Home DEVELOPER AI videos and Pictures for all: AMuse 3.0.1 Ai -Slight behavior

AI videos and Pictures for all: AMuse 3.0.1 Ai -Slight behavior

0


The free Windows program is released in AMuse Edition 3.0.1. It provides easy access to the image generation with artificial intelligence – on the local computer, without internet access.

The software is sponsored by AMD and therefore Rvenne with AI co-processor has some adaptations for CPU, but also runs on NVidia graphics card. If the accelerator hardware is missing, amuse also enables to rendate on the CPU; Of course, it takes a long time.

Software is straight on Recreational project website ready for download. Recently released version 3.0.1 old versions and AI models were uninstalled, as they are no longer compatible. Some new stable spread models are available for this. The rider was earlier created to make the video and was used for the amendment of the existing video, but now eight new models have been added to support “text to video” with “locomotion”, ie enable the video clip using the text prompt.

A very young branch of generic AI now falls into the free area. Various side conditions are predetermined, but various proposals including “headquarters” (for high quality) can be set. By default, amuse makes 2 second-second video tape with 24 images, but can also be configured here.

In the first trials, the 640 × 384 pixel requires the shape of a twice on a two-time clip, with 64 GB RAM, with 64 GB RAM, Rven 7640HS Pro and for about two and a half minutes rapidly with PCIE Gen4-SSD for about two and a half minutes. Long -term clips and high resolutions require more time. It takes a lot of tact to develop the desired results. Some models also have problems with some representations – the example above in the picture instead of a dog view, they then produced the facial of Asian people. Movements also often look unnatural.

IX Workshop: Advanced Administration of Kuberanets

AMuse invites you to try your own AI images and videos with various models. Other tasks include images with AI or “feature extraction”, from which new image materials can also be made using artificial intelligence.

AMuse to version 3.0.1 provides more models for downloads, including many that are adapted to the AMD processor. Downloading the model is better: individual selected models finish in a download queue, the download can be continued after demolition. If the models are on the computer, amuse no longer needs internet connection. With the materials arising out – the prompt is therefore on your own computers and not on foreign servers – so AI can be used under the guarantee of privacy.


(DMK)

IX Workshop: Advanced Administration of Kuberanets

NO COMMENTS

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Exit mobile version