I'd rather use an app that adds tags to files instead of arranging them in a folder structure. There's a Mac app called 'Keep It Shot' that also does file renaming (not just for documents but also for images and videos too).
@@john_blues It’s included in Settapp, which I guess makes it a little more affordable if there are other apps also of interest to you in the bundle. But you were right to point that out as using AI can quickly become expensive.
No, this actually uses llama3 in the backend to read file content to rename the files accordingly and also to give a good directory structure to the files. :)
@@TheAnubhavShrimal yes i can see that ! but in truth it does not even need the GPT model to perform the task ! < its just a collection of tools with a n AI operators brother ! so basically a fake app : If the model Only had the OS.. as a tool and it performed these tasks itsself on the os.. then yes . but the whole procedure has been programmed so the model basically is a instigator and does the user operation... im not saying its bad ! ... But yes it is what it is : ( for me it has no value as a usuable artifact for other models ) in fact why would you need 70b llama to do such a simple task ?
You can use any smaller model too :) can be a 8B or smaller model. The hackathon team just used what they found viable at that moment. And it does perform the restructuring of the directory so definitely not FAKE. It works as promised:)
@@TheAnubhavShrimal yes you can probably use a slim model 128k params : thats how much the model is actually performing : The aim i would say is to find trainable pathways : So in this instance i would train this skill : just by giving it the operating system command as tools ( it does not need them as it csan generate a script to perform the task : ) so in truth currently you can ask a model which has operating system acces to perform this task: such as code interpretor which uses jupiter notebook to pewrfrom its commands : so train the model to use the operating system by writing script to performm these task: here the model or code does some extras ( such as peek inside a few of the files !)
If you’re following the instructions I have shown in the video using Groq API, it will run on the servers. Using ollama moondream enables on device compute
> start:main > cross-env NODE_ENV=development electronmon -r ts-node/register/transpile-only . D:\llama-fs-main\electron-react-app ode_modules\electron\index.js:17 throw new Error('Electron failed to install correctly, please delete node_modules/electron and try installing again'); ^ Error: Electron failed to install correctly, please delete node_modules/electron and try installing again at getElectronPath (D:\llama-fs-main\electron-react-app ode_modules\electron\index.js:17:11) at Object. (D:\llama-fs-main\electron-react-app ode_modules\electron\index.js:21:18) at Module._compile (node:internal/modules/cjs/loader:1376:14) at Module._extensions..js (node:internal/modules/cjs/loader:1435:10) at Module.load (node:internal/modules/cjs/loader:1207:32) at Module._load (node:internal/modules/cjs/loader:1023:12) at Module.require (node:internal/modules/cjs/loader:1235:19) at require (node:internal/modules/helpers:176:18) at module.exports (D:\llama-fs-main\electron-react-app ode_modules\import-from\index.js:4:47) at Object. (D:\llama-fs-main\electron-react-app ode_modules\electronmon\bin\cli.js:7:22) Node.js v20.10.0 [webpack-dev-server] Project is running at: [webpack-dev-server] Loopback: localhost:1212/ [webpack-dev-server] On Your Network (IPv4): 192.168.3.15:1212/ [webpack-dev-server] Content not from webpack is served from 'D:\llama-fs-main\electron-react-app\public' directory [webpack-dev-server] 404s will fallback to '/index.html'
@@TheAnubhavShrimal also i am getting the error TypeError paths.forEach is not a function Call Stack buildTree renderer.dev.js:49482:11 handleBatch renderer.dev.js:49532:26 × Close
Did you run npm install and then npm start. Also ensure you’ve followed all the steps mentioned in order and have latest nodejs installed with above python 3.10
Yes, i have mentioned in the video too that it was developed as part of the LLAMA3 Hackathon :) i have used it first hand and it works well. I’ll leave the choice to you whether you wish to use it or not.
I'd rather use an app that adds tags to files instead of arranging them in a folder structure. There's a Mac app called 'Keep It Shot' that also does file renaming (not just for documents but also for images and videos too).
Keep it shot has a very limited free version, uses an OpenAI API and has a monthly subscription based on 'credits used'. Just want to prepare people.
@@john_blues It’s included in Settapp, which I guess makes it a little more affordable if there are other apps also of interest to you in the bundle. But you were right to point that out as using AI can quickly become expensive.
For sure
em this is a fake AI app Bro !
the app works without AI !
No, this actually uses llama3 in the backend to read file content to rename the files accordingly and also to give a good directory structure to the files. :)
@@TheAnubhavShrimal yes i can see that ! but in truth it does not even need the GPT model to perform the task ! < its just a collection of tools with a n AI operators brother ! so basically a fake app :
If the model Only had the OS.. as a tool and it performed these tasks itsself on the os.. then yes . but the whole procedure has been programmed so the model basically is a instigator and does the user operation...
im not saying its bad ! ... But yes it is what it is : ( for me it has no value as a usuable artifact for other models )
in fact why would you need 70b llama to do such a simple task ?
You can use any smaller model too :) can be a 8B or smaller model. The hackathon team just used what they found viable at that moment. And it does perform the restructuring of the directory so definitely not FAKE. It works as promised:)
@@TheAnubhavShrimal yes you can probably use a slim model 128k params :
thats how much the model is actually performing :
The aim i would say is to find trainable pathways :
So in this instance i would train this skill :
just by giving it the operating system command as tools ( it does not need them as it csan generate a script to perform the task : )
so in truth currently you can ask a model which has operating system acces to perform this task: such as code interpretor which uses jupiter notebook to pewrfrom its commands :
so train the model to use the operating system by writing script to performm these task:
here the model or code does some extras ( such as peek inside a few of the files !)
Make this just run on Windows, don't have skill or time for all the python this and that.
Hopefully the creators will listen to your prayers 🙏🏻
Who cares about files) try to organize my family photos
Haha that’s true. Even AI will have a hard time doing that 😂
Is it available on Windows machine?
Most likely it should work. You can go to the GitHub repository in the description and try. :)
I wish it could handle videos
I agree, that would indeed be a great addition. :)
Interesting proof of concept. I'd be interested in testing if it supported more file types.
Does Llama model run on the server or on device? because I have a very low end device.
If you’re following the instructions I have shown in the video using Groq API, it will run on the servers.
Using ollama moondream enables on device compute
Very cool! Thanks for sharing
When will mobile users get a file manager this good?
Haha true that. I bet it’s not too far off a dream 😊
I having issue when I`m placing images in my folder
Yes, looks like there are some issues. You may want to raise an issue on their github repository or search in existing issues about it.
> start:main
> cross-env NODE_ENV=development electronmon -r ts-node/register/transpile-only .
D:\llama-fs-main\electron-react-app
ode_modules\electron\index.js:17
throw new Error('Electron failed to install correctly, please delete node_modules/electron and try installing again');
^
Error: Electron failed to install correctly, please delete node_modules/electron and try installing again
at getElectronPath (D:\llama-fs-main\electron-react-app
ode_modules\electron\index.js:17:11)
at Object. (D:\llama-fs-main\electron-react-app
ode_modules\electron\index.js:21:18)
at Module._compile (node:internal/modules/cjs/loader:1376:14)
at Module._extensions..js (node:internal/modules/cjs/loader:1435:10)
at Module.load (node:internal/modules/cjs/loader:1207:32)
at Module._load (node:internal/modules/cjs/loader:1023:12)
at Module.require (node:internal/modules/cjs/loader:1235:19)
at require (node:internal/modules/helpers:176:18)
at module.exports (D:\llama-fs-main\electron-react-app
ode_modules\import-from\index.js:4:47)
at Object. (D:\llama-fs-main\electron-react-app
ode_modules\electronmon\bin\cli.js:7:22)
Node.js v20.10.0
[webpack-dev-server] Project is running at:
[webpack-dev-server] Loopback: localhost:1212/
[webpack-dev-server] On Your Network (IPv4): 192.168.3.15:1212/
[webpack-dev-server] Content not from webpack is served from 'D:\llama-fs-main\electron-react-app\public' directory
[webpack-dev-server] 404s will fallback to '/index.html'
thanks for the guide
Thank you :) glad you found it useful
@@TheAnubhavShrimal also i am getting the error
TypeError
paths.forEach is not a function
Call Stack
buildTree
renderer.dev.js:49482:11
handleBatch
renderer.dev.js:49532:26
× Close
Did you run npm install and then npm start. Also ensure you’ve followed all the steps mentioned in order and have latest nodejs installed with above python 3.10
@@TheAnubhavShrimal yes i have done those
🙇thank you man❤
I got this Error!Can fix IT?
MY computer is Windows 10 OS。
Did you see that the code was developed for a comp a few months ago and has hardly been touched since. I would not be recommending it.
Yes, i have mentioned in the video too that it was developed as part of the LLAMA3 Hackathon :) i have used it first hand and it works well. I’ll leave the choice to you whether you wish to use it or not.
You could have Claude 3.5 update it for you
scam