Clint

I just tested networkit's implementation against both igraph and networkx using floating point edge weights and got the exact same paths.

I have already expanded the code in the Trainer class for multi GPU training as well as the text introducing the reader to distributed training so it shouldn't take much...

Thank you will do

Hi Garrett, The instructions in the blog will be updated shortly. In the meantime, the recommended installation for bitsandbytes for ROCm is as follows: ``` git clone --recurse https://github.com/ROCm/bitsandbytes cd...

Support is still missing for float8_dynamic_activation_float8_weight and float8_static_activation_float8_weight on ROCm which both throw the below error: File "/usr/local/lib/python3.12/dist-packages/torchao/float8/inference.py", line 90, in addmm_float8_unwrapped_inference output = torch._scaled_mm( RuntimeError: false INTERNAL ASSERT FAILED...