repo stringlengths 7 90 | file_url stringlengths 81 315 | file_path stringlengths 4 228 | content stringlengths 0 32.8k | language stringclasses 1
value | license stringclasses 7
values | commit_sha stringlengths 40 40 | retrieved_at stringdate 2026-01-04 14:38:15 2026-01-05 02:33:18 | truncated bool 2
classes |
|---|---|---|---|---|---|---|---|---|
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/booster/plugin/dp_plugin_base.py | colossalai/booster/plugin/dp_plugin_base.py | import random
import numpy as np
import torch
import torch.distributed as dist
from torch.utils.data import DataLoader
from torch.utils.data.distributed import DistributedSampler
from .plugin_base import Plugin
class DPPluginBase(Plugin):
"""This is a base class for all DP plugins. It sets up world size and ran... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/booster/plugin/plugin_base.py | colossalai/booster/plugin/plugin_base.py | from abc import ABC, abstractmethod
from typing import Callable, Dict, Iterator, List, Optional, Tuple
import torch.nn as nn
from torch.optim import Optimizer
from torch.optim.lr_scheduler import _LRScheduler as LRScheduler
from torch.utils.data import DataLoader, Dataset
from colossalai.checkpoint_io import Checkpoi... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/booster/plugin/pp_plugin_base.py | colossalai/booster/plugin/pp_plugin_base.py | from abc import abstractmethod
from typing import Any, Callable, Iterator, Optional
import torch
from colossalai.interface import ModelWrapper, OptimizerWrapper
from .plugin_base import Plugin
class PipelinePluginBase(Plugin):
@abstractmethod
def execute_pipeline(
self,
data_iter: Iterator,... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/booster/plugin/__init__.py | colossalai/booster/plugin/__init__.py | from .gemini_plugin import GeminiPlugin
from .hybrid_parallel_plugin import HybridParallelPlugin
from .low_level_zero_plugin import LowLevelZeroPlugin
from .moe_hybrid_parallel_plugin import MoeHybridParallelPlugin
from .plugin_base import Plugin
from .torch_ddp_plugin import TorchDDPPlugin
__all__ = [
"Plugin",
... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/booster/plugin/gemini_plugin.py | colossalai/booster/plugin/gemini_plugin.py | import os
import random
from pathlib import Path
from typing import Callable, Dict, Iterator, List, Optional, Tuple
import numpy as np
import torch
import torch.distributed as dist
import torch.nn as nn
from torch.distributed.distributed_c10d import _get_default_group
from torch.optim import Optimizer
from torch.optim... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | true |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/booster/plugin/moe_hybrid_parallel_plugin.py | colossalai/booster/plugin/moe_hybrid_parallel_plugin.py | from collections import defaultdict
from types import MethodType
from typing import Callable, List, Optional, OrderedDict, Tuple
import torch
import torch.distributed as dist
from torch.distributed import ProcessGroup
from torch.nn import Module
from torch.optim import Optimizer
from torch.optim.lr_scheduler import _L... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/booster/plugin/hybrid_parallel_plugin.py | colossalai/booster/plugin/hybrid_parallel_plugin.py | import ctypes
import random
from collections import defaultdict
from contextlib import contextmanager, nullcontext
from copy import deepcopy
from functools import partial
from types import MethodType
from typing import Any, Callable, Dict, Iterator, List, Optional, OrderedDict, Tuple, Union
import numpy as np
import t... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | true |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/booster/mixed_precision/fp16_naive.py | colossalai/booster/mixed_precision/fp16_naive.py | from .mixed_precision_base import MixedPrecision
class FP16NaiveMixedPrecision(MixedPrecision):
"""
Precision for mixed precision training in FP16 using naive AMP.
Args:
log_num_zeros_in_grad(bool): return number of zeros in the gradients.
initial_scale(int): initial scale of gradient scaler.
... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/booster/mixed_precision/bf16.py | colossalai/booster/mixed_precision/bf16.py | from .mixed_precision_base import MixedPrecision
class BF16MixedPrecision(MixedPrecision):
pass
| python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/booster/mixed_precision/fp16_apex.py | colossalai/booster/mixed_precision/fp16_apex.py | from typing import Any, Optional, Union
import torch
from .mixed_precision_base import MixedPrecision
class FP16ApexMixedPrecision(MixedPrecision):
"""
Precision for mixed precision training in FP16 using apex AMP.
Args:
opt_level(str, optional, default="O1" ): Pure or mixed precision optimizat... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/booster/mixed_precision/fp16_torch.py | colossalai/booster/mixed_precision/fp16_torch.py | from typing import Callable, Optional, Tuple, Union
import torch
import torch.nn as nn
from torch import Tensor
from torch.optim import Optimizer
from colossalai.accelerator import get_accelerator
from colossalai.interface import ModelWrapper, OptimizerWrapper
from .mixed_precision_base import MixedPrecision
__all_... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/booster/mixed_precision/fp8.py | colossalai/booster/mixed_precision/fp8.py | from .mixed_precision_base import MixedPrecision
class FP8MixedPrecision(MixedPrecision):
pass
| python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/booster/mixed_precision/__init__.py | colossalai/booster/mixed_precision/__init__.py | from .bf16 import BF16MixedPrecision
from .fp8 import FP8MixedPrecision
from .fp16_apex import FP16ApexMixedPrecision
from .fp16_naive import FP16NaiveMixedPrecision
from .fp16_torch import FP16TorchMixedPrecision
from .mixed_precision_base import MixedPrecision
__all__ = [
"MixedPrecision",
"mixed_precision_f... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/booster/mixed_precision/mixed_precision_base.py | colossalai/booster/mixed_precision/mixed_precision_base.py | from abc import ABC, abstractmethod
from typing import Callable, Optional, Tuple
import torch.nn as nn
from torch.optim import Optimizer
from colossalai.interface import OptimizerWrapper
class MixedPrecision(ABC):
"""
An abstract class for mixed precision training.
"""
@abstractmethod
def confi... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/amp/__init__.py | colossalai/amp/__init__.py | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false | |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/amp/naive_amp/mixed_precision_optimizer.py | colossalai/amp/naive_amp/mixed_precision_optimizer.py | from typing import Dict, List, Optional, Tuple
import torch
from torch import Tensor, inf
from torch.nn import Module, Parameter
from torch.optim import Optimizer
from colossalai.interface import OptimizerWrapper
from .mixed_precision_mixin import BF16MixedPrecisionMixin, FP16MixedPrecisionMixin
class NaiveFP16Mix... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/amp/naive_amp/__init__.py | colossalai/amp/naive_amp/__init__.py | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false | |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/amp/naive_amp/grad_scaler/constant_grad_scaler.py | colossalai/amp/naive_amp/grad_scaler/constant_grad_scaler.py | #!/usr/bin/env python
# -*- encoding: utf-8 -*-
from .base_grad_scaler import BaseGradScaler
__all__ = ["ConstantGradScaler"]
class ConstantGradScaler(BaseGradScaler):
"""A gradient scaler which uses constant loss scale
Args:
initial_scale (float): the initial loss scale
verbose (bool): whet... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/amp/naive_amp/grad_scaler/base_grad_scaler.py | colossalai/amp/naive_amp/grad_scaler/base_grad_scaler.py | #!/usr/bin/env python
# -*- encoding: utf-8 -*-
from abc import ABC, abstractmethod
from typing import Dict
import torch
from torch import Tensor
from colossalai.accelerator import get_accelerator
from colossalai.logging import get_dist_logger
__all__ = ["BaseGradScaler"]
class BaseGradScaler(ABC):
"""A base ... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/amp/naive_amp/grad_scaler/dynamic_grad_scaler.py | colossalai/amp/naive_amp/grad_scaler/dynamic_grad_scaler.py | #!/usr/bin/env python
# -*- encoding: utf-8 -*-
from typing import Optional
import torch
from colossalai.accelerator import get_accelerator
from .base_grad_scaler import BaseGradScaler
__all__ = ["DynamicGradScaler"]
class DynamicGradScaler(BaseGradScaler):
"""A gradient scaler which uses dynamic loss scale
... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/amp/naive_amp/grad_scaler/__init__.py | colossalai/amp/naive_amp/grad_scaler/__init__.py | from .base_grad_scaler import BaseGradScaler
from .constant_grad_scaler import ConstantGradScaler
from .dynamic_grad_scaler import DynamicGradScaler
__all__ = ["BaseGradScaler", "ConstantGradScaler", "DynamicGradScaler"]
| python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/amp/naive_amp/mixed_precision_mixin/fp16.py | colossalai/amp/naive_amp/mixed_precision_mixin/fp16.py | from abc import abstractmethod
from enum import Enum
import torch
import torch.distributed as dist
from torch import Tensor
from colossalai.accelerator import get_accelerator
from colossalai.amp.naive_amp.grad_scaler import DynamicGradScaler
from .base import MixedPrecisionMixin
class OptimState(Enum):
SCALED ... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/amp/naive_amp/mixed_precision_mixin/bf16.py | colossalai/amp/naive_amp/mixed_precision_mixin/bf16.py | import torch
from torch import Tensor
from .base import MixedPrecisionMixin
class BF16MixedPrecisionMixin(MixedPrecisionMixin):
dtype = torch.bfloat16
def pre_backward(self, loss: Tensor) -> Tensor:
return loss
def pre_backward_by_grad(self, tensor: Tensor, grad: Tensor) -> Tensor:
retu... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/amp/naive_amp/mixed_precision_mixin/__init__.py | colossalai/amp/naive_amp/mixed_precision_mixin/__init__.py | from .base import MixedPrecisionMixin
from .bf16 import BF16MixedPrecisionMixin
from .fp16 import FP16MixedPrecisionMixin
__all__ = [
"MixedPrecisionMixin",
"FP16MixedPrecisionMixin",
"BF16MixedPrecisionMixin",
]
| python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/amp/naive_amp/mixed_precision_mixin/base.py | colossalai/amp/naive_amp/mixed_precision_mixin/base.py | from abc import ABC, abstractmethod
import torch
from torch import Tensor
class MixedPrecisionMixin(ABC):
"""A helper class for mixed precision training. This mixin is used in mixed precision optimizers.
Attributes:
dtype (torc.dtype): The expected dtype of the gradients.
Examples:
```p... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/utils/common.py | colossalai/utils/common.py | #!/usr/bin/env python
# -*- encoding: utf-8 -*-
import functools
import os
import random
from contextlib import contextmanager
from pathlib import Path
from typing import Callable, Optional, Set
import numpy as np
import torch
import torch.nn as nn
from colossalai.accelerator import get_accelerator
def get_current_... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/utils/timer.py | colossalai/utils/timer.py | #!/usr/bin/env python
# -*- encoding: utf-8 -*-
import time
from typing import Tuple
from colossalai.accelerator import get_accelerator
class Timer:
"""A timer object which helps to log the execution times, and provides different tools to assess the times."""
def __init__(self):
self._started = Fals... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/utils/memory.py | colossalai/utils/memory.py | from collections import namedtuple
import psutil
import torch
import torch.distributed as dist
from colossalai.utils import get_current_device
_GLOBAL_CUDA_MEM_FRACTION = 1.0
_GLOBAL_CPU_MEM_CAPACITY = -1
# copy from PatrickStar
def _get_cpu_memory_info():
ps_mem_info = namedtuple("ps_mem_info", ["total", "fre... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/utils/__init__.py | colossalai/utils/__init__.py | from .common import (
_cast_float,
conditional_context,
disposable,
ensure_path_exists,
free_storage,
get_current_device,
get_non_persistent_buffers_set,
is_ddp_ignored,
set_seed,
)
from .multi_tensor_apply import multi_tensor_applier
from .tensor_detector import TensorDetector
from ... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/utils/safetensors.py | colossalai/utils/safetensors.py | # a python safetensors serializer modified from https://github.com/huggingface/safetensors/blob/41bd1acf38ad28ac559522d40596c6c802f79453/safetensors/src/tensor.rs#L214
import json
import warnings
from dataclasses import asdict, dataclass
from typing import Dict, List, Optional, Tuple
import torch
from safetensors.torc... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/utils/rank_recorder/rank_recorder.py | colossalai/utils/rank_recorder/rank_recorder.py | import atexit
import json
import os
import shutil
import time
from typing import Dict, List
import matplotlib.colors as mcolors
import matplotlib.pyplot as plt
import torch
import torch.distributed as dist
cmap = list(mcolors.TABLEAU_COLORS.values())
LOG_FOLDER = "record.log"
MAX_WAIT_TIME = 20
class Event:
de... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/utils/rank_recorder/__init__.py | colossalai/utils/rank_recorder/__init__.py | from colossalai.utils.rank_recorder.rank_recorder import recorder
__all__ = ["recorder"]
| python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/utils/model/utils.py | colossalai/utils/model/utils.py | # This code has been adapted from the DeepSpeed library.
# Copyright (c) Microsoft Corporation.
# Licensed under the MIT License.
import functools
from typing import Optional
import torch
def substitute_init_recursively(cls, func, visited: set):
for subcls in cls.__subclasses__():
substitute_init_recurs... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/utils/model/__init__.py | colossalai/utils/model/__init__.py | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false | |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/utils/multi_tensor_apply/__init__.py | colossalai/utils/multi_tensor_apply/__init__.py | from .multi_tensor_apply import MultiTensorApply
multi_tensor_applier = MultiTensorApply(2048 * 32)
| python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/utils/multi_tensor_apply/multi_tensor_apply.py | colossalai/utils/multi_tensor_apply/multi_tensor_apply.py | # modified from https://github.com/NVIDIA/apex/blob/master/apex/multi_tensor_apply/multi_tensor_apply.py
class MultiTensorApply(object):
"""
Apply an operation to a list of tensors efficiently.
Args:
chunk_size (int): Size of a chunk.
"""
available = False
warned = False
def __i... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/utils/tensor_detector/tensor_detector.py | colossalai/utils/tensor_detector/tensor_detector.py | import gc
import inspect
from collections import defaultdict
from typing import Optional
import torch
import torch.nn as nn
LINE_WIDTH = 108
LINE = "-" * LINE_WIDTH + "\n"
class TensorDetector:
def __init__(
self, show_info: bool = True, log: str = None, include_cpu: bool = False, module: Optional[nn.Mo... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/utils/tensor_detector/__init__.py | colossalai/utils/tensor_detector/__init__.py | from .tensor_detector import TensorDetector
| python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/tensor/colo_parameter.py | colossalai/tensor/colo_parameter.py | from typing import Optional
import torch
from colossalai.tensor.colo_tensor import ColoTensor
from colossalai.tensor.param_op_hook import ColoParamOpHookManager
from .colo_tensor import _convert_output
WHITE_LIST_FUNCS = {torch.Tensor.__getitem__}
NO_HOOK_FUNCS = {torch.Tensor.is_floating_point}
def is_no_hook_op... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/tensor/shape_consistency.py | colossalai/tensor/shape_consistency.py | import math
from copy import deepcopy
from dataclasses import dataclass
from typing import Dict, List, Tuple
import numpy as np
import torch
from colossalai.auto_parallel.tensor_shard.sharding_strategy import MemoryCost, TrainCycleItem
from colossalai.context.singleton_meta import SingletonMeta
from colossalai.tensor... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | true |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/tensor/comm_spec.py | colossalai/tensor/comm_spec.py | import operator
from enum import Enum
from functools import reduce
import torch
import torch.distributed as dist
from torch.distributed import ReduceOp
__all__ = [
"CollectiveCommPattern",
"CommSpec",
]
def _all_gather(tensor, comm_spec):
"""
Implement all gather operation on device mesh based on in... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/tensor/colo_tensor.py | colossalai/tensor/colo_tensor.py | from functools import lru_cache
from typing import Callable, Set
import torch
INPALCE_MAPPING = {
torch.Tensor.add_: torch.Tensor.add,
torch.Tensor.sub_: torch.Tensor.sub,
torch.Tensor.mul_: torch.Tensor.mul,
torch.Tensor.div_: torch.Tensor.div,
}
@lru_cache(None)
def _get_my_nowrap_functions() -> S... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/tensor/utils.py | colossalai/tensor/utils.py | from typing import Dict, Iterator, List, Tuple, Union
import torch
import torch.nn as nn
from colossalai.tensor.colo_tensor import ColoTensor
def all_gather_simulator(target_pair):
"""
Simulating all-gather operation, analyze the communication cost
and simulate the influence of the DimSpec.
We don'... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/tensor/param_op_hook.py | colossalai/tensor/param_op_hook.py | from abc import ABC, abstractmethod
from contextlib import contextmanager
from typing import Any, List, Tuple
import torch
from torch.utils._pytree import TreeSpec, tree_flatten, tree_unflatten
class ColoParamOpHook(ABC):
"""
Hook which is triggered by each operation when operands contain ColoParameter.
... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/tensor/__init__.py | colossalai/tensor/__init__.py | from .colo_parameter import ColoParameter
from .colo_tensor import ColoTensor
from .comm_spec import CollectiveCommPattern, CommSpec
from .param_op_hook import ColoParamOpHook, ColoParamOpHookManager
from .utils import convert_dim_partition_dict, convert_parameter, merge_same_dim_mesh_list, named_params_with_colotensor... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/tensor/sharding_spec.py | colossalai/tensor/sharding_spec.py | import operator
from functools import reduce
import torch
from colossalai.device.device_mesh import DeviceMesh
from .utils import merge_same_dim_mesh_list
__all__ = ["_DimSpec", "ShardingException", "ShardingSpec"]
ALLGATHER_COST = 20
SHARD_COST = 5
STEP_PENALTY = 6
NAN = "nan"
class _DimSpec:
"""
Shardi... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/tensor/moe_tensor/api.py | colossalai/tensor/moe_tensor/api.py | from typing import List
import torch
import torch.distributed as dist
from torch.distributed import ProcessGroup
from .moe_info import MoeParallelInfo
def is_moe_tensor(tensor: torch.Tensor) -> bool:
"""
Check whether the given tensor is a moe tensor.
Args:
tensor (torch.Tensor): The tensor to ... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/tensor/moe_tensor/moe_info.py | colossalai/tensor/moe_tensor/moe_info.py | from colossalai.cluster import ProcessGroupMesh
class MoeParallelInfo:
"""Moe parallelism information, storing parallel sizes and groups."""
def __init__(self, ep_inside: bool, ep_size: int, dp_size: int, pp_size: int = 1):
"""
init MoeParallelInfo with ep_size, dp_size and pp_size
A... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/tensor/moe_tensor/__init__.py | colossalai/tensor/moe_tensor/__init__.py | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false | |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/tensor/padded_tensor/api.py | colossalai/tensor/padded_tensor/api.py | import torch
def _hijack_detach_and_clone(ptensor: torch.Tensor) -> torch.Tensor:
"""
Hijack the detach and clone methods of the tensor to make sure the dist_layout is copied.
Args:
tensor (torch.Tensor): The tensor to be hijacked.
Returns:
torch.Tensor: The hijacked tensor.
"""
... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/tensor/padded_tensor/__init__.py | colossalai/tensor/padded_tensor/__init__.py | from .api import init_as_padded_tensor, is_padded_tensor, to_padded_tensor, to_unpadded_tensor
__all__ = ["is_padded_tensor", "to_padded_tensor", "to_unpadded_tensor", "init_as_padded_tensor"]
| python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/tensor/d_tensor/api.py | colossalai/tensor/d_tensor/api.py | import copy
import operator
from functools import reduce
from typing import Union
import torch
import torch.distributed as dist
from torch.distributed import ProcessGroup
from colossalai.device.device_mesh import DeviceMesh
from colossalai.tensor.d_tensor.sharding_spec import DimSpec
from .layout import Layout
from ... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/tensor/d_tensor/comm_spec.py | colossalai/tensor/d_tensor/comm_spec.py | from enum import Enum
from typing import Dict
import torch
import torch.distributed as dist
from torch.distributed import ReduceOp
__all__ = [
"CollectiveCommPattern",
"CommSpec",
]
class CollectiveCommPattern(Enum):
GATHER_FWD_SPLIT_BWD = "gather_fwd_split_bwd"
ALL2ALL_FWD_ALL2ALL_BWD = "all2all_fw... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/tensor/d_tensor/layout_converter.py | colossalai/tensor/d_tensor/layout_converter.py | import math
from copy import deepcopy
from dataclasses import dataclass
from typing import Dict, List, Tuple
import torch
import torch.distributed as dist
from colossalai.context.singleton_meta import SingletonMeta
from colossalai.tensor.d_tensor.comm_spec import *
from colossalai.tensor.d_tensor.layout import Layout... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/tensor/d_tensor/utils.py | colossalai/tensor/d_tensor/utils.py | import operator
from functools import reduce
from typing import Dict
from colossalai.tensor.d_tensor.comm_spec import CollectiveCommPattern, CommSpec
from colossalai.tensor.d_tensor.layout import Layout
def get_comm_cost(layout: Layout, comm_spec: CommSpec, forward_only: bool = False) -> Dict[str, float]:
"""
... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/tensor/d_tensor/misc.py | colossalai/tensor/d_tensor/misc.py | class LayoutException(Exception):
pass
class DuplicatedShardingDimensionError(LayoutException):
pass
class ShardingNotDivisibleError(LayoutException):
pass
class ShardingOutOfIndexError(LayoutException):
pass
| python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/tensor/d_tensor/__init__.py | colossalai/tensor/d_tensor/__init__.py | from .api import (
compute_global_numel,
customized_distributed_tensor_to_param,
distribute_tensor,
distribute_tensor_with_customization,
get_device_mesh,
get_global_shape,
get_layout,
get_shard_dim_1d,
get_sharding_spec,
init_as_dtensor,
init_tensor_as_customization_distribu... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/tensor/d_tensor/sharding_spec.py | colossalai/tensor/d_tensor/sharding_spec.py | from typing import Dict, List
from ..utils import merge_same_dim_mesh_list
from .misc import ShardingOutOfIndexError
__all__ = ["DimSpec", "ShardingException", "ShardingSpec"]
ALLGATHER_COST = 20
SHARD_COST = 5
STEP_PENALTY = 6
NAN = "nan"
class DimSpec:
"""
Sharding spec for single dimension of the sharde... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/tensor/d_tensor/layout.py | colossalai/tensor/d_tensor/layout.py | import operator
from functools import reduce
import torch
from colossalai.device.device_mesh import DeviceMesh
from .misc import DuplicatedShardingDimensionError, ShardingNotDivisibleError
from .sharding_spec import ShardingSpec
class Layout:
"""Layout of a tensor.
Attributes:
device_mesh: the dev... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/checkpoint_io/checkpoint_io_base.py | colossalai/checkpoint_io/checkpoint_io_base.py | from abc import ABC, abstractmethod
from pathlib import Path
from typing import Dict, Optional, Union
import torch
import torch.nn as nn
from torch.optim import Optimizer
from torch.optim.lr_scheduler import _LRScheduler as LRScheduler
from colossalai.interface import ModelWrapper
from colossalai.logging import get_d... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/checkpoint_io/index_file.py | colossalai/checkpoint_io/index_file.py | import json
import os
from collections import OrderedDict
from pathlib import Path
from typing import Any, Dict, List, Union
from .utils import is_dtensor_checkpoint
__all__ = ["CheckpointIndexFile"]
class CheckpointIndexFile:
"""
This class is a data structure to keep the content in the index.json file for... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/checkpoint_io/general_checkpoint_io.py | colossalai/checkpoint_io/general_checkpoint_io.py | import logging
import os
from functools import reduce
from pathlib import Path
from typing import Optional
import torch.nn as nn
from torch.optim import Optimizer
from colossalai.utils.safetensors import load_flat
from .checkpoint_io_base import CheckpointIO
from .index_file import CheckpointIndexFile
from .utils im... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/checkpoint_io/utils.py | colossalai/checkpoint_io/utils.py | # coding=utf-8
import concurrent.futures
import os
import re
import warnings
from collections import abc as container_abcs
from collections import defaultdict
from itertools import chain
from pathlib import Path
from typing import Dict, Generator, Iterator, List, Mapping, Optional, OrderedDict, Tuple, Union
import tor... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | true |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/checkpoint_io/__init__.py | colossalai/checkpoint_io/__init__.py | from .checkpoint_io_base import CheckpointIO
from .general_checkpoint_io import GeneralCheckpointIO
from .hybrid_parallel_checkpoint_io import HybridParallelCheckpointIO
from .index_file import CheckpointIndexFile
from .moe_checkpoint import MoECheckpointIO
__all__ = [
"CheckpointIO",
"CheckpointIndexFile",
... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/checkpoint_io/hybrid_parallel_checkpoint_io.py | colossalai/checkpoint_io/hybrid_parallel_checkpoint_io.py | import copy
import logging
import os
from collections import defaultdict
from functools import reduce
from pathlib import Path
from shutil import rmtree
from typing import Dict, Iterator, Optional, OrderedDict, Tuple
import torch
import torch.distributed as dist
import torch.nn as nn
from torch.distributed import Proc... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | true |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/checkpoint_io/moe_checkpoint.py | colossalai/checkpoint_io/moe_checkpoint.py | import copy
import logging
import os
from pathlib import Path
from shutil import rmtree
from typing import Dict, Iterator, Optional, OrderedDict, Tuple
import torch
import torch.distributed as dist
import torch.nn as nn
from torch.distributed import ProcessGroup
from torch.distributed.distributed_c10d import get_globa... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | true |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/accelerator/npu_accelerator.py | colossalai/accelerator/npu_accelerator.py | #!/usr/bin/env python
from typing import Any, Callable, Dict, List, Optional, Tuple, Union
import torch
import torch.distributed as dist
from .base_accelerator import BaseAccelerator
try:
import torch_npu # noqa
except ImportError:
pass
__all__ = ["NpuAccelerator"]
class NpuAccelerator(BaseAccelerator)... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/accelerator/api.py | colossalai/accelerator/api.py | #!/usr/bin/env python
from collections import OrderedDict
from typing import Union
from .base_accelerator import BaseAccelerator
from .cpu_accelerator import CpuAccelerator
from .cuda_accelerator import CudaAccelerator
from .npu_accelerator import NpuAccelerator
__all__ = ["set_accelerator", "auto_set_accelerator", "... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/accelerator/cpu_accelerator.py | colossalai/accelerator/cpu_accelerator.py | #!/usr/bin/env python
import resource
from contextlib import nullcontext
from typing import Any, Callable, Dict, List, Optional, Tuple, Union
import psutil
import torch
from .base_accelerator import BaseAccelerator
__all__ = ["CpuAccelerator"]
class CpuAccelerator(BaseAccelerator):
support_set_device: bool = ... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/accelerator/cuda_accelerator.py | colossalai/accelerator/cuda_accelerator.py | #!/usr/bin/env python
from typing import Any, Callable, Dict, List, Optional, Tuple, Union
import torch
import torch.distributed as dist
from .base_accelerator import BaseAccelerator
__all__ = ["CudaAccelerator"]
class CudaAccelerator(BaseAccelerator):
"""
Accelerator class for Nvidia CUDA devices.
""... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/accelerator/__init__.py | colossalai/accelerator/__init__.py | from .api import auto_set_accelerator, get_accelerator, set_accelerator
from .base_accelerator import BaseAccelerator
from .cpu_accelerator import CpuAccelerator
from .cuda_accelerator import CudaAccelerator
from .npu_accelerator import NpuAccelerator
__all__ = [
"get_accelerator",
"set_accelerator",
"auto... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/accelerator/base_accelerator.py | colossalai/accelerator/base_accelerator.py | #!/usr/bin/env python
from abc import ABC, abstractmethod
from typing import Any, Callable, Dict, List, Optional, Tuple, Union
import torch
__all__ = ["BaseAccelerator"]
class BaseAccelerator(ABC):
support_set_device: bool = True
def __init__(self, name: str, communication_backend: str, is_synchronous: bo... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/cluster/dist_coordinator.py | colossalai/cluster/dist_coordinator.py | import functools
import os
from contextlib import contextmanager
import torch.distributed as dist
from torch.distributed import ProcessGroup
from colossalai.context.singleton_meta import SingletonMeta
class DistCoordinator(metaclass=SingletonMeta):
"""
This class is used to coordinate distributed training. ... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/cluster/__init__.py | colossalai/cluster/__init__.py | from .device_mesh_manager import DeviceMeshManager
from .dist_coordinator import DistCoordinator
from .process_group_manager import ProcessGroupManager
from .process_group_mesh import ProcessGroupMesh
__all__ = ["DistCoordinator", "ProcessGroupManager", "DeviceMeshManager", "ProcessGroupMesh"]
| python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/cluster/process_group_mesh.py | colossalai/cluster/process_group_mesh.py | import gc
import itertools
from functools import reduce
from operator import mul
from typing import Dict, List, Optional, Tuple, Union
import numpy as np
import torch.distributed as dist
from torch.distributed import ProcessGroup
from torch.distributed.distributed_c10d import GroupMember
def prod(nums: List[int]) ->... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/cluster/process_group_manager.py | colossalai/cluster/process_group_manager.py | from typing import List
import torch.distributed as dist
from torch.distributed import ProcessGroup
class ProcessGroupManager:
"""
ProcessGroupManager is used to manage the process groups in the cluster.
There are some terms used in this class:
- pg: the short name for process group
- pg... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/cluster/device_mesh_manager.py | colossalai/cluster/device_mesh_manager.py | from dataclasses import dataclass
from typing import Dict, List, Tuple, Union
import torch
import torch.distributed as dist
from colossalai.device.alpha_beta_profiler import AlphaBetaProfiler
from colossalai.device.device_mesh import DeviceMesh
@dataclass
class DeviceMeshInfo:
"""
This class is used to stor... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/testing/pytest_wrapper.py | colossalai/testing/pytest_wrapper.py | """
This file will not be automatically imported by `colossalai.testing`
as this file has a dependency on `pytest`. Therefore, you need to
explicitly import this file `from colossalai.testing.pytest_wrapper import <func>`.from
"""
import os
def run_on_environment_flag(name: str):
"""
Conditionally run a test... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/testing/comparison.py | colossalai/testing/comparison.py | from typing import Any, List, OrderedDict
import torch
import torch.distributed as dist
from torch import Tensor
from torch.distributed import ProcessGroup
from torch.testing import assert_close
from torch.utils._pytree import tree_flatten
def assert_equal(a: Tensor, b: Tensor):
assert torch.all(a == b), f"expec... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/testing/utils.py | colossalai/testing/utils.py | import gc
import random
import re
import socket
from functools import partial
from inspect import signature
from typing import Any, Callable, List
import torch
import torch.multiprocessing as mp
from packaging import version
from colossalai.accelerator import get_accelerator
def parameterize(argument: str, values: ... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/testing/random.py | colossalai/testing/random.py | import random
import numpy as np
import torch
def seed_all(seed, cuda_deterministic=False):
random.seed(seed)
np.random.seed(seed)
torch.manual_seed(seed)
if torch.cuda.is_available():
torch.cuda.manual_seed(seed)
torch.cuda.manual_seed_all(seed)
if cuda_deterministic: # slower, ... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/testing/__init__.py | colossalai/testing/__init__.py | from .comparison import (
assert_close,
assert_close_loose,
assert_equal,
assert_equal_in_group,
assert_hf_output_close,
assert_not_equal,
check_state_dict_equal,
)
from .pytest_wrapper import run_on_environment_flag
from .utils import (
DummyDataloader,
clear_cache_before_run,
f... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/legacy/constants.py | colossalai/legacy/constants.py | #!/usr/bin/env python
# -*- encoding: utf-8 -*-
ALLOWED_MODES = [None, "1d", "2d", "2.5d", "3d", "sequence"]
TENSOR_PARALLEL_MODE = "tensor_parallel_mode"
# initializer
INITIALIZER_MAPPING = {
"data": "Initializer_Data",
"tensor": "Initializer_Tensor",
"pipeline": "Initializer_Pipeline",
"embedding": ... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/legacy/core.py | colossalai/legacy/core.py | #!/usr/bin/env python
# -*- encoding: utf-8 -*-
from colossalai.legacy.context.parallel_context import global_context
__all__ = ["global_context"]
| python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/legacy/global_variables.py | colossalai/legacy/global_variables.py | from typing import Optional
class TensorParallelEnv(object):
_instance = None
def __new__(cls, *args, **kwargs):
if cls._instance is None:
cls._instance = object.__new__(cls, *args, **kwargs)
return cls._instance
def __init__(self, *args, **kwargs):
self.load(*args, *... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/legacy/__init__.py | colossalai/legacy/__init__.py | from .initialize import (
get_default_parser,
initialize,
launch,
launch_from_openmpi,
launch_from_slurm,
launch_from_torch,
)
__all__ = [
"launch",
"launch_from_openmpi",
"launch_from_slurm",
"launch_from_torch",
"initialize",
"get_default_parser",
]
| python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/legacy/initialize.py | colossalai/legacy/initialize.py | #!/usr/bin/env python
# -*- encoding: utf-8 -*-
import argparse
import os
import pprint
from pathlib import Path
from typing import Callable, Dict, Iterable, List, Optional, Tuple, Union
import torch
import torch.nn as nn
from torch.nn.modules.loss import _Loss
from torch.nn.parallel import DistributedDataParallel as... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/legacy/zero/__init__.py | colossalai/legacy/zero/__init__.py | from typing import Tuple
import torch
import torch.nn as nn
from colossalai.logging import get_dist_logger
from .init_ctx import ZeroInitContext, no_shard_zero_context, no_shard_zero_decrator
from .shard_utils import BucketTensorShardStrategy, TensorShardStrategy
from .sharded_model import ShardedModelV2
from .shard... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/legacy/zero/sharded_param/sharded_tensor.py | colossalai/legacy/zero/sharded_param/sharded_tensor.py | import torch
from colossalai.legacy.zero.gemini.stateful_tensor import StatefulTensor, TensorState
class ShardedTensor(StatefulTensor):
def __init__(self, tensor: torch.Tensor, state: TensorState = TensorState.HOLD) -> None:
r"""
A tensor sharded in multiple processes. Constructed from an existin... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/legacy/zero/sharded_param/__init__.py | colossalai/legacy/zero/sharded_param/__init__.py | from .sharded_param import ShardedParamV2
from .sharded_tensor import ShardedTensor
__all__ = ["ShardedTensor", "ShardedParamV2"]
| python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/legacy/zero/sharded_param/sharded_param.py | colossalai/legacy/zero/sharded_param/sharded_param.py | from typing import List, Optional, Tuple
import torch
from colossalai.legacy.zero.gemini.stateful_tensor import StatefulTensor, TensorState
from colossalai.legacy.zero.gemini.tensor_utils import colo_tensor_mem_usage
from .sharded_tensor import ShardedTensor
EMPTY_TENSOR_DICT = {}
def get_empty_tensor(device: tor... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/legacy/zero/init_ctx/__init__.py | colossalai/legacy/zero/init_ctx/__init__.py | from .init_context import ZeroInitContext, no_shard_zero_context, no_shard_zero_decrator
__all__ = ["ZeroInitContext", "no_shard_zero_context", "no_shard_zero_decrator"]
| python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/legacy/zero/init_ctx/init_context.py | colossalai/legacy/zero/init_ctx/init_context.py | import contextlib
import functools
from contextlib import AbstractContextManager
from dataclasses import dataclass
from typing import Optional
import torch
import torch.distributed as dist
import torch.nn as nn
from colossalai.context.singleton_meta import SingletonMeta
from colossalai.legacy.context.parallel_mode im... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/legacy/zero/shard_utils/base_shard_strategy.py | colossalai/legacy/zero/shard_utils/base_shard_strategy.py | from abc import ABC, abstractmethod
from typing import List, Optional
import torch.distributed as dist
from colossalai.legacy.zero.sharded_param.sharded_tensor import ShardedTensor
class BaseShardStrategy(ABC):
def __init__(self) -> None:
"""Abstract Shard Strategy. Use to shard a tensors on multiple GP... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/legacy/zero/shard_utils/bucket_tensor_shard_strategy.py | colossalai/legacy/zero/shard_utils/bucket_tensor_shard_strategy.py | from typing import List, Optional
import torch
import torch.distributed as dist
from torch._utils import _flatten_dense_tensors as flatten
from colossalai.accelerator import get_accelerator
from colossalai.legacy.zero.sharded_param.sharded_tensor import ShardedTensor
from .tensor_shard_strategy import TensorShardStr... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/legacy/zero/shard_utils/tensor_shard_strategy.py | colossalai/legacy/zero/shard_utils/tensor_shard_strategy.py | from typing import List, Optional
import torch
import torch.distributed as dist
from colossalai.accelerator import get_accelerator
from colossalai.legacy.zero.gemini.tensor_utils import colo_model_data_tensor_move_inline
from colossalai.legacy.zero.shard_utils import BaseShardStrategy
from colossalai.legacy.zero.shar... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/legacy/zero/shard_utils/commons.py | colossalai/legacy/zero/shard_utils/commons.py | from typing import Tuple
import torch
def get_shard(tensor: torch.Tensor, rank: int, world_size: int) -> Tuple[torch.Tensor, int]:
"""Return the local shard of a full tensor."""
# Shard using torch.chunk to match all-gather/reduce-scatter.
chunks = list(torch.flatten(tensor).chunk(world_size))
while ... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/legacy/zero/shard_utils/__init__.py | colossalai/legacy/zero/shard_utils/__init__.py | from .base_shard_strategy import BaseShardStrategy
from .bucket_tensor_shard_strategy import BucketTensorShardStrategy
from .tensor_shard_strategy import TensorShardStrategy
__all__ = ["BaseShardStrategy", "TensorShardStrategy", "BucketTensorShardStrategy"]
| python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/legacy/zero/sharded_optim/sharded_optim_v2.py | colossalai/legacy/zero/sharded_optim/sharded_optim_v2.py | # this code is inspired by the DeepSpeed library and implemented with our own design from scratch
from enum import Enum
from typing import Dict, Optional, Tuple
import torch
import torch.distributed as dist
import torch.nn as nn
from torch import Tensor
from torch.distributed import ProcessGroup
from torch.nn.paramete... | python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
hpcaitech/ColossalAI | https://github.com/hpcaitech/ColossalAI/blob/b1915d2889543949eb5b610241f1515c73df5059/colossalai/legacy/zero/sharded_optim/__init__.py | colossalai/legacy/zero/sharded_optim/__init__.py | from .sharded_optim_v2 import ShardedOptimizerV2
__all__ = ["ShardedOptimizerV2"]
| python | Apache-2.0 | b1915d2889543949eb5b610241f1515c73df5059 | 2026-01-04T14:40:19.002665Z | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.