[legacy] move communication and nn to legacy and refactor logger (#4671)

* [legacy] move communication to legacy (#4640)

* [legacy] refactor logger and clean up legacy codes (#4654)

* [legacy] make logger independent to gpc

* [legacy] make optim independent to registry

* [legacy] move test engine to legacy

* [legacy] move nn to legacy (#4656)

* [legacy] move nn to legacy

* [checkpointio] fix save hf config

* [test] remove useledd rpc pp test

* [legacy] fix nn init

* [example] skip tutorial hybriad parallel example

* [devops] test doc check

* [devops] test doc check

colossalai/nn/optimizer/cpu_adam.py

@@ -4,12 +4,10 @@ from typing import Optional
 import torch
 
 from colossalai.kernel.op_builder import CPUAdamBuilder
-from colossalai.legacy.registry import OPTIMIZERS
 
 from .nvme_optimizer import NVMeOptimizer
 
 
-@OPTIMIZERS.register_module
 class CPUAdam(NVMeOptimizer):
     """Implements Adam algorithm.
 

colossalai/nn/optimizer/fused_adam.py

@@ -8,11 +8,9 @@ Licensed under the MIT License.
 '''
 import torch
 
-from colossalai.legacy.registry import OPTIMIZERS
 from colossalai.utils import multi_tensor_applier
 
 
-@OPTIMIZERS.register_module
 class FusedAdam(torch.optim.Optimizer):
     """Implements Adam algorithm.
 

colossalai/nn/optimizer/fused_lamb.py

@@ -1,11 +1,9 @@
 # modified from https://github.com/NVIDIA/apex/blob/master/apex/optimizers/fused_lamb.py
 import torch
 
-from colossalai.legacy.registry import OPTIMIZERS
 from colossalai.utils import multi_tensor_applier
 
 
-@OPTIMIZERS.register_module
 class FusedLAMB(torch.optim.Optimizer):
     """Implements LAMB algorithm.
 

colossalai/nn/optimizer/fused_sgd.py

@@ -2,11 +2,9 @@
 import torch
 from torch.optim.optimizer import Optimizer, required
 
-from colossalai.legacy.registry import OPTIMIZERS
 from colossalai.utils import multi_tensor_applier
 
 
-@OPTIMIZERS.register_module
 class FusedSGD(Optimizer):
     r"""Implements stochastic gradient descent (optionally with momentum).
 

colossalai/nn/optimizer/hybrid_adam.py

@@ -4,13 +4,11 @@ import torch
 from torch.optim import Adam
 
 from colossalai.kernel.op_builder import FusedOptimBuilder
-from colossalai.legacy.registry import OPTIMIZERS
 from colossalai.utils import multi_tensor_applier
 
 from .cpu_adam import CPUAdam
 
 
-@OPTIMIZERS.register_module
 class HybridAdam(CPUAdam):
     """Implements Adam algorithm.
 

colossalai/nn/optimizer/lamb.py

@@ -5,10 +5,7 @@ Adapted from the pytorch-lamb library at https://github.com/cybertronai/pytorch-
 import torch
 from torch.optim import Optimizer
 
-from colossalai.legacy.registry import OPTIMIZERS
 
-
-@OPTIMIZERS.register_module
 class Lamb(Optimizer):
     r"""Implements Lamb algorithm.
     It has been proposed in `Large Batch Optimization for Deep Learning: Training BERT in 76 minutes`_.

colossalai/nn/optimizer/lars.py

@@ -5,10 +5,7 @@ from typing import Iterable
 import torch
 from torch.optim import Optimizer
 
-from colossalai.legacy.registry import OPTIMIZERS
 
-
-@OPTIMIZERS.register_module
 class Lars(Optimizer):
     r"""Implements the LARS optimizer from `"Large batch training of convolutional networks"
     <https://arxiv.org/pdf/1708.03888.pdf>`_.