github/ColossalAI
1
0
Fork 0
mirror of https://github.com/hpcaitech/ColossalAI.git synced 2025-09-10 21:40:02 +00:00
Code Issues Packages Projects Releases Wiki Activity

[misc] update pre-commit and run all files (#4752)

Browse Source 
* [misc] update pre-commit

* [misc] run pre-commit

* [misc] remove useless configuration files

* [misc] ignore cuda for clang-format
This commit is contained in:
Hongxin Liu
2023-09-19 14:20:26 +08:00
committed by GitHub
parent 3c6b831c26
commit 079bf3cb26
1268 changed files with 50037 additions and 38444 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
examples/tutorial/sequence_parallel/loss_func/bert_loss.py
View File

@@ -1,17 +1,12 @@
import torch
import torch.distributed as dist
import torch.nn as nn
import torch.nn.functional as F
from colossalai.legacy.context import ParallelMode
from colossalai.legacy.core import global_context as gpc
from colossalai.logging import get_dist_logger
from .cross_entropy import vocab_cross_entropy
class BertLoss(nn.Module):
def forward(self, lm_loss, sop_logits, loss_mask, sentence_order):
lm_loss_ = lm_loss.float()
loss_mask = loss_mask.float()

5
examples/tutorial/sequence_parallel/loss_func/cross_entropy.py
View File

@@ -1,11 +1,8 @@
import torch
from torch.cuda.amp import custom_bwd, custom_fwd
from colossalai.legacy.context.parallel_mode import ParallelMode
class _VocabCrossEntropy(torch.autograd.Function):
@staticmethod
@custom_fwd
def forward(ctx, vocab_parallel_logits, target):
@@ -59,7 +56,7 @@ class _VocabCrossEntropy(torch.autograd.Function):
# Add the gradient from matching classes.
arange_1d = torch.arange(start=0, end=grad_2d.size()[0], device=grad_2d.device)
grad_2d[arange_1d, masked_target_1d] -= (1.0 - target_mask.view(-1).float())
grad_2d[arange_1d, masked_target_1d] -= 1.0 - target_mask.view(-1).float()
# Finally elementwise multiplication with the output gradients.
grad_input.mul_(grad_output.unsqueeze(dim=-1))

17
examples/tutorial/sequence_parallel/loss_func/utils.py
View File

@@ -1,11 +1,9 @@
import torch
def ensure_divisibility(numerator, denominator):
"""Ensure that numerator is divisible by the denominator."""
assert numerator % denominator == 0, '{} is not divisible by {}'.format(
numerator, denominator)
assert numerator % denominator == 0, "{} is not divisible by {}".format(numerator, denominator)
def divide(numerator, denominator):
@@ -15,8 +13,7 @@ def divide(numerator, denominator):
return numerator // denominator
def split_tensor_along_last_dim(tensor, num_partitions,
contiguous_split_chunks=False):
def split_tensor_along_last_dim(tensor, num_partitions, contiguous_split_chunks=False):
"""Split a tensor along its last dimension.
Arguments:
tensor: input tensor.
@@ -38,12 +35,11 @@ def split_tensor_along_last_dim(tensor, num_partitions,
class VocabUtility:
"""Split the vocabulary into `world_size` chunks amd return the
first and last index of the vocabulary belonging to the `rank`
partition: Note that indices in [fist, last)"""
first and last index of the vocabulary belonging to the `rank`
partition: Note that indices in [fist, last)"""
@staticmethod
def vocab_range_from_per_partition_vocab_size(per_partition_vocab_size,
rank, world_size):
def vocab_range_from_per_partition_vocab_size(per_partition_vocab_size, rank, world_size):
index_f = rank * per_partition_vocab_size
index_l = index_f + per_partition_vocab_size
return index_f, index_l
@@ -51,5 +47,4 @@ class VocabUtility:
@staticmethod
def vocab_range_from_global_vocab_size(global_vocab_size, rank, world_size):
per_partition_vocab_size = divide(global_vocab_size, world_size)
return VocabUtility.vocab_range_from_per_partition_vocab_size(
per_partition_vocab_size, rank, world_size)
return VocabUtility.vocab_range_from_per_partition_vocab_size(per_partition_vocab_size, rank, world_size)