PipeNetwork/minimax-h3-mlx MiniMax released MiniMax-H3 two days ago - they describe it as a "a general-purpose, omni-modal generative system", which in practice means it accepts text, images, audio and video and can use them to generate up to 15 second video clips with audio included. This Python package ports it to MLX for running on Apple Silicon. I got it running on my M5 Max MacBook Pro. I cloned the repo and ran the model like this: # First download the models uvx --from huggingface_hub ✦ AI generated
The MiniMax-H3-MLX package enables running the model on Apple Silicon Macs using the MLX framework.
article author: This Python package ports it to MLX for running on Apple Silicon.
provides context · 2
MiniMax recently released H3, an omni-modal model that takes text, images, audio, or video as input and generates video clips with audio.
article author: MiniMax released MiniMax-H3 two days ago - they describe it as a "a general-purpose, omni-modal generative system", which in practice means it accepts text, images, audio and video and can use them to generate up to 15 second video clips with audio included.
The official prompting guide contains instructions for generating proper audio, which the author hadn't read before running the experiment.
article author: The prompting guide (which I didn't read prior to this experiment) has a whole bunch of information on how to get this to work.
provides context · 2
The model is resource-intensive, requiring 115 GB of downloads and taking about 45 minutes to generate a single video clip on high-end Apple hardware.
article author: It downloaded ~115 GB of model files, and the video generation took just under 45 minutes.
provides context · 2
The generated video was visually impressive, but the audio was unintelligible garbage due to the author not following the audio prompting guide.
article author: The video is impressive, but the audio is weird speech-like garbage, because I didn't provide any prompt guidance as to what the audio should be.
provides context · 2