Vikram Voleti

icon indicating a website link voletiv.github.io icon indicating an email [email protected]

icon company @mila-udem icon location Montreal, Canada icon location PhD candidate at Mila, University of Montreal

Results 26 comments of Vikram Voleti

Added TimeDistributed option

All tests passed, except on Python 2.7 with Theano - because 'mkl' could not be imported: > RuntimeError: > Could not import 'mkl'. Either install mkl-service with conda or set...

\implies does not work

I got the same, I used \Rightarrow

First page refresh causes wrong image positions

@bartosz-malec I'm facing a similar issue where the widths are 100% in all cases, could you please tell us what your JustifiedGalleryCode is?

Zero123 / Magic123 bad results on human 2D-to-3D task, on Yoga-82 dataset

Many variables in the config (LR, etc.) affect the final results quite a bit, both in Zero123 and Magic123

Proposal: using multiple guidances in training

We have a version of this implemented in zero123, because we take guidance from ref + zero123. I expanded that to take guidance from ref + SD/DF + zero123. You...

What are the inputs to the lip & audio models?

Here's the Syncnet architecture: ![screenshot from 2018-04-09 22-26-59](https://user-images.githubusercontent.com/22424247/38511254-2264ddc2-3c17-11e8-8b05-0f6e3deb8cfa.png) In my repo, I load the two branches separately - audio model, and lips model. As can be seen in the lips...

What are the inputs to the lip & audio models?

@edwinter @michiyosony I haven't used Syncnet for audio-video sync, I used it to get good video features for lipreading and other tasks. I'm not sure about the reflection part, it...

Adds guidance_val to zero123

I'm making a few more changes, please don't merge yet

question about camera matrixs for spatial coordinate conversion

In case it helps:

Zero123: Vary min_step_percent and max_step_percent for noise levels across iters

@claforte Done!