DEV Community: Rwik Mukhopadhyay

A quick look at Ai-based art generation

Rwik Mukhopadhyay — Wed, 24 Aug 2022 05:39:24 +0000

"Gone are the days when you have to spend a few precious years of your life in Art colleges, to create Art. Or you have got to born with sheer tallent": Well, we are almost there. Have a look here .

Welcome to the world of AI based art generation. Where you are only limited by your imagination. You just think about what you want to see and AI will do the rest of the work.

How it works :

In general, you may assume these art generators are basically AI systems trained in publicly available datasets. You just have to write down your imagination in a descriptive manner, and the AI system will generate art accordingly.

Here is my first attempt. I have keyed down "Wild Buffalo's dancing with bees and butterfly in desert under starry night beside fire". And the result is :

Well, it's not exactly what I wanted, but I kinda like it.
You may try to create your own art here.

Tools to explore :

https://www.midjourney.com/home/

Probably the most talked about generator. You may try it for free initially (I guess up to 25 images). They have a discord server backed by a very helpful community. Results are realistic, and you can control output up to a great extent. For example, if you want to create a comic book, you need to create same character faces in different scenario. Kris has done exactly that using midjourney. Have a look here .

https://creator.nightcafe.studio/text-to-image-art

Very simple to use. Also, you can select among popular artistic styles. I found it good for generating some nice wallpapers.

https://app.wombo.art/

This is an app, but you can also use it on the web. You can create art then sell it as NFT or buy it as high quality print. You can also remix your existing NFTs.

https://huggingface.co/spaces/dalle-mini/dalle-mini

Dall-E 2 is probably the most anticipated tool for AI based image generation through natural language processing. But sadly, it's not available yet. You may join the waitlist here . Till then, you may try dalle-mini. I didn't find it much useful compared to others, but I think I need to spend more time on this.

So how is this useful for you ?

Well, Application of AI in any field is unlimited or restricted by imagination only. Below are a few use cases :

Banner / Art work for social media.
Illustartion / Art work for articles / books.
NFT creation
Wallpaper :)
Graphic Novel

Conclusion :

I feel these AI tools will do the same to artists, what autotune did to voice artists. But still a long way to go for this. But remember, AI tools are trained in open internet filled with copyrighted artworks. So you may get copyright strike in case your artwork is similar to any copyrighted works. But, early adopters always make it big.

Clang vs GCC

Rwik Mukhopadhyay — Sun, 01 Mar 2020 04:51:55 +0000

Clang is relatively new in cpp world. Clang came out of apple’s stable and became open sourced in 2007. Apple uses LLVM extensively. For some unknown reason they choose to drop gcc’s front end , and create a new compiler front end from scratch. Thus Clang is born .
GCC’s main popularity lies in the fact that it is the only option to compile Linux kernel. So if you are working in *nix world GCC is surely your daily driver. Also GCC supports more legacy languages ( like Fortran, ADA). But on the other hand Clang+LLVM reduces compilation time for single threaded applications by 5-10% . For a more comprehensive benchmark comparison check here .

Inside Node.Js I/O

Rwik Mukhopadhyay — Sat, 29 Feb 2020 16:39:07 +0000

Node.Js introduces iteslf as a asynchronous event driven javascript runtime. To acheive this asynchronous nature , node uses a open source library called libuv.

libuv even though built for Node , it is not exclusive to Node.js. There are plenty of other projects which uses libuv.

While network i/o is asynchronous , same can't be said for disk i/o .
libuv uses single threaded operation to use event loop and produce asynchronus network i/o . Mostly lib pool apis are not thread safe as it is designed to work in single threaded mode only.

Disk i/o has some platform specific differences . Every majore APIs for disk I/O has it's own limitations, also this increases complexity. So libuv uses blocking disk operations using thread pool.