Juan A. Rodriguez

icon indicating a website link joanrod.github.io

icon company ServiceNow Research icon location Montreal icon location PhD student at Mila, ÉTS and ServiceNow Research in Montreal, Canada. I work on AI and deep learning projects

Results 3 issues of Juan A. Rodriguez

Processing in batch and GPU

1 comment

Hello, Is there any functionality in the library to process the pdfs in batches and/or using GPU to accelerate computing? If not, what would be the go to steps to...

Embedding dimensions of LDM-VQ models are different wrt to VQGAN's

I realize that the configuration of VQ autoencoders in Latent Diffusion is different than the one used in VQGAN (taming-transformers). Specifically, I see that **embed_dim** and **z_channels** have low values...

Is beta schedule correctly defined?

Hi all, Am I wrong or the beta_schedule is not correct? https://github.com/Stability-AI/stablediffusion/blob/47b6b607fdd31875c9279cd2f4f16b92e4ea958e/ldm/modules/diffusionmodules/util.py#L24 It looks like a quadratic (pow 2) rather than linear. ``` linear_start = 0 linear_end = 1 betas...

question