The development of an AI system industry for creating artistic scenarios based on abstract descriptions of acquired tempos that affect even their creators. New applications date back to several months, demonstrating an ever higher quality of r. As an example, the difficult fate of the Image Video project will do.
Google has unveiled the first work of a new AI “Imagen Video” that has developed HD video synthesis at 1280×768 resolution at 24 frames per second based on simple text descriptions. This theme builds full-fledged 3D models of production facilities with the creation of complex t t t t Oa meets iitirov sets For example, the flight of a probe through the battlefields of space pirates with unknown aliens.
Action Action Image Video Build Xs on the board with 3 levels, where is the most simple mode for the first action? Then, depending on the task of post-processing the ball, the picture succeeds step by step, produces the necessary effects, and the AI was trained on the LION-400M database, in which there are about 14 million for the “operation + video” combination and 60 million pairs of “text +picture”, but this time its main vulnerability.
In order to fully cover the accumulated cultural base, LAION-400M made an open database, so the share of accumulation in it increased, which, according to censored reserves, costs no more. Alas, it is also impossible to treat, otherwise AI loses the ability to adequately recognize requests. For example, to show compassion, you need to be able to understand what pain and suffering are. Google has not yet been able to find a solution for this problem, so they have postponed the publication of the Imagen Video code and the opening of access to this AI indefinitely.
Source: Tech Cult
