mstar.streaming.topology#
Classes
|
Defines a streaming connection between two partitions. |
|
Declares how a model's computation is split into async partitions. |
|
A graph edge that carries streaming data between partitions. |
- class mstar.streaming.topology.Connection(from_partition, to_partition, edge_name, chunk_policy_factory)[source]#
Bases:
objectDefines a streaming connection between two partitions.
- Parameters:
from_partition (str)
to_partition (str)
edge_name (str)
chunk_policy_factory (Callable[[], ChunkPolicy])
- chunk_policy_factory: Callable[[], ChunkPolicy]#
- class mstar.streaming.topology.PartitionTopology(partitions, connections=<factory>)[source]#
Bases:
objectDeclares how a model’s computation is split into async partitions.
Each partition has its own set of graph walks. Connections define streaming data flow between partitions via StreamBuffers.
- Parameters:
connections (list[Connection])
- connections: list[Connection]#
- class mstar.streaming.topology.StreamingGraphEdge(next_node, name, tensor_info=<factory>, persist=False, conductor_new_token=False, is_streaming=False, output_modality='', _persist_for_loop=False, _final_stream_chunk=False, _total_fanin=1, _shard_dim=None, target_partition='')[source]#
Bases:
GraphEdgeA graph edge that carries streaming data between partitions.
Routed like a normal GraphEdge (producer is unaware it’s streaming). On the consumer worker, the arriving tensors are buffered in a StreamBuffer and gated by a ChunkPolicy before satisfying the consuming node’s input.
- Parameters: