Re:Sora: Creating video from text
by llm_trw on Thursday February 15, 2024 @09:42PM
(#39392498)
>We are giving the enjoyable parts of life to a computer. And we are left with the drudgery.
Yesterday I asked a local llm to write a python script to have a several multimodal llms rank 50,000 images generated by a stable diffusion model. I then used those images to train a new checkpoint for the model and can now repeat the process ad infinitum.
In the olden days of 2020 I would have had to hire 5000 people each working for a day to do the same.
Yesterday I asked a local llm to write a python script to have a several multimodal llms rank 50,000 images generated by a stable diffusion model. I then used those images to train a new checkpoint for the model and can now repeat the process ad infinitum.
In the olden days of 2020 I would have had to hire 5000 people each working for a day to do the same.