[QUESTION] Does Megatron-Core supports LLAMA models?
Opened this issue · 5 comments
noob-ctrl commented
Does Megatron-Core supports LLAMA models?
ethanhe42 commented
yes
noob-ctrl commented
@ethanhe42 When transformer-impl
is local
, it reports the following error:
AssertionError: (RMSNorm) is not supported in FusedLayerNorm when instantiating FusedLayerNorm when instantiating TransformerLayer
When transformer-impl
is transformer_engine
, the following code does not seem to define RMSNorm?
So do I need to make any changes when I want to use llama?
ethanhe42 commented
You need to use mcore models. local is deprecating
noob-ctrl commented
@ethanhe42 When transformer-impl
is set to transformer_engine
, the following code does not seem to define RMSNorm?
ethanhe42 commented
It's handled by TEnorm