Artur Niederfahrenhorst

Tried to follow the steps provided in https://doc.fedml.ai/deploy/quickstart. Steps to reproduce: > conda create -c conda-forge python=3.10 -n fedml > conda activate fedml > pip install fedml > fedml model...

Since Aviary caches weights based on the model id, changing the S3 path for a given model with a given model-id that has been run before does not do anything....

LongLoRA + Flash Attention 2 causing illigal memory access

7

Thanks for providing the LongLoRA forward functions. Your flash-attn/non-flash-attn implementations of SSN show divergent behavior in my case. For a repro script, please have a look at the issue I...