Model: Add tensor_parallel_backend option

This allows for users to use nccl or native depending on the GPU setup.
NCCL is only available with Linux built wheels.

Signed-off-by: kingbri <8082010+kingbri1@users.noreply.github.com>

endpoints/core/types/model.py

@@ -90,6 +90,7 @@ class ModelLoadRequest(BaseModel):
         examples=[4096],
     )
     tensor_parallel: Optional[bool] = None
+    tensor_parallel_backend: Optional[str] = "native"
     gpu_split_auto: Optional[bool] = None
     autosplit_reserve: Optional[List[float]] = None
     gpu_split: Optional[List[float]] = Field(