[Core][Doc] Default to multiprocessing for single-node distributed case (#5230)
Co-authored-by: Antoni Baum <antoni.baum@protonmail.com>
This commit is contained in:
@@ -19,10 +19,6 @@ class MultiprocessingGPUExecutor(DistributedGPUExecutor):
|
||||
"""Python multiprocessing-based multi-GPU executor"""
|
||||
|
||||
def _init_executor(self) -> None:
|
||||
assert (
|
||||
not self.speculative_config
|
||||
), "Speculative decoding not yet supported for MultiProcGPU backend."
|
||||
|
||||
# Create the parallel GPU workers.
|
||||
world_size = self.parallel_config.tensor_parallel_size
|
||||
|
||||
|
||||
@@ -65,10 +65,11 @@ def _set_future_result(future: Union[ResultFuture, asyncio.Future],
|
||||
future.set_result(result)
|
||||
return
|
||||
loop = future.get_loop()
|
||||
if result.exception is not None:
|
||||
loop.call_soon_threadsafe(future.set_exception, result.exception)
|
||||
else:
|
||||
loop.call_soon_threadsafe(future.set_result, result.value)
|
||||
if not loop.is_closed():
|
||||
if result.exception is not None:
|
||||
loop.call_soon_threadsafe(future.set_exception, result.exception)
|
||||
else:
|
||||
loop.call_soon_threadsafe(future.set_result, result.value)
|
||||
|
||||
|
||||
class ResultHandler(threading.Thread):
|
||||
|
||||
Reference in New Issue
Block a user