Optimize Mixtral with expert parallelism (#2090)
This commit is contained in:
@@ -39,13 +39,15 @@ _MODELS = {
|
||||
}
|
||||
|
||||
# Models not supported by ROCm.
|
||||
_ROCM_UNSUPPORTED_MODELS = ["MixtralForCausalLM"]
|
||||
_ROCM_UNSUPPORTED_MODELS = []
|
||||
|
||||
# Models partially supported by ROCm.
|
||||
# Architecture -> Reason.
|
||||
_ROCM_PARTIALLY_SUPPORTED_MODELS = {
|
||||
"MistralForCausalLM":
|
||||
"Sliding window attention is not yet supported in ROCm's flash attention",
|
||||
"MixtralForCausalLM":
|
||||
"Sliding window attention is not yet supported in ROCm's flash attention",
|
||||
}
|
||||
|
||||
|
||||
|
||||
Reference in New Issue
Block a user