[legacy] clean up legacy code (#4743)

* [legacy] remove outdated codes of pipeline (#4692)

* [legacy] remove cli of benchmark and update optim (#4690)

* [legacy] remove cli of benchmark and update optim

* [doc] fix cli doc test

* [legacy] fix engine clip grad norm

* [legacy] remove outdated colo tensor (#4694)

* [legacy] remove outdated colo tensor

* [test] fix test import

* [legacy] move outdated zero to legacy (#4696)

* [legacy] clean up utils (#4700)

* [legacy] clean up utils

* [example] update examples

* [legacy] clean up amp

* [legacy] fix amp module

* [legacy] clean up gpc (#4742)

* [legacy] clean up context

* [legacy] clean core, constants and global vars

* [legacy] refactor initialize

* [example] fix examples ci

* [example] fix examples ci

* [legacy] fix tests

* [example] fix gpt example

* [example] fix examples ci

* [devops] fix ci installation

* [example] fix examples ci

tests/test_auto_parallel/test_ckpt_solvers/test_C_solver_consistency.py

@@ -6,12 +6,12 @@ import torch.fx
 import torchvision.models as tm
 
 import colossalai
-from colossalai.core import global_context as gpc
 from colossalai.fx import ColoGraphModule, ColoTracer
 from colossalai.fx._compatibility import is_compatible_with_meta
 # from colossalai.fx.passes.algorithms import solver_rotor
 # from colossalai.fx.passes.algorithms.operation import Sequence
 from colossalai.fx.passes.meta_info_prop import MetaInfoProp
+from colossalai.legacy.core import global_context as gpc
 from colossalai.testing import rerun_if_address_is_in_use, spawn
 
 if is_compatible_with_meta():

tests/test_auto_parallel/test_ckpt_solvers/test_ckpt_torchvision.py

@@ -8,12 +8,12 @@ import torchvision.models as tm
 from torch.fx import GraphModule
 
 import colossalai
-from colossalai.core import global_context as gpc
 from colossalai.fx import ColoTracer
 from colossalai.fx._compatibility import is_compatible_with_meta
 from colossalai.fx.graph_module import ColoGraphModule
 # from colossalai.fx.passes.algorithms import chen_greedy, solver_rotor
 from colossalai.fx.passes.meta_info_prop import MetaInfoProp
+from colossalai.legacy.core import global_context as gpc
 from colossalai.testing import rerun_if_address_is_in_use, spawn
 
 if is_compatible_with_meta():

tests/test_auto_parallel/test_tensor_shard/test_compatibility_with_gemini.py

@@ -13,10 +13,9 @@ from colossalai.device.device_mesh import DeviceMesh
 from colossalai.initialize import launch
 from colossalai.logging import disable_existing_loggers
 from colossalai.nn.optimizer import HybridAdam
-from colossalai.tensor.process_group import ProcessGroup
 from colossalai.testing import assert_close, rerun_if_address_is_in_use, run_on_environment_flag, spawn
 from colossalai.utils import get_current_device
-from colossalai.zero import post_process_colo_init_ctx, zero_model_wrapper, zero_optim_wrapper
+from colossalai.zero import zero_model_wrapper, zero_optim_wrapper
 
 
 class MLP(torch.nn.Module):
@@ -70,14 +69,12 @@ def check_auto_parallel_with_gemini(rank, world_size, port):
             print(strategy)
         print('=' * msg_length)
 
-    dp_process_group = ProcessGroup(rank=rank, ranks=[0, 1, 2, 3], tp_degree=2, dp_degree=2)
     gemini_config = dict(strict_ddp_mode=False,
                          device=get_current_device(),
                          placement_policy='cpu',
                          pin_memory=True,
                          search_range_m=128)
 
-    post_process_colo_init_ctx(gm, device=get_current_device(), default_pg=dp_process_group)
     gm = zero_model_wrapper(gm, zero_stage=3, gemini_config=gemini_config)
     optimizer = HybridAdam(gm.parameters(), betas=(0, 0))
     optimizer = zero_optim_wrapper(gm, optimizer, initial_scale=1)