Edgar Medina comments

Results 18 comments of


                                            Edgar Medina

problem reading JSON

Problem solved. Files must be downloaded by yourself, this error is generated by corrupted or uncompleted files.

light_head_rcnn/lib/lib_kernel/lib_fast_nms/fast_nms.so: undefined symbol:

I had the same error. I didn't solve it, but i think it's a compilation problem.

Replace BatchNorm with WeightNorm

it's true, and according to [Improved Training of Wasserstein GANs](https://arxiv.org/pdf/1704.00028.pdf) gradient penalty can lead a better performance than weight clipping or even weightNorm. did you read this paper, I would...

where is xception like network code which is written in original paper?

Hi, can someone provide only some insights about the Xception-like model? I think I can implement the training and some other stuff, but I'm not sure if I understood well...

where is xception like network code which is written in original paper?

how many layers does xception-like backbone in Light-Head R-CNN has? Are they 17 conv layers (based on Table 7) and 2 MLP (from original code)? edited: @geonseoks I computed the...

Confused about relation_prior.

Hi unfortunately no. I ran out to another project with similar performance.

Confused about relation_prior.

well, because this is not state-of-the-art nowadays, I just looked on google :D I found one from 2022 in the papers cited this one. But I dont remember at the...

split in train and test sets

Oh thanks for that. I will need it anyway. But sorry, I didn't specify the dataset because I only saw CMU -motion in this git. I was wondering about this...

split in train and test sets

Hi. Do you have a reference for that split? it is really interesting to know. Thanks in advanced.

dimension_to_use for h36m

maybe because modelling complexity and movement correlation between joints :)