanlim to martechnewser [link] [comments]

2024.05.16 00:12 Fast-Goat9040 I am planning to recreate Roxy's green room from FNAF Security Breach

^{I am planning on recreating Roxys green room, I understand it will be hard but I have a slight idea on how it may be done. First buy some materials for the frame, Making it 40% smaller would increase the probability of it being completed. It may require Alot of digging, Money, And planning. But with the right courage I think its possible. It could take months to even years}
^{The interior could be easily recreated, For example the sign that says Roxanne wolf could be recreated using neon sines, Could repaint a old go-kart with the design from the game.}
^{I may possibly create a Roxy model for it as well.}
I assure everyone I will do it

submitted by Fast-Goat9040 to FNaF [link] [comments]

2024.05.16 00:03 Kalcinator Help Needed: Persistent nvlddmkm.sys errors on GTX 980

Hi everyone,
I'm experiencing persistent "nvlddmkm.sys" errors with my MSI GTX 980. These errors occur mainly when I play graphically demanding games. For example, World of Warcraft runs perfectly without any graphical glitches, but it still crashes even when the graphics settings are at the minimum.
Here's a detailed breakdown of what I've tried and the steps I've taken so far:

Updated Drivers: I've made sure to have the latest NVIDIA drivers installed.
Stress Test: Ran a stress test with everything maxed out for an hour using OCCT, and it showed no errors.
Tested with Less Demanding Games: No crashes when playing Hearthstone or other less demanding games.
Reinstalled Drivers Using DDU: Used Display Driver Uninstaller (DDU) to completely remove and then reinstall the latest drivers.
Checked for Physical Issues: Ensured the GPU is properly seated and all cables are connected securely.
Ran System Tests: Tested CPU and other system components with no issues found.
Manual intervention: Replaced "nvlddmkm.sys" but it doesn't work (deleted now)

Despite these efforts, I continue to encounter the following error: The description for Event ID 0 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted.
The following information was included with the event:
\Device\Video3
Error occurred on GPUID: 700

Findings & Others

The nvlddmkm.sys file is missing from C:\Windows\System32\drivers
I've got no other PC to test my graphic card, but I've got my "old" GTX 960
I don't really want to intervene on the BIOS of the card
Using MSI Afterburner to slightly reduce the GPU clock and memory clock ...

If anyone has additional suggestions or has faced similar issues and found a solution, your help would be greatly appreciated. I'm at a bit of a loss and could really use some expert advice.
Thanks in advance !
Edit for System Specifications:

CPU: AMD Ryzen 5 2600X, 6-core, 12 nm, 3.6 GHz
Motherboard: Gigabyte B450M DS3H-CF, BIOS version F65c
RAM: 16GB Corsair DDR4-3200 (CMK32GX4M2E3200C16)
GPU: MSI GTX 980, 4GB GDDR5, 256-bit

submitted by Kalcinator to pcmasterrace [link] [comments]

2024.05.15 23:54 dotmax_it Full Nat/Port forwarding based on phisical interface

Hi to all, I have to solve this problem:
I have 3 device that have a fixed address (192.168.10.1). The ip addres cannot be modified in any way (don't ask me why...). I want to place the 3 devices on 3 different pfsense interfaces, and configure the wan address with a static ip (192.168.11.1) plus 2 virtual ip (192.168.11.2 and 192.168.11.3) And i want to match the wan ips with the lan ips, but based on the interface. Example: If I connect a PC on whe wan interface with ip 192.168.11.10, and I start a connection with the ip 192.168.11.1 I want to send packets to host 192.168.10.1 on interface 1 (and receive the answer back), but if I start a connection with ip 192.168.11.2 I want to send packets to host 192.168.10.1 on interface 2. Same for ip 192.168.11.3 and host on interface 3
I know for sure that microtik can do this. (I'm waiting for the model name and sample config file from a colleague who already use it) But i want to do this with pfsense. Is this possible?
Thank you

submitted by dotmax_it to PFSENSE [link] [comments]

2024.05.15 23:53 Hopeful_Anxiety7393 Advancing Nursing Practice: A Call for Continuous Improvements

In the ever-evolving landscape of healthcare, nursing stands as the cornerstone of patient care. As we navigate through the complexities of modern medicine, it becomes imperative to foster an environment of continuous improvement within the nursing profession. From refining working habits to advocating for better resources, here are some examples of nursing improvements worth discussing:

Optimizing Workflow Efficiency:

Nurses often face challenges related to time management and workflow efficiency. One improvement could involve implementing technology solutions such as electronic health records (EHR) or mobile applications to streamline documentation and communication processes. By reducing administrative burdens, nurses can dedicate more time to direct patient care, ultimately enhancing outcomes.

Enhancing Interdisciplinary Collaboration:

Effective collaboration among healthcare professionals is essential for delivering comprehensive patient care. Nurses can advocate for interdisciplinary rounds where physicians, nurses, therapists, and other team members come together to discuss patient cases and develop cohesive care plans. By fostering open communication and mutual respect, interdisciplinary collaboration can lead to improved patient outcomes and satisfaction.

Promoting Staff Well-being:

Nursing is a demanding profession that often takes a toll on the physical and emotional well-being of nurses. Implementing initiatives to support staff well-being, such as mindfulness programs, mental health resources, and flexible scheduling options, can help alleviate burnout and improve retention rates. Prioritizing the health and happiness of nurses ultimately translates to better patient care.

Advocating for Patient-Centered Care:

Nurses play a pivotal role in advocating for the needs and preferences of patients. Encouraging shared decision-making, implementing patient-centered care models, and providing cultural competency training can empower nurses to deliver care that is tailored to each individual's unique circumstances. By prioritizing the patient's voice and preferences, nurses can foster trust, enhance satisfaction, and improve health outcomes.

Embracing Evidence-Based Practices:

In an era of rapid medical advancements, it's crucial for nurses to stay abreast of the latest evidence-based practices. Establishing regular education and training programs, promoting participation in research initiatives, and creating forums for knowledge sharing can help nurses integrate evidence-based interventions into their practice. By aligning care with the best available evidence, nurses can ensure optimal outcomes for their patients.

Addressing Healthcare Disparities:

Nurses have a critical role to play in addressing healthcare disparities and promoting health equity. By advocating for equitable access to care, addressing social determinants of health, and participating in community outreach initiatives, nurses can contribute to reducing disparities among underserved populations. Through collaborative efforts, nurses can work towards creating a more just and inclusive healthcare system.
In conclusion, nursing improvements are essential for advancing the quality and effectiveness of patient care. By actively engaging in discussions and implementing innovative solutions, nurses can drive positive change within their practice environments. Let's continue to strive for excellence, advocate for our patients, and work together towards a brighter future for nursing.
Let me know your opinion in the comments and be free to say what would you change in your work enviroment! 🫶🏻

submitted by Hopeful_Anxiety7393 to nursingimprovements [link] [comments]

2024.05.15 23:50 InterestingPepe gluing and painting order. In the instructions it tells you to paint a wheel or part a certain color but it can’t be reached when I glue it together, but when I glue it after painting it won’t stick as it needs to melt through the paint. What do you guys do to paint these parts and glue them

Am i supposed to figure out what part of the part can’t be painted so it t can be glued ? For example you are building a transmission in a model car and each section needs to be painted a different color, but if you were to glue it all together you can’t really paint the tiny parts without a lot of masking tape and then you can’t even mask it correctly as it’s glued together and fragile and the tape will pull the pieces apart. Sorry I’m still new

submitted by InterestingPepe to modelmakers [link] [comments]

2024.05.15 23:49 ezxllabtw I feel like my parents care more about my grades than anything else about me.

This is mostly a vent post but I need to get what I have to say out.
Mostly, when I talk to my parents it either about my grades or my room. It always "Why aren't you getting your school work done" or "If you get a C or below your gonna get whooped". It's like I know they want me to do well in school but it also stresses my out trying to always fit up to their standards. I've been doing the best I can with my school work and grades because I mostly work in a group with other people and they haven't been doing any work so I've been doing all the work alone, and it's tiring.
I know my dad wants me to do well in life but he's always yelling to me about my grades. He's a teacher so he likes to compare me to his students or compare how they act. For example today he looked at my grades and he was like "The straight A's students at my school don't even have a phone or they pay attention and do all their work". It's so annoying because If i try to do my work I get a good 10 minutes before I can't focus on my work again.
I just wished they cared about how I feel because I feel drained. I have no motivation to do my school work or do anything for that matter. I know they love me and I love them but sometimes it just doesn't seem that way. It's like they want me to be some perfect model for them and I'm tired of it. I wish they would just talk to me or let me speak my truth to them but if I do speak my truth than I will get in trouble because I'm going to be "talking back". My mom said she wanted me to state my opinion and talk up to my dad but when I do he just yells at me or doesn't even hear me out.
Anyways I just needed to vent and get this out but if anyone has advice on what to do please tell me.

submitted by ezxllabtw to emotionalneglect [link] [comments]

2024.05.15 23:44 SIDATI666 How i've grown a business to an 8 figure valuation

Hello all,
First time posting abou this and TLDR I founded a company 18 months ago and our recent round values our business at +$10m
I was a late starter to entrepreneurialism and had a good length stint in finance (trying one startup while still being employed) before founding my current business. My family has an entrepreneurial background and because of this, I have always had ambitions for building a business but I genuinely was too afraid to take the risk (for whatever reason). Having now ran the company for just over a year, we have been a lot more successfully than i could have anticipated which has culminated in a proper valuation. While we still have a long way to go (you never know what happens), I guess you have to enjoy the milestones when they come
Some things I have learned.

Take the risk - if you are fortunate to find a problem to which you can realistically create a solution, you are on the right track. If this solution has a big enough market (it doesn't have to be a multi billion opportunity) and you can find some potential customers who confirm they would pay, you are wasting your time not taking the risk.
Markets don't have to be $100bn in size. VC's will often say that the addressable market needs to be ginormous to even attempt. While it may be true for them to make their 10x return, it is not true for you. If you can get 1% of a $100m market which has high margins and is easy to access, you can begin to build a valuable business.
Stay close to home - my first attempt at creating my own business was in a market that I saw from a distance. Because I didn't have direct business involvement in this market, I could not see the actual pitfalls of the industry and how my assumptions underlying my business model were flawed from the outset....
Don't throw good money after bad - the only good part of my first business (that failed) was that I quickly saw the flaws in my model, determined I could not solve them and closed the business, staying in my current job. I didn't take angel investment which meant that my current business was not tarred by that early opportunity.
Staffing is COMPLICATED - anyone can build the ideal model for an ideal business. Very driven people can even build the infrastructure and launch the business, but the next step is building out your staff. By definition each incremental staff member is less incentivised than the earlier staff member (less options, less responsibility). I have found running a team of staff one of the most complicated parts of our business. It is a completely different skill-set to turn your employee base into a devoted tribe and I have made plenty of mistakes along the way.
Keep the burn down - It is not sexy to boast about how much money your start-up burns. Contrary to popular belief, businesses are supposed to make money. In building the business, I have built our software in steps. For example, when I wasn't making that many invoices, I didn't need to automate invoices. Furthermore, I have built the business (a software platform) as revenue generating features first and useful back end improvements second. I was also more apprehensive adding staff in the early stages which meant our initial growth generated strain and long hours. We have now exited that part of our growth but keeping a tight ship is the only way to show fiscal responsibility.
Have a co-founder - It's fucking tough. You will be in tears, sweats and anxious at some points, at others you will feel the highest of highs. It's a grind and a rollercoaster; most people are normal human beings, you need someone to balance you out. When we went through two quiet days, I was panicking that something fundamental had gone wrong. My co-founder helped me rationalise that business isn't a straight line.
"Why can't a larger player just copy you" - you will hear this from any investor you pitch to, your employees and yourself. If that statement were true, there wouldn't be millions of small businesses in the world. Clearly there is some complexity to copycatting. Most times you don't realise that big players don't care about you and they are unlikely to copy a minnow and spend millions of dollars in the process. If you have something truly unique, a purchase is more likely.
Go into existing markets, don't create new ones - The big ideas that haven't been done before sound super sexy. "Did you know 200 million cups of coffee are drunk every day in the States, we are going to make a Deliveroo for coffee! Just 1% will make us MILLIONS". Creating a completely new market requires MASSIVE capital, MASSIVE time and constantly fails. Look for a market that exists, create a business that segments that market by solving a particular problem. Your customers already know you, they already feel the problem, when you present the solution, some will buy in.
Know your value precisely if you're exiting - If you're considering exiting your business, knowing its precise value is crucial. (easter egg) Our valuation tool can give you an initial estimate of your business's worth. For a more precise valuation, once you fill out the form, an M&A expert from our team will reach out to provide you with a detailed, free valuation if you're interested.

I'm sure I am missing quite a few other important lessons. It would be great to hear from other founders

submitted by SIDATI666 to EntrepreneurRideAlong [link] [comments]

2024.05.15 23:42 Caw-zrs6 Some things I noticed regarding Corporate Clash.

A few days ago, I went and made a CC account because Toontown is a game I used to play back in the day, but much like what the title says, there's some things that made me feel a little confused, to say the least.
The first of which is the lack of boosts from the fire hydrants, trashcans, and mailboxes. I remember back in the original game, whenever a fight was initiated near one of these three objects, they would give a boost to certain Gags (i.e. fire hydrant boosting Squirt Gags). Why are these boosts no longer a thing in CC?
The second is the fact that some of the Cogs don't possess updated head models. Now this is the thing that really has me scratching my head considering these new heads for the Cogs just look so good, and they help breathe new life into otherwise lifeless machines. Plus some of the Cogs could honestly do with a new look.
As an example, the Cold Callers. Back when the original Toontown was still a thing, there were these trading cards of various things regarding the game, such as characters, Gags, etc. One of these cards was for the Cold Caller, which depicted its head as an ice cube. In the game though it's literally just a more blue version of the Short Change's head.
Anyways, just felt like sharing some observations I had regarding this version of Toontown, feel free to discuss in the comments if you want.

submitted by Caw-zrs6 to Toontown [link] [comments]

http://swiebodzin.info