github/ColossalAI
1
0
Fork 0
mirror of https://github.com/hpcaitech/ColossalAI.git synced 2025-09-04 18:40:28 +00:00
Code Issues Packages Projects Releases Wiki Activity

[misc] refactor launch API and tensor constructor (#5666)

Browse Source 
* [misc] remove config arg from initialize

* [misc] remove old tensor contrusctor

* [plugin] add npu support for ddp

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* [devops] fix doc test ci

* [test] fix test launch

* [doc] update launch doc

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
This commit is contained in:
Hongxin Liu
2024-04-29 10:40:11 +08:00
committed by GitHub
parent 91fa553775
commit 7f8b16635b
223 changed files with 294 additions and 403 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
tests/test_tensor/test_comm_spec_apply.py
View File

@@ -178,7 +178,7 @@ def check_all_reduce_in_flatten_device_mesh(device_mesh, rank):
def check_comm(rank, world_size, port):
disable_existing_loggers()
launch(config={}, rank=rank, world_size=world_size, host="localhost", port=port, backend="nccl")
launch(rank=rank, world_size=world_size, host="localhost", port=port, backend="nccl")
physical_mesh_id = torch.arange(0, 4)
assert rank == dist.get_rank()

2
tests/test_tensor/test_dtensor/test_comm_spec.py
View File

@@ -124,7 +124,7 @@ def check_all_reduce_bwd(process_groups_dict, rank):
def check_comm(rank, world_size, port):
disable_existing_loggers()
launch(config={}, rank=rank, world_size=world_size, host="localhost", port=port, backend="nccl")
launch(rank=rank, world_size=world_size, host="localhost", port=port, backend="nccl")
physical_mesh_id = torch.arange(0, 4)
assert rank == dist.get_rank()

2
tests/test_tensor/test_dtensor/test_dtensor.py
View File

@@ -21,7 +21,7 @@ class TestModel(torch.nn.Module):
def check_dtensor(rank, world_size, port):
disable_existing_loggers()
launch(config={}, rank=rank, world_size=world_size, host="localhost", port=port, backend="nccl")
launch(rank=rank, world_size=world_size, host="localhost", port=port, backend="nccl")
test_model = TestModel(8, 8).to("cuda")
original_tensor = torch.rand(4, 8).to("cuda")
compare_output = test_model(original_tensor)

6
tests/test_tensor/test_dtensor/test_layout_converter.py
View File

@@ -20,7 +20,7 @@ mesh_shape = (2, 2)
def check_one_step_transform(rank, world_size, port):
disable_existing_loggers()
launch(config={}, rank=rank, world_size=world_size, host="localhost", port=port, backend="nccl")
launch(rank=rank, world_size=world_size, host="localhost", port=port, backend="nccl")
# [[0, 1],
# [2, 3]]
device_mesh = DeviceMesh(physical_mesh_id, mesh_shape, init_process_group=True)
@@ -82,7 +82,7 @@ def check_one_step_transform(rank, world_size, port):
def check_layout_converting(rank, world_size, port):
disable_existing_loggers()
launch(config={}, rank=rank, world_size=world_size, host="localhost", port=port, backend="nccl")
launch(rank=rank, world_size=world_size, host="localhost", port=port, backend="nccl")
dim_partition_source = {1: [0, 1]}
dim_partition_target = {0: [0, 1]}
device_mesh = DeviceMesh(physical_mesh_id, mesh_shape, init_process_group=True)
@@ -141,7 +141,7 @@ def check_layout_converting(rank, world_size, port):
def check_layout_converting_apply(rank, world_size, port):
disable_existing_loggers()
launch(config={}, rank=rank, world_size=world_size, host="localhost", port=port, backend="nccl")
launch(rank=rank, world_size=world_size, host="localhost", port=port, backend="nccl")
dim_partition_source = {1: [0, 1]}
dim_partition_target = {0: [0, 1]}

2
tests/test_tensor/test_mix_gather.py
View File

@@ -296,7 +296,7 @@ def check_two_all_gather_RS01(device_mesh, rank):
def check_comm(rank, world_size, port):
disable_existing_loggers()
launch(config={}, rank=rank, world_size=world_size, host="localhost", port=port, backend="nccl")
launch(rank=rank, world_size=world_size, host="localhost", port=port, backend="nccl")
physical_mesh_id = torch.arange(0, 8)
assert rank == dist.get_rank()

2
tests/test_tensor/test_padded_tensor.py
View File

@@ -10,7 +10,7 @@ from colossalai.testing import rerun_if_address_is_in_use, spawn
def check_padded_tensor(rank, world_size, port):
disable_existing_loggers()
launch(config={}, rank=rank, world_size=world_size, host="localhost", port=port, backend="nccl")
launch(rank=rank, world_size=world_size, host="localhost", port=port, backend="nccl")
original_tensor = torch.rand(32, 64).to("cuda")
device_mesh = DeviceMesh(torch.Tensor([0, 1, 2, 3]), (2, 2), init_process_group=True)

2
tests/test_tensor/test_shape_consistency_apply.py
View File

@@ -11,7 +11,7 @@ from colossalai.testing import rerun_if_address_is_in_use, spawn
def check_apply(rank, world_size, port):
disable_existing_loggers()
launch(config={}, rank=rank, world_size=world_size, host="localhost", port=port, backend="nccl")
launch(rank=rank, world_size=world_size, host="localhost", port=port, backend="nccl")
physical_mesh_id = torch.arange(0, 4)
mesh_shape = (2, 2)