Wenxuan Cao

how can I use mmsegmentation to optimize the model？

1

I‘m new to mmsegmentation and I need to optimize the model, so I wonder how to implement Pruning, Quantization and Knowledge Distillation in mmsegmentation. I didn't find them in document....

增加了HGA模型和训练

[distGB] fix local variable 'sorted_idx' referenced before assignment in convert_partition.py

3

The code will crashed in convert_partition.py::_process_partition_gb when sorted_etypes is False. In this pr, I add the code when condition is False.

DistDGL v2.4 Training Error when num_samplers>0

4

## 🐛 Bug When we run dgl/tools/launch.py, it will return Failures if num_samplers = 1 and return _frozen_importlib._DeadlockError when num_samplers > 1. ## To Reproduce Steps to reproduce the behavior:...

[DistGB] save as graphbolt graph directly after partition

31

Simplify the graphbolt graph generation process, now we no longer need to save the DGLGraph to generate the FusedCSCSamplingGraph in partition process