Roman Suvorov (@windj007) profile picture

Deep Learning Research @ https://xitter.gq/t.co/YZ8uB4Cavn

Dubai
Joined August 2020
Photo by Roman Suvorov (@windj007)Photo by Roman Suvorov (@windj007)Photo by Roman Suvorov (@windj007)Photo by Roman Suvorov (@windj007)Photo by Roman Suvorov (@windj007)Photo by Roman Suvorov (@windj007)Photo by Roman Suvorov (@windj007)Photo by Roman Suvorov (@windj007)Photo by Roman Suvorov (@windj007)
Roman Suvorov retweeted
The fastest image model on the market is here. Zero lag in your workflow. Recraft V4.1 Flash produces images in 1.3 seconds. Pick Flash from the model picker, generate, then hit Refine to clean up the details. Check it out and give it a try ↓
🤖 Made with AI
27
26
47
229
581,500
Roman Suvorov retweeted
Two weeks since launch, and the public benchmarks are in. Recraft is officially the #1 independent image generation lab 🚀
14
12
8
103
1,785,840
Roman Suvorov retweeted
Say hello to V4.1 This model is built for images that captivate you. Photorealism is more human, gradients are dreamier, and new illustration styles are now possible. Test it out in Recraft Studio today and see what you can create.
37
47
24
548
3,340,112
Roman Suvorov retweeted
Clean vector graphics straight from the prompt. Yes — real SVG. Export and refine in Figma or Illustrator. Recraft V4 Vector doesn’t just generate images. It generates usable design assets.
11
10
277
125,433
Roman Suvorov retweeted
V4 Vector is built for real design workflows. Clean path structure. SVG export. Print-ready (300 DPI, CMYK). Generate → export → refine
1
3
1
82
450,217
Roman Suvorov retweeted
Generate fully scalable graphics with Recraft V4. Posters, characters, icons, illustrations... whatever you need. → clean structure. → crisp lines. → production-ready. Export as SVG or PDF for print. Or open it in Figma / Illustrator and tweak it however you want.
3
6
2
84
3,661
Roman Suvorov retweeted
Recraft V4 is low-key insane for logos. Have you tested it yet? One prompt → multiple usable directions.
6
14
211
475,446
Roman Suvorov retweeted
When aesthetics are built in, you don’t need a 200-word prompt. “close-up of an Asian model, orange background” V4 understands lighting, color harmony, and composition by default.
4
4
1
78
3,289
Roman Suvorov retweeted
We're excited to announce Recraft V4, our latest image model developed with designers to bring true visual taste to AI generation 🔥 Find out more about the model: go.recraft.ai/model-launch
23
40
17
326
50,935
Roman Suvorov retweeted
Today our founder and CEO @dorogushanna is giving a talk at the AI User Conference in SF. Tune in for: → Advancements in AI image generation technology → Features we recently launched at Recraft → Our roadmap for 2024 Watch on @aiusergroup's Linkedin Page at 2:50pm PT.
2
1
6
844
We've been working on the new model for several months and this is huge leap forward. I'm excited about Recraft 20B has became available and looking forward to receiving any feedback.
Today we launched Recraft 20B, our new foundation model on @ProductHunt. 🎉 We now offer our users visual style control, improved anatomy, enhanced prompt quality, and more! Share your thoughts with us on Product Hunt. producthunt.com/posts/recraf…
3
162
Just wow! Wanna try!
This Post is from an account that no longer exists. Learn more
113
👷🏻‍♂️ We are working hard to help designers solve their tasks more efficiently with Recraft. Some people already find it useful!
Thank you, 50,000 beloved users, for generating 1,500,000 illustrations, icons, avatars, logos, and other beautiful designs. Love, the Recraft team ♥
1
2
145
Roman Suvorov retweeted
Recraft AI: all-in-one suite to 💫 make you vector graphics or ✏️ edit your own assets. On Product Hunt today! 🚀Feature rundown in the thread:
1
3
2
9
2,557
Today, we are launching @recraft_ai on Product Hunt! We have big updates, style presets, new website, and many other features! We would appreciate your support!
3
2
17
873
Generative AI is getting more and more friendly for non-tech people. Happy to help building that bridge!
This quoted post is unavailable.
1
83
🤯
📢📢 Multi-shot view synthesis from our group. We extend our previous StereoLayers for an arbitrary number of input images with blazing fast inference and high quality. Look at SIMPLI interactive demo samsunglabs.github.io/MLI/ Seconds for a new scene and rendering in a browser💥
LaMa is on the cover 😮
Awesome AI R&D content on Computer Vision News of January 2022. rsipvision.com/ComputerVisio… Subscribe for free on page 70. Enjoy the reading! #AI #ComputerVision #artificialintelligence #deeplearning #MedicalImaging @RSIPVision
3
14
Roman Suvorov retweeted
Stereo Magnification with Multi-Layer Images abs: arxiv.org/abs/2201.05023 project page: samsunglabs.github.io/Stereo…
7
26
So excited about being a part of @ykilcher's video! That was a really interesting experience to me. BTW, which other tasks should we apply LaMa-Fourier architecure to?
🔥Paper Explanation & Author Interview🔥 LaMa uses Fourier Convolutions for large mask in-painting. The results here are really crazy🤯LaMa can easily fill very large holes with complex background patterns. You can even try it yourself! Watch here: youtu.be/Lg97gWXsiQ4
1
7
2
45