it took a lot of time and i kept doing the same steps over and over because you didn't speak in a clear way so some times i didn't understand you in general I'm not programmer or anything of that and i didn't install visual studio before but in the same time i didn't watch any one else because i loved your way i don't know why keep going you are amazing honey
When I later wanted to train the gaussian-splat I got an error about cl not being in my path, so I went into environment variables and added "C:\Program Files\Microsoft Visual Studio\2022\Community\VC\Tools\MSVC\{VERSION_NR}\bin\Hostx64\x64" to the path and now it works.
Hi! I have the same error here. I am trying to do the same as you, but no success. Super newbee here in regards to this world, especially to programming. Can you please write steps on what is needed to be done? Hope for your reply... I've also posted that issue to Github but no answers for now.
Great video! Have you done an install with zip-nerf? Have tried to do the setup for days but having an issue on compilation: #error: You need C++17 to compile PyTorch
Hi Johnathan, amazing series of videos I have been following and learning so much from them thank you. Question, I updated to Nerfstudio 24 as you showed here no problem, but I did that for the new functionality of the viewer and render options. I still want to access previously trained .ply files - but when I try to use the same code / technique from your older videos on how to run GS in NS - it doesn't work the same way. Is it still possible to run pre-trained .ply GS via the new Nerfstudio viewer - I can't see a Gaussian Splat file in the script folder for example, so not sure if you can? Thanks.
Hello, I think I have installed the nerfstudio successfully and tried to train my first nerf model. But getting this error 'ns-download-data' is not recognized as an internal or external command, operable program or batch file when running download some test data command. I think it's a command problem or I didn't installed nerfstudio right?
that's great and all, but the problem really begin with data processing, as "ffmpeg" isn't in path, then COLMAP isn't installed. you're describing the easy part only ;)
@@thenerfguru maybe i binge watched too much of "curb your enthusiasm" lately, but is it really an installation, if the core functions are not working? :D isn't it just a nerf viewer at this stage? i'd say that creating is at the core of the studio experience of nerfstudio. anyway - colmap on windows is quite random and a true pain.
@@nAtic86 Where are you finding it hard to install? If you use binaries, you just download them and path them to your system variables. No code compiling or anything. Same goes for FFMPEG.
Hi Johnathan, thank you for your video! Im wondering, could you help me with this. I got $2.000 , and i want to buy a laptop that works with NERFSTUDIO and UNREAL ENGINE. What laptop can i get for this?
i already have nerfstudi on my system and this apps running well, do i need to delete the apps, delete environment and start from 0? or what step to save time, i only need update to the latest version of nerfstudio
Unfortunately, I don’t think that the GPU will perform well. But I’m not 100% sure. The original splatting project required RTX GPUs, Nerfstudio may be different. I would ask on the Discord before wasting time.
with tiny-cudy-nn I basically get this error \AppData\Local\Temp\pip-req-build-m9oa1ybh/dependencies/fmt/include\fmt/format-inl.h(11): fatal error C1083: File(Include) "algorithm": No such file or directory basically the fmt folder is empty it seems and when I googled this error I found what looks like an answer but the guy that solved the problem only said that I should manually download those missing folders and place them in the "corresponding" folder but I have no clue what that means. When I run that command it searches for that folder in a temporary folder in temp, that gets deleted after the command is run and also always has a different name. So there is no way to put it in a "corresponding" folder. It's kinda frustrating that exactly this part is not specified by that commenter and since I'm not deep into any of the technicallities I am pretty helpless. Do you know how I can solve this error? I have no idea how to install a local github repo as well and found nothing that helped me.
I am trying here... But I got error: "(...) /14.38.33130/include\crtdefs.h(10): fatal error C1083: Cannot open include file: 'corecrt.h': No such file or directory" or at one misses math.h file. I tried to solve it with a vcvars64.bat solution, reinstalling VisualStudio, adding Windows Universal CRT SDK... nothing seems to help :(
@@JakubPeka1 I saw a few promising threads when I searched the problem on Google. They are all fairly old though. Looks like CUDA isn't finding the right file location.
![PURPEECH - นี่ฉันเองคนที่... (It's me) [Official MV]](http://i.ytimg.com/vi/NuhHd0hOOg8/mqdefault.jpg)
it took a lot of time and i kept doing the same steps over and over because you didn't speak in a clear way so some times i didn't understand you in general I'm not programmer or anything of that and i didn't install visual studio before
but in the same time i didn't watch any one else because i loved your way i don't know why
keep going you are amazing honey
When I later wanted to train the gaussian-splat I got an error about cl not being in my path, so I went into environment variables and added "C:\Program Files\Microsoft Visual Studio\2022\Community\VC\Tools\MSVC\{VERSION_NR}\bin\Hostx64\x64" to the path and now it works.
Good solution. This is a common issue I have found.
Hi! I have the same error here. I am trying to do the same as you, but no success. Super newbee here in regards to this world, especially to programming. Can you please write steps on what is needed to be done? Hope for your reply... I've also posted that issue to Github but no answers for now.
Issue is fixed. Anyone can find solution at Issues tab of nerfacto github. Issue number is 2820. Hope it will help someone
Muchas gracias Maestro, un saludo desde Arauca Colombia.
You're welcome!
Great video!
Have you done an install with zip-nerf? Have tried to do the setup for days but having an issue on compilation:
#error: You need C++17 to compile PyTorch
just one gotcha I didn't wanna mess with my Nvidia driver so in Cuda installation you can select only CUDA to install and leave the drivers alone
Good tip!
Hi Johnathan, amazing series of videos I have been following and learning so much from them thank you. Question, I updated to Nerfstudio 24 as you showed here no problem, but I did that for the new functionality of the viewer and render options. I still want to access previously trained .ply files - but when I try to use the same code / technique from your older videos on how to run GS in NS - it doesn't work the same way. Is it still possible to run pre-trained .ply GS via the new Nerfstudio viewer - I can't see a Gaussian Splat file in the script folder for example, so not sure if you can? Thanks.
C:\Users\Administrator>conda create --name nerfstudio -y python=3.8
'conda' 不是内部或外部命令，也不是可运行的程序
或批处理文件。
C:\Users\Administrator>
Hello, I think I have installed the nerfstudio successfully and tried to train my first nerf model. But getting this error 'ns-download-data' is not recognized as an internal or external command, operable program or batch file when running download some test data command. I think it's a command problem or I didn't installed nerfstudio right?
that's great and all, but the problem really begin with data processing, as "ffmpeg" isn't in path, then COLMAP isn't installed. you're describing the easy part only ;)
I covered installing Nerfstudio. In my follow up video, I talk about those two things!
@@thenerfguru maybe i binge watched too much of "curb your enthusiasm" lately, but is it really an installation, if the core functions are not working? :D isn't it just a nerf viewer at this stage? i'd say that creating is at the core of the studio experience of nerfstudio. anyway - colmap on windows is quite random and a true pain.
@@nAtic86 Where are you finding it hard to install? If you use binaries, you just download them and path them to your system variables. No code compiling or anything. Same goes for FFMPEG.
Please make a tutorial for Mac Book Pro M2 Chip. Thanks Jonathan
I have a plan for that. It’s a different toolset, but makes Gaussian splats.
@@thenerfguru that would be great. I will be waiting it 😁 i think many people will apriciate it. Thanks Jonathan
Thanks a lot !! :-)
You’re welcome!
Hi Johnathan, thank you for your video! Im wondering, could you help me with this. I got $2.000 , and i want to buy a laptop that works with NERFSTUDIO and UNREAL ENGINE. What laptop can i get for this?
"error: metadata-generation-failed
× Encountered error while generating package metadata.
╰─> See above for output."
Any fixes with tiny-cuda?
That is a new one to me. I suggest asking in the original GitHub Project discussion area.
i already have nerfstudi on my system and this apps running well, do i need to delete the apps, delete environment and start from 0? or what step to save time, i only need update to the latest version of nerfstudio
Cool thanks
🙌
So if I am using Mac I can't download it?
What is min requirements
I have 4800h ryzen cpu and gtx1660ti
Is that will work good or slow
Unfortunately, I don’t think that the GPU will perform well. But I’m not 100% sure. The original splatting project required RTX GPUs, Nerfstudio may be different. I would ask on the Discord before wasting time.
@@thenerfguru ok i will as i need it alot for my graduation project ❤️
@@thenerfguru can i use 4070 intead of 3090 ? newer card with low watt
@@panonesia You certainly can! The only caveat is that the less VRAM you have, the fewer images you can process.
@@thenerfguru can you make video comparison? with nerfacto or nerfacto-big, how much vram need for process? let say with 200 image need 6gb vram etc
Very usufull but how you lunch it !?
Haha, you run it via command line.
with tiny-cudy-nn I basically get this error \AppData\Local\Temp\pip-req-build-m9oa1ybh/dependencies/fmt/include\fmt/format-inl.h(11): fatal error C1083: File(Include) "algorithm": No such file or directory
basically the fmt folder is empty it seems and when I googled this error I found what looks like an answer but the guy that solved the problem only said that I should manually download those missing folders and place them in the "corresponding" folder but I have no clue what that means. When I run that command it searches for that folder in a temporary folder in temp, that gets deleted after the command is run and also always has a different name. So there is no way to put it in a "corresponding" folder. It's kinda frustrating that exactly this part is not specified by that commenter and since I'm not deep into any of the technicallities I am pretty helpless. Do you know how I can solve this error? I have no idea how to install a local github repo as well and found nothing that helped me.
I am trying here...
But I got error: "(...) /14.38.33130/include\crtdefs.h(10): fatal error C1083: Cannot open include file: 'corecrt.h': No such file or directory" or at one misses math.h file.
I tried to solve it with a vcvars64.bat solution, reinstalling VisualStudio, adding Windows Universal CRT SDK... nothing seems to help :(
Hmmm, that’s a new one to me. I suggest asking in the Nerfstudio Discord or on the Nerfstudio GitHub discussion board.
@@thenerfguru I will :) think about trying on other device to be sure ;)
@@JakubPeka1 I saw a few promising threads when I searched the problem on Google. They are all fairly old though. Looks like CUDA isn't finding the right file location.
@@thenerfguru yeah I have spent few hours browsing and reinstalling but still :) thanks for your help :)) I will try again tomorrow :)