When it comes to AI, many of us prefer open source models. They are not always competitive with commercial models but the HunyuanVideo I2V model comes close. It is an image to video generation model that generates realistic videos from your images. You can always download and run it locally. For a 720p resolution video, you are going to need an NVIDIA GPU with CUDA. You are going to need at least 60GB GPU memory for 720p videos.
Most of us can’t afford such a GPU. The good news is Tensor.art is already offering this model. You simply have to upload your image, enter your prompt, and the rest is easy. You get daily credits on the site and can buy 3000 credits for under $10. My generation cost 34.02 credits, so it wasn’t too bad.
[HT]
