MAGAer13

You can initalize the model on cpu with float32. After the initialization, convert the model into half then put it on GPU.

Sure.

Hi, for the SEEDBench and Q-Bench, we treat this as the Multiple Choice Question for open-generation. You can refer to the MMBench (which is also a MC benchmark) for reference,...

Hi, all! Just quick notice about our latest version of mPLUG-Owl which remove the dependency of apex! And it is more easy to use

Scheduled in next month.

We will release mPLUG-Owl-2 recently, which is the next generation of mPLUG-Owl. As well as the pre-trained weight / paper.

> > We will release mPLUG-Owl-2 recently, which is the next generation of mPLUG-Owl. As well as the pre-trained weight / paper. > > Is the "mplug-2" you recently released...

Hi, the model on huggingface space is the advanced version of mPLUG-Owl which natively support video with temporal related module as input without treating video as multiple frames. The video...

Yes, specially we add the local temporal modeling module proposed mPLUG-2, and add a trajectory learnable queries in visual abstractor module. We will include this in the new version of...

> really need it ! We will release the video version in this week!