search

ivipsourcecode / dxslam

Other
418 stars 75 forks source link

Get features from a GPU-trained hf-net #28

Closed Gpetrak closed 2 years ago

Gpetrak commented 2 years ago

First of all, thank you for your great work. I want to use DXSLAM to get features from a re-trained hf-net in GPU. I compare the two graphs (yours optimized in OpenVino and mine trained in GPU) in tensorboard and I observed that the two graphs have a significant difference. The optimized graph has two extra parts: simple_nms and top_k_keypoints which include the two tensors used in GetFeature.py script: pred/simple_nms/radius:0 pred/top_k_keypoints/k:0 where they don't exist in the re-trained hfnet. Is this difference a modification which was made for some reason, or it is the optimized version by OpenVino ?

How could I get features from a re-trained in GPU Hf-net ?

Gpetrak commented 2 years ago

After more research in the issue, I realized that these tensors don't exist in my re-trained model of Hf-net. The initial vanilla Hfnet model includes these tensors. Thus, it's something wrong with the re-trained model deployment. Because it is not an issue of DXSLAM I close this issue.