Ashwin Murali comments

Results 16 comments of


                                            Ashwin Murali

Webdataset support for Tensorflow

Thanks for your prompt reply. Sure, I will debug the issue shortly and leave an update. "(Separately, we're working on native Tensorflow support for WebDataset.)" This is great news! Do...

LoadMapAdnLocalize crashes

Hi @JCastel358 , Did you check if the path specified in the "map_url" parameter exists? Best,

Export to ONNX

Hi @mlzxy, I have the same issue as well. I am not able to export the model to onnx. Please help if anyone has tried something similar. Thanks

Long inference time

I also have the same problem. Has anyone found a fix? Thanks

Inference speed issue

Hi @YanxingLiu, Were you able to figure out the reason and improve the inference speed? Your answer would greatly help. Many Thanks,

Training on a dataset with one class.

I think, you have to include categories from COCO, PASCAL or LVIS datasets as base classes. So, if you have 1 novel class, you have to include 3 categories from...

calibration of depth value

I also have a similar issue. I have a real depth image that has sparse depth values. Is it possible to rectify the raw depth using the ZoeDepth's predicted depth...

Minimum dataset size of Novel Objects

Thanks for your reply @tgjantos , If I have a backbone that is able to detect my specific objects well, then few images per object would be enough to train...

Details of Hardware used for "Real Robot Experiment"

Thanks @mlzxy for your reply, I tried the vit-l open vocabulary model on my custom dataset and it produced good results with around 0.8FPS in Jetson Orin device. I'm looking...

How to produce mask for new image, in few-shot inference?

Hi @nguyenthekhoig7, I generated my masks using Segment Anything Model. You can also use SimpleClick. Best,