mirror of
https://github.com/hpcaitech/ColossalAI.git
synced 2025-09-07 20:10:17 +00:00
[FP8] rebase main (#5963)
* add SimPO
* fix dataloader
* remove debug code
* add orpo
* fix style
* fix colossalai, transformers version
* fix colossalai, transformers version
* fix colossalai, transformers version
* fix torch colossalai version
* update transformers version
* [shardformer] DeepseekMoE support (#5871)
* [Feature] deepseek moe expert parallel implement
* [misc] fix typo, remove redundant file (#5867)
* [misc] fix typo
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
---------
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
* [Feature] deepseek support & unit test
* [misc] remove debug code & useless print
* [misc] fix typos (#5872)
* [Feature] remove modeling file, use auto config. (#5884)
* [misc] fix typos
* [Feature] deepseek support via auto model, remove modeling file
* [misc] delete useless file
* [misc] fix typos
* [Deepseek] remove redundant code (#5888)
* [misc] fix typos
* [Feature] deepseek support via auto model, remove modeling file
* [misc] delete useless file
* [misc] fix typos
* [misc] remove redundant code
* [Feature/deepseek] resolve comment. (#5889)
* [misc] fix typos
* [Feature] deepseek support via auto model, remove modeling file
* [misc] delete useless file
* [misc] fix typos
* [misc] remove redundant code
* [misc] mv module replacement into if branch
* [misc] add some warning message and modify some code in unit test
* [misc] fix typos
---------
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
* [Hoxfix] Fix CUDA_DEVICE_MAX_CONNECTIONS for comm overlap
Co-authored-by: Edenzzzz <wtan45@wisc.edu>
* [Feat] Diffusion Model(PixArtAlpha/StableDiffusion3) Support (#5838)
* Diffusion Model Inference support
* Stable Diffusion 3 Support
* pixartalpha support
* [HotFix] CI,import,requirements-test for #5838 (#5892)
* [Hot Fix] CI,import,requirements-test
---------
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
* [Feature] Enable PP + SP for llama (#5868)
* fix cross-PP-stage position id length diff bug
* fix typo
* fix typo
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
* use a one cross entropy func for all shardformer models
---------
Co-authored-by: Edenzzzz <wtan45@wisc.edu>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
* [ShardFormer] Add Ulysses Sequence Parallelism support for Command-R, Qwen2 and ChatGLM (#5897)
* add benchmark for sft, dpo, simpo, orpo. Add benchmarking result. Support lora with gradient checkpoint
* fix style
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
* fix eval
* hotfix citation
* [zero] support all-gather overlap (#5898)
* [zero] support all-gather overlap
* [zero] add overlap all-gather flag
* [misc] fix typo
* [zero] update api
* fix orpo cross entropy loss
* [Auto Parallel]: Speed up intra-op plan generation by 44% (#5446)
* Remove unnecessary calls to deepcopy
* Build DimSpec's difference dict only once
This change considerably speeds up construction speed of DimSpec objects. The difference_dict is the same for each DimSpec object, so a single copy of it is enough.
* Fix documentation of DimSpec's difference method
* [ShardFormer] fix qwen2 sp (#5903)
* [compatibility] support torch 2.2 (#5875)
* Support Pytorch 2.2.2
* keep build_on_pr file and update .compatibility
* fix object_to_tensor usage when torch>=2.3.0 (#5820)
* [misc] support torch2.3 (#5893)
* [misc] support torch2.3
* [devops] update compatibility ci
* [devops] update compatibility ci
* [devops] add debug
* [devops] add debug
* [devops] add debug
* [devops] add debug
* [devops] remove debug
* [devops] remove debug
* [release] update version (#5912)
* [plugin] support all-gather overlap for hybrid parallel (#5919)
* [plugin] fixed all-gather overlap support for hybrid parallel
* add kto
* fix style, add kto data sample
* [Examples] Add lazy init to OPT and GPT examples (#5924)
Co-authored-by: Edenzzzz <wtan45@wisc.edu>
* [ColossalChat] Hotfix for ColossalChat (#5910)
* add ignore and tiny llama
* fix path issue
* run style
* fix issue
* update bash
* add ignore and tiny llama
* fix path issue
* run style
* fix issue
* update bash
* fix ddp issue
* add Qwen 1.5 32B
* refactor tokenization
* [FIX BUG] UnboundLocalError: cannot access local variable 'default_conversation' where it is not associated with a value (#5931)
* cannot access local variable 'default_conversation' where it is not associated with a value
set default value for 'default_conversation'
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
---------
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
* fix test data
* refactor evaluation
* remove real data path
* remove real data path
* Add n_fused as an input from native_module (#5894)
* [FIX BUG] convert env param to int in (#5934)
* [Hotfix] Fix ZeRO typo #5936
Co-authored-by: Edenzzzz <wtan45@wisc.edu>
* [Feature] Add a switch to control whether the model checkpoint needs to be saved after each epoch ends (#5941)
* Add a switch to control whether the model checkpoint needs to be saved after each epoch ends
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
---------
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
* fix style
* fix style
* fix style
* [shardformer] hotfix attn mask (#5945)
* [shardformer] hotfix attn mask (#5947)
* [Feat] Distrifusion Acceleration Support for Diffusion Inference (#5895)
* Distrifusion Support source
* comp comm overlap optimization
* sd3 benchmark
* pixart distrifusion bug fix
* sd3 bug fix and benchmark
* generation bug fix
* naming fix
* add docstring, fix counter and shape error
* add reference
* readme and requirement
* [zero] hotfix update master params (#5951)
* [release] update version (#5952)
* [Chat] Fix lora (#5946)
* fix merging
* remove filepath
* fix style
* Update README.md (#5958)
* [hotfix] Remove unused plan section (#5957)
* remove readme
* fix readme
* update
* [test] add mixtral for sequence classification
* [test] add mixtral transformer test
* [moe] fix plugin
* [test] mixtra pp shard test
* [chore] handle non member group
* [zero] solve hang
* [test] pass mixtral shardformer test
* [moe] implement transit between non moe tp and ep
* [zero] solve hang
* [misc] solve booster hang by rename the variable
* solve hang when parallel mode = pp + dp
* [moe] implement submesh initialization
* [moe] add mixtral dp grad scaling when not all experts are activated
* [chore] manually revert unintended commit
* [chore] trivial fix
* [chore] arg pass & remove drop token
* [test] add mixtral modelling test
* [moe] implement tp
* [moe] test deepseek
* [moe] clean legacy code
* [Feature] MoE Ulysses Support (#5918)
* moe sp support
* moe sp bug solve
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
---------
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
* [chore] minor fix
* [moe] init moe plugin comm setting with sp
* moe sp + ep bug fix
* [moe] finalize test (no pp)
* [moe] full test for deepseek and mixtral (pp + sp to fix)
* [chore] minor fix after rebase
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
* [chore] solve moe ckpt test failure and some other arg pass failure
* [moe] remove ops
* [test] fix test: test_zero1_2
* [bug] fix: somehow logger hangs the program
* [moe] deepseek moe sp support
* [test] add check
* [deepseek] replace attn (a workaround for bug in transformers)
* [misc] skip redunant test
* [misc] remove debug/print code
* [moe] refactor mesh assignment
* Revert "[moe] implement submesh initialization"
This reverts commit 2f9bce6686.
* [chore] change moe_pg_mesh to private
* [misc] remove incompatible test config
* [misc] fix ci failure: change default value to false in moe plugin
* [misc] remove useless condition
* [chore] docstring
* [moe] remove force_overlap_comm flag and add warning instead
* [doc] add MoeHybridParallelPlugin docstring
* [moe] solve dp axis issue
* [chore] remove redundant test case, print string & reduce test tokens
* [feat] Dist Loader for Eval (#5950)
* support auto distributed data loader
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
* support auto distributed data loader
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
* fix tp error
* remove unused parameters
* remove unused
* update inference
* update docs
* update inference
---------
Co-authored-by: Michelle <qianranma8@gmail.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
* [lora] lora support hybrid parallel plugin (#5956)
* lora support hybrid plugin
* fix
* fix
* fix
* fix
* fp8 operators for compressed communication
cast_to_fp8, cast_from_fp8, all_reduce_fp8
* fix scaling algorithm in FP8 casting
* support fp8 communication in pipeline parallelism
* add fp8_communication flag in the script
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
* fix typo
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
* shardformer fp8
* fix rebase
* remove all to all
* fix shardformer fp8 communication training degradation
* [fp8] support all-gather flat tensor (#5932)
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
* fix
* Update low_level_optim.py
---------
Co-authored-by: YeAnbang <anbangy2@outlook.com>
Co-authored-by: Haze188 <haze188@qq.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Edenzzzz <wenxuan.tan@wisc.edu>
Co-authored-by: Edenzzzz <wtan45@wisc.edu>
Co-authored-by: Runyu Lu <77330637+LRY89757@users.noreply.github.com>
Co-authored-by: Guangyao Zhang <xjtu521@qq.com>
Co-authored-by: YeAnbang <44796419+YeAnbang@users.noreply.github.com>
Co-authored-by: Hongxin Liu <lhx0217@gmail.com>
Co-authored-by: Stephan Kö <stephankoe@users.noreply.github.com>
Co-authored-by: アマデウス <kurisusnowdeng@users.noreply.github.com>
Co-authored-by: Tong Li <tong.li352711588@gmail.com>
Co-authored-by: zhurunhua <1281592874@qq.com>
Co-authored-by: Insu Jang <insujang@umich.edu>
Co-authored-by: Gao, Ruiyuan <905370712@qq.com>
Co-authored-by: hxwang <wang1570@e.ntu.edu.sg>
Co-authored-by: Michelle <qianranma8@gmail.com>
Co-authored-by: Wang Binluo <32676639+wangbluo@users.noreply.github.com>
Co-authored-by: HangXu <hangxu0304@gmail.com>
This commit is contained in:
flybird11111
GitHub
189
tests/test_legacy/test_moe/test_moe_load_balance.py
Normal file
189
tests/test_legacy/test_moe/test_moe_load_balance.py
Normal file
@@ -0,0 +1,189 @@
|
||||
import pytest
|
||||
import torch
|
||||
import torch.distributed as dist
|
||||
|
||||
import colossalai
|
||||
from colossalai.booster import Booster
|
||||
from colossalai.booster.plugin import LowLevelZeroPlugin
|
||||
from colossalai.booster.plugin.low_level_zero_plugin import LowLevelZeroModel
|
||||
from colossalai.legacy.moe.manager import MOE_MANAGER
|
||||
|
||||
# from colossalai.shardformer.layer.moe import apply_load_balance
|
||||
from colossalai.tensor.moe_tensor.api import is_moe_tensor
|
||||
from colossalai.testing import rerun_if_address_is_in_use, spawn
|
||||
from tests.test_moe.moe_utils import MoeGradientHandler, MoeModel
|
||||
|
||||
|
||||
def split_ddp_grad(grad, world_size):
|
||||
with torch.no_grad():
|
||||
grad = grad.clone().detach().flatten()
|
||||
padding_size = (world_size - grad.numel() % world_size) % world_size
|
||||
if padding_size > 0:
|
||||
grad = torch.nn.functional.pad(grad, [0, padding_size])
|
||||
splited_grad = grad.split(grad.numel() // world_size)
|
||||
return splited_grad
|
||||
|
||||
|
||||
def run_fwd_bwd(model, data, label, criterion, optimizer, enable_autocast=False):
|
||||
model.train()
|
||||
with torch.cuda.amp.autocast(enabled=enable_autocast):
|
||||
if criterion:
|
||||
y = model(data)
|
||||
loss = criterion(y, label)
|
||||
else:
|
||||
loss = model(data, label)
|
||||
loss = loss.float()
|
||||
|
||||
if isinstance(model, LowLevelZeroModel):
|
||||
optimizer.backward(loss)
|
||||
else:
|
||||
loss.backward()
|
||||
return y
|
||||
|
||||
|
||||
def run_zero_optim_test(local_rank, world_size, stage=1):
|
||||
criterion = torch.nn.CrossEntropyLoss()
|
||||
|
||||
MOE_MANAGER.__init__()
|
||||
MOE_MANAGER.setup(
|
||||
parallel="EP",
|
||||
)
|
||||
zero_model = MoeModel(enable_load_balance=True)
|
||||
zero_optimizer = torch.optim.Adam(zero_model.parameters())
|
||||
plugin = LowLevelZeroPlugin(stage=stage, precision="bf16", verbose=True)
|
||||
booster = Booster(plugin=plugin)
|
||||
zero_model, zero_optimizer, _, _, _ = booster.boost(zero_model, zero_optimizer)
|
||||
|
||||
MOE_MANAGER.__init__()
|
||||
MOE_MANAGER.setup(parallel="EP")
|
||||
torch_model = MoeModel()
|
||||
for zero_param, torch_param in zip(zero_model.parameters(), torch_model.parameters()):
|
||||
torch_param.data.copy_(zero_param.data)
|
||||
torch_optimizer = torch.optim.Adam(torch_model.parameters())
|
||||
torch_model = torch_model.cuda().bfloat16()
|
||||
grad_handler = MoeGradientHandler(torch_model)
|
||||
|
||||
# run to update expert load
|
||||
data = torch.randn(16, 4).cuda().bfloat16() / 1000 / (local_rank + 1)
|
||||
label = torch.randint(0, 4, (16,)).cuda()
|
||||
|
||||
# run torch model twice
|
||||
run_fwd_bwd(torch_model, data, label, criterion, None)
|
||||
grad_handler.handle_gradient()
|
||||
torch_optimizer.step()
|
||||
torch_optimizer.zero_grad()
|
||||
run_fwd_bwd(torch_model, data, label, criterion, None)
|
||||
grad_handler.handle_gradient()
|
||||
|
||||
# get optim and load status in zero model
|
||||
run_fwd_bwd(zero_model, data, label, criterion, zero_optimizer)
|
||||
zero_optimizer.step()
|
||||
zero_optimizer.zero_grad()
|
||||
with torch.no_grad():
|
||||
origin_out = zero_model(data)
|
||||
|
||||
# load balance
|
||||
apply_load_balance(zero_model, zero_optimizer)
|
||||
|
||||
# run again to test
|
||||
zero_out = run_fwd_bwd(zero_model, data, label, criterion, zero_optimizer)
|
||||
torch.allclose(origin_out, zero_out)
|
||||
|
||||
# assert optim
|
||||
torch_optimizer.step()
|
||||
torch_out = run_fwd_bwd(torch_model, data, label, criterion, None)
|
||||
zero_optimizer.step()
|
||||
zero_out = run_fwd_bwd(zero_model, data, label, criterion, zero_optimizer)
|
||||
assert torch.allclose(zero_out, torch_out, atol=3e-5), f"zero_out:{zero_out}\ntorch_out{torch_out}"
|
||||
|
||||
|
||||
def run_hybrid_zero_optim_test(local_rank, world_size, stage=1):
|
||||
criterion = torch.nn.CrossEntropyLoss()
|
||||
data = torch.randn(16, 4).cuda()
|
||||
label = torch.randint(0, 4, (16,)).cuda()
|
||||
|
||||
MOE_MANAGER.__init__()
|
||||
MOE_MANAGER.setup(parallel=None)
|
||||
torch_model = MoeModel()
|
||||
torch_optimizer = torch.optim.Adam(torch_model.parameters())
|
||||
torch_model = torch_model.cuda()
|
||||
|
||||
MOE_MANAGER.__init__()
|
||||
MOE_MANAGER.setup(
|
||||
max_ep_size=2,
|
||||
use_ep_inside=False,
|
||||
parallel="EP",
|
||||
)
|
||||
zero_model = MoeModel(enable_load_balance=True)
|
||||
extra_dp_group = MOE_MANAGER.parallel_info_dict[2].dp_group
|
||||
ep_rank = dist.get_rank(MOE_MANAGER.parallel_info_dict[2].ep_group)
|
||||
ep_size = MOE_MANAGER.parallel_info_dict[2].ep_size
|
||||
for zero_param, torch_param in zip(zero_model.parameters(), torch_model.parameters()):
|
||||
if is_moe_tensor(zero_param):
|
||||
num_expert = torch_param.data.shape[0]
|
||||
zero_param.data.copy_(
|
||||
torch_param.data[ep_rank * (num_expert // ep_size) : (ep_rank + 1) * (num_expert // ep_size)]
|
||||
.detach()
|
||||
.clone()
|
||||
)
|
||||
else:
|
||||
zero_param.data.copy_(torch_param.data.detach().clone())
|
||||
zero_optimizer = torch.optim.Adam(zero_model.parameters())
|
||||
plugin = LowLevelZeroPlugin(stage=stage, precision="fp32")
|
||||
plugin.zero_optim_kwargs["moe_extra_dp_process_group"] = extra_dp_group
|
||||
booster = Booster(plugin=plugin)
|
||||
zero_model, zero_optimizer, _, _, _ = booster.boost(zero_model, zero_optimizer)
|
||||
|
||||
# run torch for twice
|
||||
run_fwd_bwd(torch_model, data, label, criterion, None)
|
||||
torch_optimizer.step()
|
||||
torch_optimizer.zero_grad()
|
||||
run_fwd_bwd(torch_model, data, label, criterion, None)
|
||||
torch_optimizer.step()
|
||||
|
||||
# run zero
|
||||
run_fwd_bwd(zero_model, data, label, criterion, zero_optimizer)
|
||||
zero_optimizer.step()
|
||||
zero_optimizer.zero_grad()
|
||||
with torch.no_grad():
|
||||
origin_out = zero_model(data)
|
||||
|
||||
# load balance
|
||||
apply_load_balance(zero_model, zero_optimizer)
|
||||
|
||||
# assert out
|
||||
zero_out = run_fwd_bwd(zero_model, data, label, criterion, zero_optimizer)
|
||||
torch.allclose(origin_out, zero_out)
|
||||
|
||||
# assert optim
|
||||
zero_optimizer.step()
|
||||
zero_out = run_fwd_bwd(zero_model, data, label, criterion, zero_optimizer)
|
||||
torch_out = run_fwd_bwd(torch_model, data, label, criterion, None)
|
||||
# TODO: high atol, check if bug exists
|
||||
assert torch.allclose(zero_out, torch_out, atol=8e-4), f"zero_out:{zero_out}\ntorch_out{torch_out}"
|
||||
|
||||
|
||||
def run_dist(rank, world_size, port):
|
||||
colossalai.launch(
|
||||
rank=rank,
|
||||
world_size=world_size,
|
||||
host="localhost",
|
||||
port=port,
|
||||
backend="nccl",
|
||||
)
|
||||
run_zero_optim_test(rank, world_size, stage=1)
|
||||
run_zero_optim_test(rank, world_size, stage=2)
|
||||
run_hybrid_zero_optim_test(rank, world_size, stage=1)
|
||||
run_hybrid_zero_optim_test(rank, world_size, stage=2)
|
||||
|
||||
|
||||
@pytest.mark.skip(reason="moe need to be refactored")
|
||||
@pytest.mark.dist
|
||||
@pytest.mark.parametrize("world_size", [4])
|
||||
@rerun_if_address_is_in_use()
|
||||
def test_moe_load_balance(world_size):
|
||||
spawn(run_dist, world_size)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
test_moe_load_balance(world_size=4)
|
||||
Reference in New Issue
Block a user