[gemini] improve compatibility and add static placement policy (#4479)

* [gemini] remove distributed-related part from colotensor (#4379)

* [gemini] remove process group dependency

* [gemini] remove tp part from colo tensor

* [gemini] patch inplace op

* [gemini] fix param op hook and update tests

* [test] remove useless tests

* [test] remove useless tests

* [misc] fix requirements

* [test] fix model zoo

* [test] fix model zoo

* [test] fix model zoo

* [test] fix model zoo

* [test] fix model zoo

* [misc] update requirements

* [gemini] refactor gemini optimizer and gemini ddp (#4398)

* [gemini] update optimizer interface

* [gemini] renaming gemini optimizer

* [gemini] refactor gemini ddp class

* [example] update gemini related example

* [example] update gemini related example

* [plugin] fix gemini plugin args

* [test] update gemini ckpt tests

* [gemini] fix checkpoint io

* [example] fix opt example requirements

* [example] fix opt example

* [example] fix opt example

* [example] fix opt example

* [gemini] add static placement policy (#4443)

* [gemini] add static placement policy

* [gemini] fix param offload

* [test] update gemini tests

* [plugin] update gemini plugin

* [plugin] update gemini plugin docstr

* [misc] fix flash attn requirement

* [test] fix gemini checkpoint io test

* [example] update resnet example result (#4457)

* [example] update bert example result (#4458)

* [doc] update gemini doc (#4468)

* [example] update gemini related examples (#4473)

* [example] update gpt example

* [example] update dreambooth example

* [example] update vit

* [example] update opt

* [example] update palm

* [example] update vit and opt benchmark

* [hotfix] fix bert in model zoo (#4480)

* [hotfix] fix bert in model zoo

* [test] remove chatglm gemini test

* [test] remove sam gemini test

* [test] remove vit gemini test

* [hotfix] fix opt tutorial example (#4497)

* [hotfix] fix opt tutorial example

* [hotfix] fix opt tutorial example
This commit is contained in:
Hongxin Liu
2023-08-24 09:29:25 +08:00
committed by GitHub
parent 285fe7ba71
commit 27061426f7
82 changed files with 1008 additions and 4036 deletions

View File

@@ -1,8 +1,9 @@
import pytest
import torch
from torch.distributed.distributed_c10d import _get_default_group
import colossalai
from colossalai.tensor import ColoTensor, ColoTensorSpec, ProcessGroup
from colossalai.tensor import ColoTensor
from colossalai.testing import parameterize, rerun_if_address_is_in_use, spawn
from colossalai.zero.gemini.chunk import ChunkManager
from tests.test_tensor.common_utils import debug_print
@@ -15,19 +16,18 @@ CPU_MEM = {True: {True: 0, False: 0}, False: {True: 512, False: 0}}
@parameterize('keep_gathered', [True, False])
@parameterize('pin_memory', [True, False])
def exam_chunk_memory(keep_gathered, pin_memory):
pg = ProcessGroup()
debug_print([0], "keep_gathered: {}, pin_memory: {}".format(keep_gathered, pin_memory))
params = [ColoTensor(torch.rand(8, 8), spec=ColoTensorSpec(pg)) for _ in range(3)]
params = [ColoTensor(torch.rand(8, 8)) for _ in range(3)]
config = {2: dict(chunk_size=128, keep_gathered=keep_gathered)}
chunk_manager = ChunkManager(config)
assert chunk_manager.total_mem['cpu'] == 0
assert chunk_manager.total_mem['cuda'] == 0
process_group = _get_default_group()
for p in params:
chunk_manager.register_tensor(p, 'param', 2, pin_memory=pin_memory)
chunk_manager.register_tensor(p, 'param', 2, process_group, pin_memory=pin_memory)
chunk_manager.close_all_groups()
assert chunk_manager.total_mem['cpu'] == CPU_MEM[keep_gathered][pin_memory]
assert chunk_manager.total_mem['cuda'] == CUDA_MEM_0[keep_gathered]