Codys12

maxBatchSize() parameter for Governance and Governance settings

2

**🧐 Motivation** Having a maxBatchSize parameter that is checked by the Governor contract would limit proposals to a specific number of operations. **📝 Details** This parameter could be defined in...

Position Embedding with Seq > 512

1

I am trying to run Llama-3.1-8B with a seq > 512, and I get this error. Do I have to manually set position embeddings to get this to work? ```...

Is it possible to initialize a model in data parallel and stream one layer to multiple GPUs for high batch sizes?