pyun-ram

[question] Problems when running batchensemble on TPU

1

Hi, Thanks for sharing the awesome codebase! I am trying to run batchensemble on TPU with CoLab, but did not make it. When I run the following commands in CoLab,...

Cannot get the similar result as the paper reported

6

The mAP on VOC2007 is as follows (19 classes and 19+1 classes): | Category | mAP (40k, 1000000 img) | |---------------+--------------------------| | aeroplane | 3.9 | | bicycle | 9.1...

[ADD] K-FAC for BatchNormNd (#259)

2

Extension to support BatchNormNd (eval) K-FAC Resolves f-dangel/backpack/issues/259 Auxiliary: - The kfac quantity contains only one element, and represents the GGN approximation. - It only supports the evaluation mode. -...

KFAC support in BatchNorm (eval mode)

16

Hi, Thanks for the repo! This is really a nice work. I am planning to calculate the KFAC with backpack. But it raises the following error: ``` NotImplementedError: Extension saving...

Should this line be 'yz_distances[:, iz] = np.linalg.norm(yc - z_embeded, 2, axis=1)' instead? If it is 'yz-z_embed', then it is not a GMM but a uni-modal Gaussian. https://github.com/ZiningWang/Inferring-Spatial-Uncertainty-in-Object-Detection/blob/3fc7c8d78e538805d928ba084c1c4570f795babc/utils/metric_utils.py#L97

Add DockerfileCUDA102

1

## Description of the change > Description here ## Type Of Change - [ ] Bug Fix (non-breaking change that fixes an issue) - [x] New Feature - [ ]...

[feature] Dockerfile to support CUDA-version pytorch-cpp

2

Thanks for the nice code! Here is a Dockerfile to support CUDA-version pytorch-cpp. Hope it helps when you want to run the code with GPUs. ```Dockerfile FROM nvidia/cuda:10.2-cudnn7-devel-ubuntu18.04 LABEL maintainer="[email protected]"...