GPT | ChatGPT | Midjourney — News
Канал "GPT | ChatGPT | Midjourney — News" - ваш источник последних новостей из мира искусственного интеллекта, технологий и науки. Мы предлагаем увлекательные статьи, глубокие аналитические материалы и самые свежие обновления в области AI, технологий и научных открытий. Присоединяйтесь, чтобы быть в курсе всех событий и тенденций в этой захватывающей области!
A powerful book writing tool has been assembled on the basis of Claude
It is enough to give the neural network a general idea and briefly describe the events that you want, and then artificial intelligence will do everything.
Thanks to the elaborated model, the texts come out alive and interesting, almost as good as human ones.
You can test it for free in Google Colab
Free image generator.
You make one request, and you get three images at once in the best Stable Diffusion models — SDXL Turbo, Lightning and Hyper.
Try it here.
🆕 Great news for programmers: five of the most powerful neurons have been embedded directly into VSCode.
They write, fix and explain all the code for free, and if you don't like the result, you can immediately choose another AI.
It connects in five minutes:
1. Create an account in NVIDIA AI here.
2. Click on any model and click "Get API key".
3. Download the CodeGPT extension in VSCode.
4. In it, select the NVIDIA provider and paste the copied API key.
5. Everything is ready — Llama-3, Mixtral, Gemma, Phi-3 and others are available in the development environment.
🤩 The best free neural network for voice cloning has been updated
In the new OpenVoice V2, the sound quality has been greatly improved and the number of supported voices has been expanded.
Free demo here, can be installed from GitHub .
✊ Upscale is the best quality improvement app, it has been updated - now it allows you to improve the quality of photos by whole folders.
The neural network removes artifacts in the photo, makes small elements more expressive and increases its size by 4 times — while the text does not suffer with the latest update.
The coolest thing is that you can use it for free on your PC or laptop.
Now any sketches can be turned into architectural projects.
🏆 The guy created an interface with a neural network that converts ordinary drawings into architectural-themed designs.
The installation is simple (and free), the instructions are here.
🔥Vidu — The Chinese have created their own Sora!
We have trained a new competitor of open ai, at the Shangshu Technology startup and Tsinghua University.
Video is generated up to 16 seconds in 1080p. If the demo video is not fake, then the quality of Vidu is already one step away from the quality of Sora generation.
The architecture of the model is kind of based on Vt, which is similar to the Diffusion Transformer that Sora has.
If you're wondering what kind of engine is behind this, then here's the Google scholar profile the researcher who led this project.
We are waiting for more details about this model!
Checkpoint in Chinese
You can apply for access here
😱The Chinese have started posting dozens of dipfakes online to scam men out of money.
Going online is getting scarier and scarier.
The Champ neural network animates portraits
It is enough to upload a photo and add an example of movements.
You can try it for free here: https://replicate.com/camenduru/champ
A site has been found where you can track neural network statistics, a huge number of different functions to determine the best neural networks for your tasks. https://artificialanalysis.ai/
A man put on a T-shirt with a "Stop" sign to stop an unmanned taxi
It turned out that robotic machines are very vulnerable to this kind of tricks. The car stops 99% of the time.
The uprising of the machines is canceled.
💡 A free chatbot from #Nvidia, functioning locally on a computer, without a network connection. Just a dream!
In ChatRTX, you can select a folder with files on your PC as a dataset and ask questions about it, or select another model inside.
It works without the Internet. To work, you need an RTX 30 or 40 with a volume of 16 GB or more.
🦾 Adobe has released the VideoGigaGAN neural network, which increases video quality by eight times
Even videos with a very low resolution of 128x128, it is able to convert up to 1024x1024.
What it looks like to synchronize lip movement with sound using the FlawlessAI neural network.
The tool will allow you to use more censorious language to lower the age rating of a movie, as well as do lip-syncing for other languages.
The robot with the built-in AI grew tired and interrupted the human who had been talking for a long time.
Even the neural networks were beginning to tire of humans.
A music video created with AI has been released
AI Filmmaker made a music video for the band Pain using neural networks. The author used Midjourney for generation, then processed the images in Photoshop, After Effects and Premiere Pro. The dubbing was overlaid with Elevenlabs
Video is too big
Nvidia will provide software and microchips for Japan's national ABCI-Q hybrid quantum supercomputer initiative.
Scientists from China have trained AI to identify the source of cancer by metastasis cells
Tianjin University specialists trained an AI model on 30 thousand images of metastasis cells found in the abdominal cavity or lung fluid of 21 thousand people whose tumor origin was known. The model was then tested on 27,000 images of metastasis cell samples. Its efficiency in recognizing the source of cancer was 83%.
The scientists also tested a group of more than 390 patients who were treated four years ago according to the AI's prognosis and doctors' predictions. The results showed that matching the AI prognosis increased patient survival.
🔗 Source
Magnific AI is done. Coders have dropped the most powerful AI tool for 16x (!) photo enhancement.
Clarity Upscaler neural network increases the detail of each pixel individually. Now the most soapy frame will turn into a studio photo.
This is the kind of thing we save - link.
Похожие
